Choose timezone
Your profile timezone:
Powered by Indico
v3.3.2-pre
Welcome to the Indico interface for the 2018 IEEE IPMHVC. At this site, you will be able to perform the following actions related to the conference:
Further information about the conference, including the categories of the technical program, important dates, and contact information, we refer you to the main conference website.
Breakfast in Exhibitors Area
James C. Dickens, Nicholas A. Wilson, Vincent Meyers, Daniel Mauch, Sergey Nikishin, Ravi Joshi, and Andreas A. Neuber
Center for Pulsed Power and Power Electronics
Texas Tech University
Lubbock, TX 79409 USA
Wide bandgap semiconductor photoconductive switches (PCSSs) have unique switching capabilities that push beyond any other type of switching device. In particular, PCSSs were demonstrated to switch tens of kV with picosecond response times on turn-on and turn-off as well as very low jitter. These unique capabilities make PCSSs a promising technology across a broad range of applications, including medical and industrial accelerators, high power RF generation via photonic to RF conversion, THz generation, and sub-picosecond rise-time pulse generation. While multiple technological issues have plagued PCSS switching, its limited quantum efficiency (defined here as ratio of photons to ultimately transferred charge carriers) is still its prominent weakness.
Texas Tech University has had an ongoing research program in uncovering the dominant physics and improving wide bandgap PCSSs for the past decade. Details of this very active research effort will be presented including generational switch performance, fabrication and material technology as well as device and material characterization. In addition, an introduction to PCSS device physics, wide bandgap material types and device modeling will be included as a brief primer, which will also address a brief summary of PCSS efforts from other institutes.
To further the goal of optimized material design for pulsed power components, we aim to achieve a fundamental understanding of the model nonlinear dielectric material BaTiO3, through concurrent experimental and theoretical study. This effort is intended to lead to improved synthesis and design control, and a validated model to enable material predictions. The ultimate goal is to improve energy storage and discharge characteristics through design models for dielectric materials that translate results from the molecular level to macroscopic device level.
We examined the bonding and arrangement of selected ligands and molecules adsorbed on BaTiO3 surfaces using both density functional theory (DFT) and experimental techniques. Bond strengths and reaction mechanisms were examined in detail. Initial DFT results showed that the energy involved in the polarization response in a ferroelectric BaTiO3 slab is similar to or greater than the local adsorption or reaction energy of the molecule with the top layer of the slab. Several approaches are applied for systematically distinguishing these effects and extracting molecular desorption barriers, including thermogravimetric analysis (TGA) and Fourier-transform infrared spectroscopy (FTIR) experiments.
This work was supported by AFOSR FA9550-16DCOR281. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia LLC, a wholly owned subsidiary of Honeywell International Inc. for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.
For several years the University of Missouri has been developing compact capacitors for use in high voltage pulsed power/directed energy applications. The dielectric employed in this development is a proprietary nanocomposite, nanodielectric material, MU100. The material was originally developed for use in dielectric loaded antennas, however, due to various material properties, the nanocomposite has shown promise in development of compact high voltage capacitors. Prior work has shown small scale samples of the high permittivity nanocomposite dielectric material to have an average dielectric strength of 225 kV/cm with peak breakdown fields in excess of 325 kV/cm. When scaling up to accommodate application specific voltages, failure modes become more pronounced due to volume effects of the nanocomposite and field enhancement factors at the electrode dielectric interfaces. This paper will present how the material was scaled from small scale samples up to compact capacitor prototypes capable of repeatable performance at 250 kV with lifetimes greater than 104 shots.
Keywords: Nanodielectrics, Dielectric, Capacitors, High Voltage, Nanocomposites, Pulsed Power System
Distribution A: Approved for Public Release
Surface trap parameter and distribution can significantly affect the development of the secondary electron emission avalanche (SEEA), the most widely accepted mechanism for flashover in vacuum. To further understand the influence of traps on the flashover of polymeric materials under different degrees of aging, polytetrafluoroethylene (PTFE), polymethylmethacrylate (PMMA) and polyamide (PA6) were aged by repeated single-shot flashover, which was excited by a microsecond pulsed generator. The finger-like plane-electrodes were used in this experiment with 1.2 mm gap. The trap parameters of specimens experienced 1000, 2000, 3000 times of flashovers, were respectively tested by means of isothermal surface potential decay (ISPD) and the flashover voltage was also recorded. The results show that the deeper energy level trap occurs and the trap charge density is lower after 1000 times of flashovers for PTFE and PA6. The deep trap and shallow one exist simultaneously in three materials. The trap parameters of PMMA don’t change obviously compared with PTFE and PA6. The trap energy level is deeper and the trap charge density is higher after 3000 times of flashovers for three materials. In order to analyze the effect of trap parameters on discharge development, the dynamics process of charge trapping and detrapping were analyzed based on the charge transport model in dielectrics with single trap level and two discrete trap levels. The simplied calculation results show that the deeper the trap level is, the quantity of trapped charges is more in polymeric materials. The trapped probability of charged particles for shallow level trap is small in insulating materials.
Electric pulses (EPs) have multiple medical applications that depend upon their impact on plasma membrane integrity and ion and molecular transport [1]; however, the motion of the ions during the EPs, which may impact the resulting phenomena, remain poorly understood. This study reports the net ion motion during the application of multiple nanosecond EPs (NSEPs) to a Jurkat cell suspension by determining the electrical conductivity by measuring the applied voltage and resulting current during the final EP of the train. The cells were suspended in a high conductivity buffer, growth media (GM), and a low conductivity buffer (LCB) [2] and exposed to trains of one, five, and fifteen 60 ns EPs of fixed energy while 300 ns pulses of fixed energy were applied to LCB. For 60 ns EPs, the extracellular conductivity increased for the higher conductivity buffers and decreased for the LCB, indicating ion motion out of and into the cells, respectively. Applying a train of 300 ns EPs to LCB also increased extracellular conductivity. Calculations indicate that the variation in extracellular conductivity observed during the shorter EP and lower conductivity buffer arises due to non-electrical effects, such as shocks waves, membrane temperature gradients, and colloid-osmotic swelling. The potential significance of these multiphysics phenomena on fundamental biophysical phenomena and the implications on biomedical applications will be discussed.
[1] M. L. Yarmush, A. Golberg, G. Serša, T. Kotnik, and D. Miklavčič, “Electroporation-based technologies for medicine: Principles, applications and challenges,” Ann. Rev. Biomed. Eng., vol. 16, pp. 295-320, 2014.
[2] A. L. Garner et al., “Ultrashort electric pulse induced changes in cellular dielectric properties,” Biochem. Biophys. Res. Commun., vol. 362, no. 1, pp. 139–44, 2007.
Extraction of lipids from microalgae is a major barrier to the industrial production of lipid-derived biodiesel, motivating efforts to implement treatment strategies to replace or supplement solvent extraction. Pulsed electric field (PEF) treatment has shown promise in weakening the cell membrane to facilitate greater lipid yield by subsequent solvent extraction. Previous studies have assessed microsecond duration pulses with fields between 10 and 60 kV/cm [1].
This study assesses the impact of pulse duration on PEF facilitated lipid extraction for nanosecond electric pulses. In particular, we assess the impact of total energy delivered by fixing the pulse duration (60 ns) and electric field (60 kV/cm) and applying 10, 50, 100, 200, and 300 pulses. Lipid extraction increased significantly compared to the unpulsed control by 14.6%, 16.8%, and 19.2% following 10, 50, and 100 pulses, respectively. However, applying 300 pulses resulted in a statistically significant 9.4% decrease compared to control, suggesting a peak efficiency in PEF enhanced lipid extraction. The implications of these results will be discussed and compared to ongoing experiments at 100 microseconds and 5 kV/cm with the same energy delivered.
Recent research has highlighted multiple opportunities for activated platelet rich plasma (PRP)/platelet gel in wound healing [1]. This entails drawing the patient’s blood, centrifuging it to separate the PRP, and activating it by adding bovine thrombin (BT), the current state of the art clinical platelet activator, and CaCl2 to trigger platelet activation (growth factor release and clotting). The activated PRP is then applied onto a patient’s wound. While efficient, BT may trigger side effects [2] and create challenges for cost, repeatability and workflow.
Electric pulse stimulation (EPS) is a potential candidate to replace BT by mitigating these drawbacks [3]. While initial studies suggested that EPS required CaCl2 for platelet activation, subsequent research demonstrated that one can achieve growth factor release without clotting with EPS and no added CaCl2 [3]. This observation could open new opportunities for injectable PRP since BT triggers growth factor and clotting and cannot be used for these applications while one could use EPS to activate PRP before injection to facilitate growth factor delivery at the injury site.
The current study reports the relationship between electrical parameters (pulse width, pulse amplitude, conductive vs. capacitive coupling) and CaCl2 level on growth factor release and clotting. Depending upon CaCl2 level and EPS, one can achieve various levels of growth factor release with no clotting or growth factor release with clotting. Furthermore, clotting time and clot mechanical strength may be controlled by choosing appropriate CaCl2 level and pulse parameters. For instance, combining 5.35 mM of CaCl2 with a 1.5 kV, 300 ns pulse triggers no clotting and little growth factor release while applying a 1.75 kV, 5 us pulse causes growth factor release with no clotting. Adding more CaCl2 to the microsecond pulses eventually triggers clotting. This demonstrates the potential for EPS as the first ever designer platelet gel technique.
[1] K. M. Lacci and A. Dardik, “Platelet-rich plasma: Support for its use in wound healing,” Yale J. Biol. Med., vol.83, pp. 1-9, 2010.
[2] D. L. Diesen and J. H. Lawson, “Bovine thrombin: history, use, and risk in the surgical patient,” Vascular, vol. 16, Suppl. 1, pp. S29-36, 2008.
[3] V. B. Neculaes, A. S. Torres, A. Caiafa, B. D-L. Lee, and A. L. Garner, “Platelet activation and growth factor release using electric pulses,” US Patent 9 452 199, Sep. 27, 2016
To meet the needs of high voltage pulsed electric field for tumor therapy and tissue ablation，a design scheme of high voltage composite pulsed wave source combining classic Marx generator and all solid state switching device is proposed in this paper.Using all solid state switching devices instead of traditional spark gap switches, the unipolar Marx generator is the core of circuit. The output of composite pulse square wave can be realized by the relay switching. In this paper, the structure, working process, control strategy and load adaptability of the circuit are analyzed in detail. using solid state MOSFET as the main switch device, the main circuit part of the pulse source is developed. The corresponding control circuit and switch synchronous trigger circuit are designed, and the signal transmission and the isolation of strong and weak electricity are realized through optical fiber and isolated power supply module.Compared with the conventional high-voltage pulse source, the scheme has a more compact circuit structure and good load adaptability. It achieves the advantages of higher output pulse front, higher pulse frequency, pulse width, and adjustable voltage amplitude.The experimental results show that the system can generate high voltage pulse source amplitude range of 0 ~ 7 K V, 0 ~ 1kHz frequency, pulse width range 200ns ~ 1μs square wave pulse, low voltage can generate the amplitude range of 0 ~ 1 K V, 0 ~ 10kHz frequency, pulse width range of 1 μs ~ 100 μs square wave pulse. In order to find the best electrical parameters, it provides the hardware support for the tissue ablation of composite pulsed electric field and the treatment of tumor.
Irreversible electroporation (IRE) uses ~100 μs pulsed electric fields to permanently disrupt cell membranes for tumor treatment. Combining short high-voltage (SHV) pulses with long low-voltage (LLV) pulses, a new modality of the pulsed electric fields exhibits the synergistic effect which could enhance the cell cytotoxic effect. Here, the ablation by 2D monolayers mimic in vitro model and rabbit liver in vivo model was investigated. For 2D monolayers mimic in vitro model, combining SHV pulses with LLV pulses produced larger ablation area than that by either SHV pulses or LLV pulses applied respectively. Even when applying the same pulse number with either SHV pulses or LLV pulses, combining SHV pulses with LLV pulses also produce larger ablation region. For rabbit liver in vivo model, Combining SHV pulses with LLV pulses yielded an ablation area of 50.70 mm2, which was increased by 112.58% relative to that after SHV pulses applied alone and 134.18% relative to that after LLV pulses applied alone. With the same pulse number, the ablation area by SHV+LLV pulses was still 75.92% larger than that after LLV pulses. Particularly, when the lag time between the SHV pulses and LLV pulses protocols was adjusted to 100 s, the ablation region would be further decreased (32.21%). However, SHV pulses and LLV pulses sequence mattered, LLV + SHV pulses could not decrease cell viability. The combining SHV pulses with LLV pulses exhibited the synergistic effect, which is that the SHV pulses has a stronger electric field that creates a larger electroporated area in liver tissue and make it more susceptible to the subsequent LLV pulses, then resulting in highly efficient tissue ablation.
On your own for lunch
Laboratory equipment is sometimes required to be calibrated in-situo and, in those scenarios, a pulse with frequency content exceeding the intended voltage source is needed. Moreover, for sensors that are highly attenuated, the calibration pulser must also be high voltage and very reproducible. Because the calibration pulser is intended to be used across a wide variety of experiments, it has been made to be independent of the load impedance- known as a self-matching circuit. This variable high voltage self-matching pulser gives the user the ability to analyze his or her equipment using a consistent pulse. The self-matching aspect of this device is what makes it so unique and consistent. The pulser is made using a transmission line whose impedance is matched to several resistors. These resistors are responsible for absorbing the reflections that are generated by the load/transmission line interface. Because of this, the users output will show the voltage pulse across the load and the reflections at this interface will travel to the transmission line/resistor interface where the pulse will be absorbed due to matching impedances. In practice however, there are slight reflections generated due to a mismatch between the resistors and the transmission line, but because the pulse itself is so consistent the device can still be used for calibration purposes. The variable high voltage self-matching pulser can be used between .5 kilovolts and 30 kilovolts and although the amplitude of the pulse may change with the variably the shape of the pulse does not change. It has proven to be a useful calibration tool as well as a pulsed source.
As DC distribution system is developing rapidly, DC series arc fault has become a severe threat to safe operation of DC systems, since DC arc cannot extinct by itself. Moreover, DC series arc fault is difficult to be located, as the high frequency components of the arc current that are superimposed on the normal DC current cannot be extracted in some cases.
In this paper, a DC series arc fault location method is proposed by using time lag between parallel capacitor current pulses at both ends of the DC cable. Capacitors are utilized to couple the high frequency current that can indicate the presence of the arc fault and avoid the interference of normal DC current. High frequency current arrives at two capacitors paralleled with the cable ends at different time when series arc fault occurs in the cable. The arriving time of the fault current pulse can be accurately determined by using the rising edge of the pulse. Thus, the time lag can be used to locate the arc fault.
Models of DC distribution system with different cable configurations are built in PSCAD. Then DC arc test platform is constructed consisting of DC power supply, cables (300 m) and load. Capacitors are parallel with the two ends of the cables. Rogowski coils are adopted as the capacitor current sensors in the system, and bandwidth of the coils is from 100 Hz to 1000 kHz, according to the characteristics frequency band of the fault capacitor current. The arc faults are generated at different locations.
For arc faults at different locations, the largest location error is less than 15 m. The test results indicate that the proposed method can locate DC series arc fault accurately.
Insulator surface flashover in vacuum due to pulsed and DC voltages has been thoroughly investigated, however the characteristics of surface flashover under multiple, short duration, high-voltage pulses are not well understood. A test stand has been developed at Los Alamos National Lab to further understand the effects of these conditions on the surface flashover strength of various insulators. The test stand is designed to apply voltages of up to 350 kV across a variable spacing, uniform field,electrode gap. Models indicate the electric field varies by less than ±2 percent over a 4 cm cylindrical radius at up to 4 cm gap spacing. The electrodes are designed to accommodate insulator test samples ranging in axial length from 1 - 4.5 cm, and up to 8 cm diameter. Measurements of breakdown field and current will be made with an E-dot probe and Pearson current transformer respectively. This paper details the design, modeling, and testing plan for this multi-pulse test-bed.
Charges accumulated on the dielectric surface under DC field can lead to the overstress of the insulator due to the local field distortion, and may even initiate the surface flashover. Therefore, it is very necessary to study the surface charge accumulation and decay on polymer materials under DC voltage in order to get a theoretical basis for the design of DC composite insulators, considering that polymer materials are widely used in the electrical field. Since then, this paper chooses silicon rubber as the study object and focuses on the accumulation and decay of surface charge on it.
Since the output of the Kelvin probe is the electric potential instead of the charge density. In this paper, the general principle of the surface charge calculating method is analyzed based on the active capacitance probe method and the image processing technique. The experimental subject is a shift-invariant system and the transfer function matrix H is simplified based on the properties of it. Furthermore, two-dimensional Fourier transform and Wiener filter techniques are applied in this paper, thus the relationship between potential and charge density can be processed in spatial frequency domain. Afterwards, the accuracy of the measurement and inverse algorithm are verified by Dust Figure Method.
On the one hand, the surface charge accumulation is rapid and surface potential reaches nearly the applied voltage on the silicon rubber material which has a quite serious accumulation charge distribution. On the other hand, we inject charge into the surface of the silicon rubber sample by using a needle-plate electrode to study the decay phenomenon. The study is conducted under different humidity conditions, which reaches that the surface charge mainly decays by volume resistivity under low humidity and the initial surface charge distribution changes its shape when decay process occurs in high humidity air.
The interruption performance of the Molded Case Circuit Breaker (MCCB) is important because this circuit breaker is closest protection device to consumer at the power distribution system. Basically, the interruption process of the MCCB is follows: when the fault current inflows, the trip unit detects it and operates stator to separate fixed electrode and moving electrode. Next, the arc stretches with bending toward the splitter plate after the arc discharge occurs between electrodes. finally, the arc is extinguished between plates by cooling, dividing and stretching when it reaches the splitter plate. In whole process of interruption, the factors that the arc stretches with bend toward the splitter plate are the Lorentz force, the arc runner and the gas pressure generated by hot-gas. Therefore, by improving these factors, the arc extinguishment at the splitter plate can be more fast and it leads to advance an interruption performance of the MCCB.
In this paper, to improve the Lorentz force, an external magnetic field is applied to both side of electrodes. The experiment circuit to make the over-current for activating the interruption performance of the MCCB is implemented and the experiments applied an external magnetic field are done by using ready-made product. As a result, the time between contact separation and splitter plate reach is shorter and the interruption process is faster. It is expected that these results help an optimization of design and improvement of performance.
In South China and Southwest China, 10 kV switchgear running under high temperature and high humidity, which is easier to induce switchgear partial discharge and then benefit the development of insulation fault, seriously affect the safe operation of power grid. At this stage, the protective measures of the switchgear insulation are limited to preventive measures and can’t protect the insulation of the switchgear. Furthermore, semiconductor materials have been applied to improve the electric field distribution, which can effectively inhibit the occurrence of partial discharges. In this paper, the application of the semiconductor coating in the electric field distortion of the switchgear was simulated, the best coating scheme and the parameters of the semiconductor coating were obtained. Simultaneously, artificial contamination experiments were carried out with different combinations of semiconductor material and RTV coating to verify the effectiveness of partial discharge inhibition of 10kV switchgear. The application of both semiconductor material and RTV Coating can effectively inhibit the occurrence of the partial discharge of 10kV switchgear.
Dielectrophoresis has become a very important tool in the aspect of utilizing microscopic dielectric particles in recently years. More and more developments have been achieved in various subjects, such as: chemistry, life science, astronomy and engineering. In these developments, manipulation of these small dielectric particles is the main problem due to quantifying FDEP (dielectrophoretic forces). Point-Dipole (PD) and Maxwell Stress Tensor (MST) methods are commonly used for quantifying in current researches. PD method is a simplified way which can be used in the condition that particles are all homogenous and spherical, while their existences don’t change external electric fields. MST method is a more exact way for the calculation which is able to consider particles’ shape and the deformation of particles to external electric fields, however, it is limited to homogenous particles. In order to overcome the limitations above, a new method VEM (Volumetric Element Method) was introduced into the calculation. With the method, FDEP on a homogenous and non-homogenous particles with different shapes (sphere, ellipsoid, square) were calculated under an electric field. Differences among the three methods are compared and analyzed. Finally, it was found that PD method is not suitable when the deformations of particles on external electric fields were very strong, but VEM and MST still kept a good performance. What’s more, it indicated that variations of shapes and homogeneity changing the volume and dielectric constant of particles would influence FDEP acting on particles.
To enable employment of 3D printed polymeric components in high voltage pulsers and related support systems, compatibility with materials commonly used for insulation and cooling in these systems must be known. Polymeric test samples, printed with four common 3D printing techniques (fused deposition modeling, stereolithography, polymer jetting, and selective laser sintering) were exposed to water or insulating oil environments for extended durations. Preliminary results of compatibility testing of 3D printed parts will include comparisons of material absorption, embrittlement, and changes in dielectric strength between exposed samples and a control group of unexposed samples. Water absorption tests were performed after a 24 hour exposure period. Mechanical and dielectric strength testing was performed on test samples exposed to insulating oil environments for a nine month period.
with the development of power grid, the practicability and reliability of the evaluation method of transformer, as the key equipment of power network, are paid more and more attention. Aiming at the problem that some evaluation indexes are not easy to quantify in transformer evaluation system, a transformer condition assessment method based on cloud matter element and cooperative game is proposed in this paper. Firstly, a relatively complete evaluation index system is established, and the weight of each index is calculated by AHP, entropy and neural network, and then the weights of the three methods are combined by cooperative game , and the weight of each index are obtained . Secondly, the state of transformer is evaluated by cloud model and matter-element extension theory. Finally, a case analysis is carried out to verify the rationality and effectiveness of the method.
Partial discharge(PD) in gas-insulated switchgear(GIS) includes electrical signals with broad bandwidth. In order to simulate and study the propagation characteristics of specific frequency band, pulses with rise time on the order of subnanosecond are required. We have developed a pulse generator which provides 240ps rise time with an amplitude up to 2kV into a 50Ω coaxial capactive load with an extensive variation of amplitude. The pulse generator consists of a mercury switch. The rise time of the output pulse from the generator is reduced shorter than 300ps due to precise designs of PCB circuit board, tail-cut capacitor and a head resistance. The tail-cut capacitor and the head resistance determine the duration and amplitude of the pulse, meanwhile a slide rheostat can adjust the amplitude independently. By adjusting the input voltage and parameters of adjustable components, different FWHM, rise time and amplitude of the output pulse are obtained. The shortest rise time 240ps is recorded with a 480ps pulse duration while the longest 980ps is recorded with 10ns pulse duration. It is possible to study different propagation characteristics of both high and low frequency signals as well as large and short band discharge signals.
Nonlinear transmission lines (NLTLs) have been the subject of several studies that have shown their suitability in high-speed systems for a wide variety of applications as pulse compression, phase shifter, frequency multiplier, pulsed radar, battlefield communication disruption and also as a high power microwave source, holding in this case an additional way to generate radio frequency (RF) signals without using vacuum electronic tubes, which normally requires heating filament and bias power supplies. NLTLs, which are designed for the RF generation, uses nonlinear dielectric and magnetic materials that are arranged to form a nonlinear medium that can be dispersive or continuous (non-dispersive). This paper presents a summary of the main research on the recent NLTLs development and provides an analysis of the influence of the nonlinear material characteristics on the performance of NLTLs, pointing that there is a lack of nonlinear dielectric and magnetic materials that would allow the achievement of NLTLs with better RF conversion efficiency and the operation at higher frequencies even under adverse environmental conditions. The nonlinear dielectric and magnetic materials for NLTLs applications need to have characteristics such as highly nonlinear behavior, low losses, and thermal stability.
*Work Supported by US Air Force Office of Scientific Research under contract no. FA9550-18-1-0111.
I. Background
Transformer’s insulation structure profoundly determines device’s lifetime. As novel method, Frequency Domain Spectroscopy, which based on the principle of dielectric response, attracts researchers all over the world for its advantages of richer information details, excellent interference immunity and non-destructive. However, Frequency Domain Spectroscopy's characteristic parameters related to insulating condition are not easy to extract, which makes it limited to utilize on site. Based on that, The concept of dielectric modulus is introduced to assess the transformer's insulation condition in this thesis.
II. Method
Based on theoretical analysis and deduction, the physical meaning of dielectric modulus spectrum is explained. Then, the differences and connections between frequency domain spectroscopy and dielectric modulus spectrum are investigated. After that, we prepared oil immersed paper samples with water content of 1%, 2%, 3%, 4% and 6% and obtained their dielectric modulus spectrums by IDAX206. Meanwhile, the influence of temperature is studied by testing dry samples under different temperatures.
III. Results
An obvious relaxation peak related to dielectric material's polarization process reveals in the low band of dielectric modulus spectrum rather than frequency domain spectroscopy. By extracted the frequency of relaxation peak, which is easy to be gained , as the characteristic parameter, the water content of oil-paper insulating structure could be quantitatively evaluated by fitting equation. The relationship between temperature and the characteristic parameter obeys Arrhenius formula， by which the influence of temperature on the dielectric modulus spectrum could be eliminated.
Epoxy resin of saturable reactor for UHVDC converter valves works in high temperature condition for long-term, and it also suffers from bipolar exponential decay pulse voltage erosion. In order to study the electro-thermal aging law of epoxy resin, the electro-thermal aging tests under pulse aging and sinusoidal aging are carried out. Pulse aging voltage is set as bipolar exponential decay pulse with amplitude of 4.5kV, repetition rate of 50Hz, positive rise time of 1.2us, positive width of 12us, negative rise time of 24us and negative width of 74us. Sinusoidal aging voltage is set as sinusoidal voltage with amplitude of 4.5kV and repetition rate of 50Hz. Aging temperature is set as 110 degrees centigrade. Aging time is set as 20h to 140h with a step length of 20h. The results show that with the increase of aging time, the dielectric constant and dielectric loss tangent value of epoxy resin increase obviously in the 10-1~103 Hz frequency range, and the aging degree of pulse voltage is lower than that of sinusoidal voltage. Therefore, the margin of the design according to the sinusoidal voltage aging is higher than that according to the bipolar pulse voltage under normal operating conditions.
Series arcs can occur in dc power systems whenever an energized wire supplying power to a load is interrupted. This can happen when a wire snaps abruptly or when a load connection becomes loose. Rather than ending in a simple open circuit, the inductance in series with the load establishes an arc, which inserts an additional impedance in between the source and load. The arc impedance decreases load current and complicates detection while still posing a severe hazard. For this reason, the interaction between series arcs and load power converters is of interest. Load converters act as a dynamic input impedance to the arc thus complicating efforts to accurately decouple the interface of arc and load. This paper performs an empirical study of the fundamental interaction between a series arc and fixed impedance, RC load. Both transient and steady state arc behavior are experimentally studied for a fixed series inductance while varying dc supply voltages, load currents, load capacitances as well as terminal gap distances for a spring based separation mechanism. Tests are repeated multiple times at each parameter set to account for the random nature of the arc. It is found that load current and load capacitance dominantly affect the occurrence of sustained series arcs. Results have contributed to formation of a generic arc transient waveform that can be used for further study of the complex interactions between series arcs and power-converter loads.
PRIMA (Padova Research on ITER Megavolt Accelerator) is the ITER Neutral Beam Test Facility in Padova, Italy, for the development of the ITER Neutral Beam Injectors (NBI). It comprises two experiments: MITICA, the full-scale prototype of the NBI, designed to produce and neutralize a 40 A negative ions beam, accelerated up to the energy of 1 MeV, and SPIDER, the full-size negative ions source of the NBI.
ITER NBI includes a Radio Frequency (RF) plasma source where plasma is produced by the inductive coupling with coils wound around complex vacuum chambers called drivers. Each coil is fed at 1 MHz up to a power of 100 kW, which corresponds to a voltage of about 12 kV rms, with nominal plasma parameters.
The voltage hold off in vacuum of the beam source components is one of the most critical issues connected to the fulfillment of the requirements for ITER, not only for the particles acceleration system subjected to very high dc voltage (up to 1 MV) but also for the RF circuits of the plasma source and in particular the RF drivers.
The development of a simple, accessible and flexible device called “High Voltage Radio Frequency Test Facility” (HVRFTF) was launched to effectively characterize the voltage hold off of the RF drivers of SPIDER and MITICA, subjected to radiofrequency E-fields at low pressure.
The experimental arrangement worked out to reproduce the driver operating conditions includes a vacuum vessel capable to host different types of Devices Under Test (DUT), a gas injection and pumping system to supply the desired gas species up to the test pressure and a RF circuit designed to produce the high voltage.
The first DUT tested with the HVRFTF is composed of planar circular electrodes, a configuration not directly relevant for the driver but widely treated in literature and significant for the validation of the basic test assessment.
This paper presents the results obtained during the first operations of the HVRFTF with the DUT in an Argon atmosphere in the pressure range 10-3 – 102 Pa. A voltage up to 10 kV rms, at the frequency of 1 MHz was applied to the electrodes, with gaps of 0.1 mm – 40 mm.
Breakdowns were observed for pressures higher than 0.2 Pa, associated with the appearance of a glow discharge between the high voltage electrode and the grounded vacuum vessel, as expected according to the Paschen’s law. For lower pressures the measuring equipment did not detect any breakdown, although sparks were noticed between the electrodes.
Transient plasma ignition (TPI) uses short (typ. nanosecond durations) high voltage pulses to generate highly non-equilibrium plasmas for combustion ignition. TPI allows lean-fuel combustion, improves ignition or combustion efficiency with potentially reduced emission [1,2]. This study evaluates the effect of pulse duration and rise time of >10 kV pulsed plasmas on the ignition plasma formation and combustion chemistry in a static chamber containing methane and dry air at atmospheric pressure. Repetitive, nanosecond, up to 20 kV pulses with different pulse durations (e.g. 10 ns, 30 ns, and 100 ns) and rise times (2 ns, 8 ns, and 80 ns) were used to generate transient plasmas between a pin-to-plate electrode configuration that was the same as a typical spark-plug igniter. Ignition delay, peak pressure and pressure rise time were compared among the transient plasmas driven by different nanosecond pulses. Hydroxyl radicals as important indicators to evaluate combustion chemistry were evaluated using optical emission spectroscopy and laser induced fluorescence. The work has been supported by the Department of Energy (STTR) and in part by the Air Force Office of Scientific Research (FA9550-17-1-0257).
[1] F. Wang et al., "Transient plasma ignition of quiescent and flowing air/fuel mixtures," in IEEE Trans. on Plasma Sci. 33(2), (2005) 844-849.
[2] D. Singleton et al., “The role of non-thermal transient plasma for enhanced flame ignition in C2H4–air,” in J. Phys. D: Appl. Phys. 44 (2011) 022001.
Characteristics of a plasma generated in an arc discharge are investigated. In a discharge, various processes contribute to overall characteristics. Electron chemistry and photonic processes each play a role in establishing the discharge environment based on background pressure and gas species involved. Photonic processes have been incorporated into a PIC-DSMC plasma modeling code, showing effects of including these processes on the discharge current and generating simulated photo-emission spectra. A high-pressure arc discharge experiment was set up to attempt to elucidate mechanisms of charged species interaction during discharge and in the post-arc environment and to compare with model prediction. Discharges generated with inert and reactive gases (nitrogen and air) are studied to determine how the presence of readily reactive species alter the discharge characteristics. In addition, micron-size particles (conductive and dielectric) are injected to study dusty discharge/plasma environment. These particles, with differing electrical characteristics, provide additional sources of interaction with charged species generated in the plasma. Photodetectors and optical emission spectroscopy are used to probe the plasmas and characterize their spectral responses. Furthermore, differentially-charged species in the post-arc environment interact via local electric field, resulting in current flow. Model can simulate/isolate various processes, and discharge behavior can be inferred by measuring dI/dt and compared with predicted observables, showing FFT components associated with this localized current flow due to charged species interaction.
Surface charge accumulation on insulator is one of the main restrictive factors of DC gas insulated pipeline transmission line (DC-GIL). The current research on the accumulation and dissipating mechanism of surface charge has the following deficiencies on the solid side. Firstly, the volume conductivity of insulators are usually taken as a fixed value,so the comprehensive effect of electric field intensity and temperature on this parameter is not taken full consideration. Secondly, the calculation models of current density and surface charge density on the solid side ignore the effects of dielectric relaxation, which also affects the calculation of surface charge density. So in this paper, the mathematical model of solid conductivity based on electric field and temperature has been studied, and the model of volume current density and charge density has been reestablished in combination with dielectric relaxation. When the solid side is dominant, the temperature , electric field and dielectric relaxation characteristics of surface charge accumulation and dissipation are simulated and analyzed. The results indicate that the steady state value of surface charge density increases exponentially with the increase of the initial value of electric field and temperature, and the charge saturation time also increased. The increase of the electric field intensity change rate leads to a exponential decrease of the steady state value of surface charge density and the saturation time. The saturation time increases by 3% and the steady state value of surface charge density increases by 0.9% in the influence of dielectric relaxation.Considering only the charge dissipation of the solid side, the higher the temperature, the more favorable for the dissipation of the surface charge. Moreover, the dissipation rate increases exponentially with temperature. The effect of field strength on charge dissipation depends on the initial charge density .Dielectric relaxation slows the dissipation rate by 28%.
The phenomenon of metal removal from optical discs using pulsed power has been investigated. Pulsed power with 35.3 J/pulse was applied to concentric ring electrodes placed on the optical discs. The protective layer containing the metal layer was completely separated from the plastic substrate by about 30 shots of pulsed power. In order to clarify the mechanism of the metal removal, discharge and shock wave were observed with a high speed camera and a pulse-laser Schlieren system, respectively. As a result, a fan shape discharge was observed after a breakdown took place between the electrode and the metal layer at the first shot. The fan shape discharge expanded until the current become peak, and then it gradually disappeared. The protective layer at the part of the fan shape discharge and its surrounding was separated from plastic substrate. At the second and the following shots, similar discharge appeared until the metal layer around the high voltage electrode was completely separated. Then, the fan shape discharge appeared at the tip of a surface discharge which bridged between the high voltage electrode and the isolated metal layer. On the other hand, the shock wave was mainly observed around the discharge between the electrode and the metal layer. The spread of the fragments of the protective layer containing the metal layer were also observed and followed the shock wave. From the above, it was revealed that the fan shape discharge and the shock wave affected the metal separation from the optical discs.
There are plenty of factors that influence partial discharge (PD) phenomena of insulating materials in high voltage direct current (HVDC) system, one significant and severe problem is the distorted voltage caused by harmonics. The minor insulation of valve side windings for converter transformer withstands complex electrical stress, which is fundamental AC voltage superimposed with harmonics. The impact of harmonics on insulation systems in terms of electrical losses, lifetime and dielectric material degradation has been previously considered in many researches, a better understanding of the PD characteristics in oil-paper insulation becomes increasingly practical importance.
Partial discharge tests were performed on an experimental platform with needle-plane and sphere-plane electrode, which were designed to simulate the metallic protrusion and surface insulation defect. Statistical phase distributions and typical parameters of PDs were acquired by detector using pulse current method. By means of harmonic order and total harmonic distortion (THD), as well as phase shift angle, the presence of harmonics can be quantified in a simplest case to measure its impact on PD.
The test results showed that various harmonic compositions superimposed on the fundamental sinusoidal waveform have a significant impact on PD intensity. The maximum discharge pulses Qmax was associated with the voltage slope steepness dU/dt and followed Umax, whereas the value of composed testing voltage Umax had a crucial relationship with harmonics content and phase shift. In addition, in the case of phase-resolved measurements, dU/dt influenced the proper interpretation of patterns and derived statistical parameters.
It is concluded that a knowledge of harmonic content is essential and critical in conducting a proper assessment of PD impact. The intention of this paper is to underline the importance of the awareness of harmonic component inside the applied voltage, especially when PD measurements are related to pattern recognition and diagnose criteria.
Oil-immersed pressboard are used in transformer to separate high voltage winding and low voltage winding, as well as providing structural support. Thus during normal operation of a transformer, oil-immersed pressboard will be subjected to the action of mechanical stress and constantly aging. Moreover, when the transformer suffers a sudden short-circuit, huge axial force will act on the pressboard near winding ends, affecting insulation properties of the pressboard, leading to failure of the transformer. Therefore, it is necessary to study the evolution of insulation properties of oil-impregnated paperboard under different mechanical stress. In this paper, normal mechanical stress experiments on oil-immersed pressboard were performed, and partial discharge tests of insulation pressboard after pressed were carried out in order to obtain its partial discharge inception voltage, breakdown voltage and partial discharge characteristics. Material testing machine is used to apply mechanical stress between 0 and 100 MPa to the pressboard and the strain curve is recorded as well. Insulation characteristics of pressed oil-immersed pressboard were obtained by performing both short-term and long-term partial discharge tests under point-plane electrodes. Experimental results indicates that mechanical stress on pressboard will reduce its insulation ability. As the applied mechanical stress increases, the breakdown voltage and withstand voltage time of pressboard will continue to decrease. Partial discharge inception voltage stays substantially when the mechanical stress is relatively low, but decreases rapidly when the mechanical stress is higher than 60MPa. When the mechanical stress is higher, the maximum discharge magnitude and average discharge magnitude will be higher, but the distribution phase of discharges has little to do with the mechanical stress. Confocal microscopy observations on pressed pressboard indicate that damage on fiber inside pressboard caused by mechanical stress may be the cause of reducing insulation ability.
Purpose/Aim
Metal enclosed switchgears, which have been widely used, play an important role in power distribution network. However, its frequent faults are becoming an increasingly serious threaten to the grid security. Prior to faults, partial discharge (PD) would happen due to some insulation defects under electrical, thermal, and mechanical stress with the increase in operation time. Thus, PD detection is an effective way to identify potential faults and inspect insulation defects for high-voltage switchgear. Occurring inside a switchgear, PD would emit electromagnetic (EM) wave that propagates and leaks to the outside, and it also excites the surface current on the metal wall of the switchgear. As a result of current and impedance of the material, it would produce a transient earth voltage (TEV) that could be detected by TEV sensor. Therefore, PD monitoring based on the non-intrusive TEV method has the great engineering application value for estimating insulation condition of switchgear.
Experimental/Modeling methods
A wireless distributed on-line monitoring system based on TEV detection method is proposed in this paper. The overall system includes signal preprocessing unit, data acquisition unit, wireless module and terminal analysis platform. Firstly, depend on the distributed TEV sensors on each switchgear, the partial discharge signals are induced, then through the circuits of filter, amplification and demodulation, the analog signal could be obtained by the acquisition unit of 1Ms/s sampling rate. The wireless module is mainly used for the transmission of not only signal data but also control signal between fore-end data collector and back-end control platform. Finally, the signal data is processed and analyzed by terminal analysis platform.
Results/discussion
The results show that the TEV PD online monitoring system has the characteristics of high reliability, anti-interference ability, less maintenance, etc. In addition, it could assess the PD location and insulation condition effectively to realize fault early warning of switchgears in time, which is of great significance for ensuring the safe operation of switchgears.
The effective identification of ultrasonic direct wave signal is the key to the success of the partial discharge(PD) location of electrical equipment, and it is also a hot spot of research at home and abroad. Based on this, this paper proposes a method based on improved support vector machine (Improved SVM) to identify PD ultrasonic direct wave signal. The parameters of support vector machine are optimized by improved particle swarm optimization (PSO) algorithm, and then the direct wave signal is extracted. Firstly, the time domain waveform of PD ultrasonic direct wave and aliasing wave signals received by the ultrasonic sensor are modeled, and extracting their various characteristics. Secondly, by observing and comparing, the fractal box dimension characteristics of direct wave and aliasing wave signals are selected and the database is established, then use cross validation to select training set and test set. Finally, Simulation Research on the recognition of PD ultrasonic direct wave signal based on ISVM algorithm is carried out, then the direct wave signal is extracted. The simulation results show that compared with the traditional PD ultrasonic direct wave signal extraction method, this method performs fast and the accuracy of classification is high.
There is a growing interest in the development of dielectric materials of high voltage breakdown strength used in compact pulse forming lines (PFLs) to drive high power microwave (HPM) sources and pulsed lasers. Therefore, much attention has been paid to polymer-ceramic composites, where normally the ceramic powders with high dielectric constant are dispersed in a polymer matrix of high breakdown strength (BDS). Therefore, we have tested composite samples made of barium titanate (BaTIO3) as the ceramic powder and epoxy or polymethyl-metacrylate (PMMA) as the matrix polymer. Because of the polymer mixture, we have measured higher BDS compared to the pure ceramics of the order of 2.55 MV/cm for the epoxy-barium titanate composite (with ε= 40) and 420 kV/cm for the PMMA-barium titanate composite (with ε=25). However, as pointed by [1] the use of barium titanate in the PFL dielectric can distort the pulse waveform generated on the load as BaTiO3 is highly nonlinear, i.e. not leading to a rectangular output pulse as desired. To circumvent that this paper addresses this issue by proposing the manufacturing of the polymer-ceramic composite based on PZT (lead-zirconium-titanate) ceramics of weak nonlinearity for use in PFLs. This proposal will be discussed by showing the Weibull plots of the polymer composites and the nonlinear characterization tests of the PZT and BaTiO3 ceramics.
+Work supported by US Air Force Office of Scientific Research under contract no. FA9550-18-1-0111.
[1] L.S.C Bendixsen and P.W. Smith, “Very low impedance forming lines built from ferroelectric tiles,” in Proc. of the 2005 IEEE Pulsed Power Conference, Monterey, Monterey, CA, Jun. 2005, pp. 1333-1336.
Abstract: Charge accumulation on the surface of polymer insulators under DC voltage usually presents two different distribution characteristics, which can be called "dominant pattern" and "charge speckle" pattern respectively. Under "dominant pattern", the charge distribution is uniform and the polarity is the same as the applied voltage; However, the charge density of "charge speckle "is obviously different from the surrounding region, and the charge polarity may be the same as the applied voltage polarity or the opposite. Firstly, the charge accumulation mechanism of the "charge speckle" pattern is different from that of the "dominant pattern", which can’t reflect the essential characteristics of the charge accumulation of insulating materials. Secondly, the "charge speckle" may has a serious effect on the distortion of the surface electric field and reduce flashover voltage along the surface, so it needs to be studied specially. Based on theoretical analysis and experimental methods, the causes and characteristics of "charge speckle" pattern on insulating materials under different conditions are studied, and corresponding measures are put forward to reduce the influence of "charge speckle".
Keywords: DC voltage; polymer; Charge accumulation; Dominant pattern; charge speckle; GIL
References: Zhang B., Gao W., Qi Z., Wang Q., and Zhang G., IEEE Transactions on Instrumentation and Measurement (2017), 66, 3316-3326.
Dayu Li1 , Yicen Hou1 , Boya Zhang1 , Wenqiang Gao1 , Guixin Zhang1
1 Department of Electrical Engineering, Tsinghua University, Beijing, 100084, China Corresponding author's e-mail: guixin@mail.tsinghua.edu.cn
Atmospheric pressure plasma jets (APPJs) are suitable for uneven surface and larger-scale treatment, which have been a research hotspot in recent years,ex.surface modification for high voltage insulation. But the interaction between individual jets is the main concern impeding their wide application. In this paper, the temporal-spatial distribution of reactive species in one-dimensional plasma jet array driven by nanosecond pulse power were studied. The effect of input gas and jet spacing on jet interaction were evaluated by electrical, optical and fluidal measurements. Results showed that: (1) The repulsive force between He plasma jet was stronger than Ar plasma jet; (2) The interaction force become weaker with longer jet distance, in our case, with distance of 20 mm, no obvious jet interaction was observed; (3) Although center jet travelled into a shorter distance compared to jets outsides, the emission intensity of N2 second positive band and N2+ first negative band showed a slower dissipation rate than that of outside jets; (4) N2+ first negative band was very sensitive to jet distance, which showed a faster dissipation rate with shorter jet distance.
Aiming at the μs-class response speed and the requirement of pulsed high-current test of the ampere-level current amplitude, a mathematical model of self-integrating Rogowski coil was established. The amplitude-frequency characteristic of self-integrating Rogowski coil was analyzed. The influence of magnetic core material, the structural parameters and electromagnetic parameters can be researched. The Rogowski coil was made which can measure the μs-class response speed and the pulsed high-current test of the ampere-level current amplitude. According to the performance evaluation of Rogowski coil, a Rogowski coil performance parameter test platform was set up, and the Rogowski coil performance parameters were checked. Finally, the calibration factor of Rogowski coil was 499.4A / V, which can accurately measure the rise time greater than 2.6μs, Current amplitude up to kA pulse high current.
Kraft paper, as a main kind of insulation statistic in power transformer, is enduring various effects form outside all the time. These effects contain heat, mechanical stress, water and etc. The main influence factor has been pointed to be heat. The cellulose chain in paper degrades under the effect of heat, which turns to the result that Kraft paper become crisper. At this condition, if paper get stress or friction, it can fracture easily. And that announces the end of the power transformer’s life. To analyze this condition, paper simulates the mechanical stress on different thermal aging paper. The experiment based on pulsed current method used needle-plate structure at partial discharge test. Under step-up applied AC voltage, developing processes of partial discharge and the characteristic of discharge were found from experiment. The results show that the mechanical stress has different influence on the different thermal aging pressboard. Especially for the middle thermal aging paper. The results can provide some help to further study on combined effect of mechanical and thermal stress for oil-immersed Kraft paper.
While gas breakdown is generally driven by Townsend avalanche, field emission dominates for microscale gaps [1]. Recently derived closed form solutions unify these mechanisms with good agreement to experiment while analytically demonstrating the transition between them [2]. One weakness of these approaches is that they fit the field enhancement factor empirically, preventing a priori predictions of field emission driven microscale breakdown. This becomes particularly critical since the field enhancement factor may vary between samples due to variations in surface roughness.
Previous experiments measured the breakdown voltage of a flat plate electrode and a sharp-tipped electrode for microscale gaps at atmospheric pressure [3]. We extend this study by modifying the surface roughness of the flat electrode and measuring the breakdown voltage and current as a function of gap distance and time. Preliminary results suggest that differences in surface roughness may lead to 50% changes in breakdown voltage for 5 um gaps. Furthermore, we will present a new geometry to assess the impact of tip aspect ratio, gap distance, gas, and pressure will permit a full parametric study to benchmark to theory. Implications on gas breakdown for microscale and smaller gaps and the conceptual understanding of breakdown regimes will be discussed.
[1] D. B. Go and A. Venkattraman, “Microscale gas breakdown: ion-enhanced field emission and the modified Paschen’s curve,” J. Phys. D: Appl. Phys., vol. 47, art. no. 503001, 2014.
[2] A. M. Loveless and A. L. Garner, “A universal theory for gas breakdown from microscale to the classical Paschen law,” Phys. Plasmas, vol. 24, art. no. 113522, 2017.
[3] M. A. Bilici, J. R. Haase, C. R. Boyle, D. B. Go, and R. M. Sankaran, “The smooth transition from field emission to a self-sustained plasma in microscale electrode gaps at atmospheric pressure,” J. Appl. Phys. vol. 119, art. no. 223301, 2016.
The spark paths of SF6 are tortuous in non-uniform electric field, for the randomness of gas discharge and the influence by the space charge. Meanwhile, the breakdown voltage appears non-monotonous trend with the variation of pressure. This phenomenon is called as corona stabilization because it is caused by the space charge generated by corona discharge. This paper focuses on the relationship between the corona stabilization and the tortuous spark paths. In this paper, we built a spark path observation system by the three-dimensional reconstruction method. The deviation angle of spark paths was used to describe the tortuous degree. The breakdown characteristic under positive DC voltage was measured. The result shows the tortuous degree of the spark paths appears in different breakdown region. In the streamer region, the deviation angle of spark paths is closed to zero, that is, the spark paths are almost perpendicular to the plane electrode. Once the leader discharge appears, the deviation angle increases suddenly. And the deviation angle decreases with the pressure until the breakdown enters the corona-free region. In the corona-free region, the deviation angle depends on the surface topography of the electrode.
Purpose/Aim
Partial discharge (PD) is the main reason for the insulation deterioration inside the electrical equipment. It’s necessary and valuable to locate the PD source occurred in electric power system timely and accurately. At present, ultrasonic array location method is an effective way in partial discharge detection. With high signal gain, strong anti-interference ability as well as high resolution, this method has been widely used to locate the partial discharge source. As we all know, the partial discharge ultrasonic array sensor with good acoustic performances is the basis of PD detection and location. And indeed, there is still much room for improvement in array arrangement using the ultrasonic array location method, which needs further exploration.
Experimental/Modeling methods
In this paper, a partial discharge location technique based on double helix ultrasonic sensor array is proposed, and the accuracy of this method has been verified by simulation and lab experiment. Firstly, the static beam-pattern was analyzed for several kinds of arrays to get the acoustic performances, which explains the great acoustic characteristics of double helical array in theory contrasted with other circular array; Secondly, this paper uses MATLAB to establish an acoustic finite element model, which employs the double helix sensor array to simulate the acoustic characteristics of a single narrowband ultrasonic source. Next, this paper uses MUSIC algorithm to draw the corresponding spatial spectrum and locate the PD source. Finally, the lab experiment has been executed to verify the simulation in a similar process.
Conclusions
The results show that the uncertainty of ultrasonic positioning based on double spiral ultrasonic sensor array is less than 7% both in simulation and lab experiment. It means that this array can successfully locate the PD source, and can also meet the requirements of engineering.
Wide-bandgap GaN optically-controlled switches have the potential for driving down the cost and size, and improving the efficiency and capabilities of high voltage pulsed-power applications. The key scientific challenge will be measuring the high-field photo-conductive (PC) properties of these materials to determine if they will operate in a high-gain and/or sub-bandgap triggering mode, like what has been observed in GaAs.
600 µm gap devices from GaN wafers from Kyma and Ammono were fabricated for laser testing. The laser used was a Nd:YAG Q-switched system doubled to produce emission at 532nm to 12mJ energy. With relatively low voltages applied the switch, linear photoconductive currents were measured through the pulse in response to exposure to the laser pulses. As increasing voltage is applied to the PCSS, above a threshold persistent conductivity is measured that lasts well beyond the duration of the laser pulse and discharges the charging transmission line in the system. This occurs at threshold fields of 10-15 kV/cm.
This persistent photoconductivity is a distinguishing characteristic of what is known as “lock-on” effect in photoconductive switches, most notably known in GaAs-based devices. Once lock-on is initiated from laser generated electron-hole pairs, as long as sufficient field remains applied to the switch, avalanche carrier generation is sustained to maintain conduction even after the laser illumination is removed. This effect is the basis for highly-efficient photoconductive switching requiring relatively little laser energy. In other experiments with the GaN devices, high-gain switching has been initiated with as little 35µJ laser energy thus far.
Another distinguishing characteristic of lock-on switching effect known from characterization in GaAs is the formation of filamentary current channels. These filaments can be imaged due to the emission of recombination radiation from electrons and holes in the plasma within the filaments, and have been recorded in the GaN switches.
Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia LLC, a wholly owned subsidiary of Honeywell International Inc. for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE‐NA0003525.
When driving dynamic loads such as the pulsed electron beam device (GESA), an adjustable output waveform of the driving pulsed power source is desirable. To meet the demand of a maximum output voltage of 120kV at pulse currents of up to 600A for pulse length of up to 100µs, a new Marx-type power modulator is currently under development at IHM. The modular design combined with a gate-boosting circuit enables the use of cost-efficient off-the-shelf IGBT devices as switching elements. Together with a low inductance layout of the capacitor leads and the arrangement of stages, a fast voltage rise time below 100ns across the load is ensured. By implementing an optical bus, the control signal distribution is simplified. Fault detection and clearance as well as EMI robustness is achieved by implementing a control unit on every stage. Individual switching sequences for each stage allow for step-wise arbitrary output waveforms. This contribution presents selected aspects of the design process and their validation in a generator arrangement driving dynamic loads.
The thyratron has been used as a switch in pulsed-power applications for almost a century. In the last 20 years, as a result of developments pioneered at DTI, most new modulator applications have transitioned away from thyratrons to solid-state switching. As the cost and capabilities of solid-state modulators has improved, their adoption has grown rapidly. With the continued evolution and reliability of solid-state pulsed-power systems, virtually all new pulsed-power systems are designed around solid-state capabilities.
Thyratrons have a lifetime of only ten to twenty thousand hours and require periodic adjustment of their reservoir heater voltage. The solid-switch’s extremely long life and reliability offers an attractive replacement opportunity for many thyratron-based systems, with orders of magnitude longer lifetime and no regular maintenance. Replacing thyratrons with solid-state switches that last 20 years or more without maintenance would enable significant savings over an accelerator’s life. Until recently, however, solid-state switches have not historically been capable of handling the voltage, current, and risetime required to replace thyratrons.
Under a recently completed DOE Phase II SBIR, DTI has developed and demonstrated a thyratron replacement switch for the SLAC National Accelerator Laboratory. Unlike the series string of insulated-gate bipolar transistors (IGBTs) used in most of DTI’s modulators, this switch was designed using arrays of series- and parallel-connected commercial IGBTs. It has successfully demonstrated full operating capability at the SLAC thyratron specifications of 48 kV, 6.3 kA, and 1 µs risetime. In addition, the switch has the potential to improve accelerator performance by reducing peak-to-peak pulse jitter to a level five times shorter than is typical for thyratrons. This demonstrated jitter of 1.5 ns has the potential to significantly improve the performance of the Linac Coherent Light Source (LCLS) beam, and increase the HV stability in the accelerator.
Solid state klystron modulators are typically based on oil-immersed pulse transformers due to their high performance, robustness, simplicity and straightforward design. However, the size of such transformers are highly impacted by pulse power, output voltage, pulse length, and required rise time; key parameters which are difficult to combine in long pulse high power linac applications.
In this paper, pulse transformer design models for two winding configurations (single layer winding and pancake winding), including calculation of parasitic elements, are developed and validated in a 3D finite element analysis environment. These models are then employed in a global optimal design procedure used to study the evolution of pulse transformer footprint, magnetic core volume, oil tank volume, and efficiency as pulse length is increased from 100 µs to 4 ms while constraining maximum pulse rise time and overshoot. The impact of required pulse power and output voltage is also studied.
The single layer winding based on standard enameled round wire is first investigated without externally imposed restrictions on size to study the limitations of this first winding technique. This is followed by a study of the same configuration but with constraints on transformer size imposed mainly by manufacturability and maintainability, in which it is seen that sub-optimal rise time and therefore transformer size is attained for long pulse high power applications. Consequently, the pancake winding configuration is evaluated under the same conditions, demonstrating that, although more complex and costly, its flexibility allows a more compact design.
Finally, a pulse transformer rated for pulse amplitude 115 kV, output current 25 A, pulse length 2.8 ms, and 0-99% rise time <300 µs is designed, demonstrating the design procedure and showcasing limitations experienced in design. Its performance is assessed in circuit simulation whereas the validity of the derived parameters is demonstrated through finite element analysis.
The CLIC study is investigating the technical feasibility of an electron-positron collider with high luminosity and a nominal centre-of-mass energy of 3 TeV. Pre-damping rings and damping rings (DRs) will produce ultra-low emittance beam with high bunch charge. The DR kicker systems must provide extremely stable field pulses to avoid beam emittance increase. The DR extraction kicker system consists of a stripline kicker and two pulse modulators. Specifications for the electromagnetic field pulses require that the modulator generates pulses of 900 ns flattop duration, +/-12.5 kV and 305 A, with ripple and droop of not more than 200 ppm (+/-2.5 V) with respect to a reference waveform. Inductive adder topology has been chosen for the pulse modulators: the output waveform can be adjusted by applying analogue modulation methods. Two full-scale, 12.5 kV, 20-layer, prototype inductive adder have been designed, built and tested at CERN. One of these has also been tested with eight additional layers, to facilitate two operation modes: 12.5 kV pulses for extraction kicker operation and 17.5 kV for dump kicker operation. An automated control system for droop and ripple compensation, based on Labview software, has been designed and implemented for the prototype modulators. The prototypes are planned to be tested with a prototype stripline kicker, installed in a beamline, at ALBA Synchrotron Light Source in Spain. The results of laboratory tests, comparisons of results from different measurement techniques and an analysis of the limits of these techniques are presented.
Four identical pulse generators (pulsers) provide fast beam deflection in the Low Energy Beam Transport (LEBT) beam chopping system at the Spallation Neutron Source (SNS) Linac. The existing pulsers have become obsolete and need to be replaced to improve overall performance of the chopper system. A new bipolar pulse generator has been designed, and a prototype has been built and tested at the SNS. The generator uses fast push-pull transistor BEHLKE switch modules. It can produce a 1MHz burst of alternate ± 2.5 kV with adjustable pulse widths during the 1-ms macro-pulse at a burst repetition frequency of 60 Hz with rise/fall times of less than 40 ns into a 100 pF capacitive load. The performance enhancements of this new pulser in comparison with existing and previous versions of the pulse generator are discussed in the paper. The pulser can be used in applications requiring a series of bipolar ± 3.5 kV pulses, or unipolar pulses of positive or negative polarity with amplitude up to 5 kV and various pulse widths with a minimum of 150 ns. The maximum switching frequency, number of pulses, or burst repetition frequency in burst mode is limited by maximum power dissipation in the switches and other components. The features of this new pulser along with the results of timing control adjustments, timing and temperature measurements, and HV full power testing on the test stand are presented.
Wide bandgap (WBG) semiconductor devices bring numerous advantages to power converters due to their inherent superior characteristics and high performance under harsh operating conditions. Currently, traditional silicon switching devices are approaching their practical limits in terms of fast switching, high voltage, and high temperature operation. To overcome the limitations of the existing Si technologies, silicon carbide (SiC) power devices are chosen because they exhibit a higher energy bandgap with a lower on-state resistance and therefore are effectively able to improve the power density and increase efficiency of power conversion for high-temperature and high-switching environments. This paper presents an efficient transformer-less high-gain floating boost converter quipped with SiC cascode JFET for high-voltage applications. The switching behaviors of both Si and SiC cascode JEFT devices are evaluated using a double-pulse test (DPT) technique at different gate resistances, device currents, and junction temperatures. In addition, the efficiency of the converter design with SiC power devices is investigated and compared with the Si-based converter under different switching frequencies and load conditions. This work uniquely presents a characterization of the new generation SiC cascode JFET manufactured by USCi and then demonstrates the benefit of using SiC JEFETs in the high-gain DC-DC converter. The results show SiC power devices utilized in the converter can reduce the total semiconductor loss and improve overall converter efficiency.
We describe a new class of electric motor/generator technology with specific power in excess of 10kW/kg for a variety of applications including terrestrial and airborne hybrid and all electric propulsion.
The motor concept takes advantage of the fortuitous match between the recently developed low voltage, high current, small footprint MOSFET technology and the elemental (single turn) voltage and current required at the motor’s gap.
The motor topology uses arrays of inverter Controlled Turn-less electro-mechanical Structures (CTS) where individual inverters control individual turn-less 3-phase motor elements (TLS) where the TLS is made of 3 single conductor shorted at the opposite end from the inverter and functions as a 3-phase motor winding. Each CTS measures around 6mm in the direction of motion and corresponds to a N-S pole pair. Since specific power is inversely proportional to the pole size, it is here enhanced accordingly.
This motor/generator concept has been configured for a variety of applications including terrestrial and airborne propulsion and generation, and various actuators for flight control and robotics where long strands of TLSs can mimic muscle fiber with much higher force density obviating the need for gear reduction. Furthermore, the integration of a new class of Li-Ion high specific power batteries with each CTS provides the double function of both the inverter capacitor and energy reserve providing an enhanced overall performance.
We will describe the underlying technology and discuss its implementation in various applications with emphasis on airborne propulsion.
Many cathode-driven beam injection systems require a low average power highly stable and flexible modulator to generate an electron beam. The stability of the initial lower-power beam is vital to overall system efficiency. Stangenes Industries has developed a modulator system that is comprised of an 18-stage Marx modulator driving the primary of a pulse transformer within the confines of a 19” 5U rack with a 20” depth. The system operates with single-phase 115/230VAC input with <15A of input current. The modulator is a 50KW peak power 150W average load power 18-stage solid-state Marx modulator with pulse-to-pulse output variability of 3-40KV with <100V of output resolution. The pulse width is adjustable up to 5us and rep-rate is 500pps+ for driving loads up to 300pF with an isolated filament via a bifilar winding in the transformer. The system operates over the entire voltage range with a fixed charge voltage.
The system controls are FPGA and MPU based and are Ethernet compatible with live diagnostic monitoring and control. Low-latency hardwired optical isolated and differential signaling receives off-system interlocks and an N-bit signal to choose between several operational states with preset output voltage, rep-rate and delay. These operation states can be adjusted inter-pulse at rep-rates up to 500pps. The trigger is also received via differential signaling with a propagation delay between signal reception and 10% pulse rise of <1us. The system incorporates a pulse transformer with a bifilar winding to provide 40W of DC filament power. All connections are made in the back of the unit with no water-cooling requirements.
Due to the inefficiency, high price, and environmental unsafety of the conventional recovery methods, the production decrease of the oil well is an increasingly serious problem all over the world. Therefore, the need for enhanced oil recovery (EOR) techniques to recover a higher proportion of the oil has become imperative. The electrohydraulic shock waves with steep fronts and high intensity are utilized to generate micro-cracks near the oil well and enhance the permeability of the reservoir. A novel enhanced oil recovery technology based on repetitive electrohydraulic shock waves is presented in this paper. An observable enclosure with the ability of withstanding 50MPa hydrostatic pressure is developed to study the influence of hydrostatic pressure on the mechanism, controlling, and application of the repetitive electrohydraulic shock waves. The paper presents experimental results and optical observation for the characteristics and the development of the streamers under high hydrostatic pressure. The phenomena of the discharge as well as controlling of the electrohydraulic shock waves as a function of hydrostatic pressure are presented. In order to further evaluate the effect of fracturing the rocks and then enhancing the permeability, experiments are carried out on hollow cylinder specimens under different hydrostatic pressures. X-ray scans and microtomography have been used to analyze the dynamic response characteristic and the evolution of the micro-cracks of the specimens. These micro-cracks contributed to a great reduction of the difficulty and energy consumption in the oil recovery process. Results points out that design of prototype is dependent from the hydrostatic pressure and confirms the potential of this technology to be a promising EOR method in low permeability reservoir.
Affordable access to deep geothermal heat, available anywhere on Earth but at costs depending on geological conditions near the surface and equipment access, is the most promising solution for humanity's need to deeply, then completely, decarbonize its total energy supply, as quickly as we prudently and profitably can. Nascent Electro Pulse Boring (EPB)technology, proven in concept in test boring to 200 m in Europe, is very promising for low-cost access to deep (6 - 10 km) geothermal, but a major R&D & Demonstration program will be needed for evolutionary equipment design based on progressively-deeper field boring tests. This project will attempt to organize and fund that program. EPB is a rock breaking technology for both sedimentary and crystalline bedrock, compatible with conventional pumped mud cuttings removal, with potential for high rate of penetration (ROP) with relatively light, transportable, low-power-consumption equipment. It may be capable of directional drilling, to replace deep rock fracking and the high operating pumping costs of conventional EGS with deep hole branching and multiple thermosiphons, delivering water to the surface that is hot and copious enough for both electricity generation, by steam or organic Rankine cycle (ORC), and district heating and cooling systems (DHCS).
The critical EPB component is the Down Hole Pulse Generator (DHPG), which must operate at full-depth (~ 300 C) while providing ~ 10-20 pps, ~ 10 ns, at 1-10 KJ, 500-1000 KV, with DHPG lifetime of 10^5 - 10^7 pulses before refurbishing. When design and construction of the DHPG is funded and completed, progressively-deeper boreholes may be constructed at a test site like USDOE "FORGE".
This may be an ideal commercial application of DEW technology developed by several defense contractors. EPB commercialization may be the "transformative" and "disruptive" technology needed for benign, inexhaustible, baseload electricity and thermal energy, nearly anywhere on Earth.
To influence the weather is a long time dream of mankind. Current cloud seeding technology mostly relies on chemicals such as silver iodide and hygroscopic salts, which may have negative impacts on environment and human health. Here we present a new method for triggering macroscopic water precipitation in air. Experiments demonstrated that high-voltage generated corona discharge was able to induce rain and snow formation in atmospheric pressure air. The effect was confirmed by precipitation experiments performed in a 1 m3 and a 15,000 m3 cloud chamber, where the absolute water content was close to the actual values in natural cumulonimbus clouds. With the presence of electrical charges, the collision efficiency between the water droplets was increased by over one order of magnitude, accelerating the coalescence process and possibly leading to the rain and snow formation otherwise impossible to form. The electric-based technology presented in this paper provides an effective, inexpensive, and environmental friendly tool for triggering water precipitation, and may open up new opportunities for producing water precipitation in large open space.
We have studied the water purification by pulsed discharge in atmospheric gas including water droplets. The water purification is caused due to decomposition of organic compounds in water by OH radical generated in the pulsed discharge space.
Although there are many reaction path from plasma generation to OH radical reaction with an organic compound, it is not known which path is dominant. To aim to know it, OH radical was measured, setting density of ozone and plasma as parameter. The pulsed discharge plasma was generated in mix gas of argon and oxygen. Ozone was produced by the pulsed discharge in the reactor and supplied from ozonizer. To control ozone density in the reactor, the ratio of argon and oxygen in the reactor and the amount of supplied ozone were adjusted, with feeding back measured ozone concentration. To control plasma density, peak, width, and repletion ratio of pulsed voltage applied at electrode in the reactor were varied using a solid state compact linear transformer driver (LTD). OH radical was measured by fluorescence method using the disodium salt of terephthalic acid.
The result showed that the amount of OH radical depended on ozone density. Therefore, the reaction path including ozone is one of dominant path which occurs from plasma generation to OH radical reaction with an organic compound. Additionally, synergistic effect by plasma is also discussed in our presentation.
Diamond Cross Ranch
Weihua Jiang, Akira Tokuchi, Taichi Sugai, Keita Inagawa, Takayuki Hangai, and Genta Sagisaka
Extreme Energy-Density Research Institute, Nagaoka University of Technology, Japan
Solid-state linear transformer driver (LTD) technology is being studied at Nagaoka University of Technology. It is a different approach to compact pulsed-power source development from that based on traditional magnetic pulse compression. Pulsed power generators based on the LTD scheme are characterized by modular structure and inductive output adding. These characteristics have brought higher flexibility and versatility to pulsed power generators, allowing more functional and more efficient applications in many industrial areas.
Solid-state LTDs are in principle similar to the large, spark-gap switched LTDs, although they have been developed for very different purposes with totally separated operation parameters. Solid-state LTDs use power semiconductor devices as switches and the switching-off capability and the highly repetitive potential of the semiconductor devices have very much distinguished solid-state LTDs from the large ones.
It has been demonstrated that the LTD output adding can be utilized as a means of output waveform variation, by adequately controlling the relative timing of different modules. With all module switches controlled by binary signals generated by a logical circuit board, the solid-state LTD system is controlled electronically, allowing us to explore the possibility of "smart pulsed power" sources.
Solid-state LTDs are developed for industrial applications. Development projects are underway as our collaboration with different industrial partners. The advancement in pulsed power sources is being turned into that in performance and productivity.
DTI successfully tested the first of three klystron modulators for the CLARA at Daresbury Laboratories, UK, on February 20, 2018. This high power modulator is designed to pulse 70 MW-class klystrons at an average beam power of 250 kW. Each system provides 450 kV, 545 A cathode pulses, with a 3.0 µs flattop better than ±0.02%. Installation of the first modulator in with its klystron is scheduled for the spring of 2018, with delivery of the next two modulators in the summer of 2018.
The customer-supplied klystron is mounted directly on the oil-filled modulator tank, which contains the pulse circuitry and a passive pulse corrector with automated adjustment. This unique passive circuitry delivers the extremely flat output pulse required for advanced accelerator applications. The modulator tank lid is split into two separate sections for ease of service. The primary section of the tank houses the solid-state switches and main storage capacitor. The secondary section of the tank houses the pulse transformer and tube socket. The modulator is energized by a standard DTI switching power supply, providing 40 kV, 250 kW average power.
The CLARA systems are hybrid modulators, with the 40 kV solid state switch driving the primary of a 12:1 pulse transformer at over 6 kA peak. This design provides high efficiency, simple operation, and substantial margins with respect to output voltage, current, pulse width, PRF, and average power for high reliability. A secondary function of the solid state switch is circuit protection, eliminating the need for a crowbar. When an arc occurs, the fault is sensed and the switch will open in less than 1 µs to disconnect high voltage from the klystron cathode. The current rate of rise is limited to a safe value by the inductance of the pulse transformer. System controls and low power support circuitry are mounted in a standard rack.
For laboratory-scale experiments on the treatment of biological material by pulsed electric fields an eight-stage Marx-type pulse generator has been developed and set up. Its stages comprise IGBT switches in full-bridge configuration for generating bipolar pulses. The generator is grounded at its center allowing for a ground-symmetric operation of a flow chamber for material treatment. The generator has been designed for a pulse current of 600 A, and a maximum stage voltage of 1 kV enabling the use of inexpensive off-the-shelf components. It delivers rectangular pulses of adjustable polarity and length in a range between 1 µs and 10 µs at a pulse repetition rate up to 200 Hz. Each stage combines two modules each equipped with two IGBT switches in half-bridge configuration, the related control circuitry, one charging switch, and the stage capacitor. A circuit design focusing on low inductance enables a full-load current rise of a stage within approximately 100 ns. The contribution describes selected design aspects of the generator and presents measurement results obtained during commissioning.
Nanocrystalline material has become one of the most efficient and prolific materials for use in high-frequency pulse transformers. Nanocrystalline cores successfully bridge the gap between laminated steel and ferrite with high saturation and low loss, however, the cost and quality of the material is variable. The manufacturing and quality control procedures of the material has a direct impact on the core loss. MK Magnetics is constantly developing improved core manufacturing procedures with nanocrystalline material and a variety of other standard materials. At Stangenes Industries an experiment was devised to test core loss in two of the industry-available nanocrystalline materials used at MK Magnetics. Two cores of similar size were manufactured with an identical annealing process, only differing the core material. The cores were then tested at range of frequencies to compare core loss. This data is used used to confirm quality as well as experimentally verify a commonly used equation for core loss in nanocrystalline cores. An expanded variety of core materials could be subjected to the same test procedure to both verify the quality and serve to update core loss charts constructed around less accurate test methods.
Numerical methods are frequently employed for high voltage designs. The results usually determine the size of parts with safety factors. Charge transport based numerical codes are sometimes implemented to better understand the high voltage design related to the breakdown performance, however such approaches require computationally expensive compared to straight forward electric field computations. Notice that any numerical model should be verified (validated) with measurements with measurements to establish the model and release it for other to use in industrial settings. In some cases validation tests could be expensive due to time and man power needed for manufacturing a part and test set-up. A decrease the number of sample variations would be appreciated for quick success. In this contribution we will present a numerical model for determining how high voltage parts can be designed for best performance in air for reliable performance. The numerical method is based on the finite element method and utilizes a straight forward electric field solver. The results from the simulations are compared to actual measurements to determine whether a rule for air breakdown could be established just by considering numerical simulations as an early step to modify high voltage designs.
According to the preliminary study, the Laser Triggered Vacuum Switch (LTVS) might be an advanced high-voltage, high-intensity pulse switch. It will have the advantages of short trigger delay time and good isolation from the main circuit. The study of the trigger mechanism of LTVS is conducive to both the optimization and the miniaturization of it, and beneficial to promote its practical process. In this paper, the LTVS which has a composite target electrode made of potassium chloride and titanium is studied. The energy loss of the target before and after the trigger is compared by X-ray spectroscopy. And based on the result, the composition of the plasma mass generated by LTVS is analyzed. The trigger mechanism of LTVS is proposed for laser ablation according to the trigger physical phenomenon and the analysis above. Meanwhile, the interaction between laser and the target is divided into three stages: the heat transfer stage, target reaction stage and plasma formation stage. In the experiment, both the electron microscope images of target ablation and the switch delay at different wavelengths and different trigger energy are studied. The results show that the hypothesis about trigger mechanism is in good agreement with the experimental data. The research about the trigger mechanism of the LTVS is of great guiding significance for its optimization and application.
Rectangular high-voltage narrow pulses are preferred in most Dielectric barrier discharge (DBD) loads and corona discharges in water. Usually at least two groups of switches are required to generate rectangular pulses over capacitive loads. One group discharges the energy-storage capacitors to the load and the other group discharges the energy stored in the stray capacitor of the capacitive load so as to chop off the slow tailing of pulses. Therefore, two independent drive circuits are required for these pulse generators. In this paper, a novel drive circuit combined with time-delay circuits, which can drive many different groups of switches with only one drive circuit, is proposed. Using magnetic transformers with common primary winding, many synchronized driving signals can be generated. With some special time-delay circuits, these synchronized driving signals are modulated into many groups of different driving signals. Experiments prove that rectangular and stepped pulses can be generated using this single drive circuit.
The AT-HVCM, achieved by series-stacking the dual-function resonant filter capacitors following each rectification stage, is the baseline design for the PPU project at the SNS. The project requires three AT-HVCM modulators to power 28 additional 700kW klystrons and associated high beta cavities in order to accelerate 38mA of beam current to 1.3GeV. Each modulator produces an 82 kV, 120A output pulse of 1.35ms duration at a 60Hz repletion rate. The AT design approach improves upon the existing HVCM topology, reducing stresses on the H bridge IGBTs using zero voltage (ZVS) and current switching (ZCS) techniques and on the resonant filter capacitors by limiting voltage reversal. This paper reviews the prototype development progress, currently focused on evaluating system reliability, and discusses some design techniques to achieve the modulator performance requirements.
A magnetic field diagnostic system has been developed that enables measuring the transient magnetic field diffusing through conductors with wall thickness on the order of the skin depth. Typical commercial magnetic field probes are limited in their ability such that they are only able to measure in a single direction at a time, and they may require multiple measurements in different orientations to account for the external common mode noise.
The magnetic field testbed generates a sinusoidal current with a peak value of ~38 kA, with a ringing frequency of ~5.5 kHz, through a two turn excitation coil. The field diffuses through the walls of a hollow conductive structure placed roughly 5 cm away from the center of the coil. The thickness of side and bottom walls is kept constant, while the top plate thickness is varied to alter the rate at which the magnetic field diffuses through the conductor. Field amplitudes inside the shielded volume are attenuated by multiple orders of magnitude compared to the external field, which complicates measurements. That is, the noise introduced through magnetic interference may easily be on the order of the measured signal unless a proper diagnostic design is chosen.
A 3D differential B-dot probe with sub-µs rise time was successfully developed. The probe consists of three differential pairs of multi-turn pickup coils, which allows accurate capture of tri-axial magnetic fields. Previously developed probes constructed using single-turn differential coils had a lower signal-to-noise ratio and were more susceptible to stray magnetic fields in comparison to the newly constructed probes. The significant common mode noise of the commercially available probe has been virtually eliminated through employing a differential measurement technique. With the existing noise level a minimum dB/dt of 25 times less than the external field of 950 T/s is measured with high fidelity.
Two-dimensional as well as 3-D electromagnetic calculations of transient magnetic field diffusion are compared to experimental results providing unique insight into the impact of pertinent parameters on the diffusion process. For simulation, the finite element method (FEM) is employed while a pulsed high magnetic field testbed operating at a peak current of 40 kA is utilized to generate experimental magnetic field diffusion data. The complexity of the diffusion of the transient magnetic field through the walls of conductive shielding structures is elucidated in detail. Most relevant is the diffusion condition where conductor thickness is on the order of one skin depth, a situation typical to many pulsed power systems.
Any conductive structure subject to a time-varying magnetic field gives rise to an eddy current distribution in the conductor that opposes the incident field. This current distribution decays with a characteristic time constant, similar to that of a loop with inductance L and resistance R, resulting in a finite delay and attenuation of the diffused field. For practical geometries subject to highly divergent external magnetic fields, these complex induced current distributions are at best difficult to predict analytically and shielding performance even more so.
Clearly, the magnitude and direction of the diffused field are highly dependent on a variety of geometrical and material parameters, including wall thickness, conductivity, permeability, and excitation frequency. Further, conductive loading of a shielding structure’s internal volume will result in significant degradation of the shielding performance due to compression of the internal field. For instance, simulation of a cylindrical shielding geometry with a wall thickness of approximately one skin depth yielded a factor of five increase in the diffused field magnitude when the internal non-conductive volume is reduced by ninety percent. The introduction of imperfections, such as slots and holes, within the structure walls, is also considered.
In order to study the nitrogen-oxygen reaction caused by intense pulse arc, in this paper, the average temperature and the radius of arc channel are calculated by the method of subsection exponential fitting in Elenbaas-Heller arc model, which is related to the arc current. The calculated result shows that when the arc current exceeds 30 kA, the arc channel average temperature is saturated near 40000 K. The calculated arc radius is related to the peak and gradient of the current, as the arc radius increases, the volume of the arc channel expands, which will heat more reaction gas. The variable of peak current range from 150 kA to 300 kA and the other variable of pulse width between 1 ms and 3 ms are investigated by two experiments respectively, and the molar quantity of nitride is measured in the 4 L seal cavity with 2.3 atm dry air by flue gas analyzer. The experiment results show that when the current pulse width keeps 700 μs and the discharge peak current exceeds 240 kA, the concentration of nitrogen dioxide contents stable near 565 ppm and the concentration of nitric oxide increases progressively. According to the nitrogen-oxygen reaction principle, under the experiment parameter, the amount of nitride increases with the augment of the calculated arc radius.
The implementation and demonstration of laser-collision induced fluorescence (LCIF) generated in atmospheric pressure helium environments is presented in this communication. As collision times are observed to be fast (~ 10 ns), ultrashort pulse laser excitation (< 100 fs) of the 2(3)S to 3(3)P (388.9 nm) is utilized to initiate the LCIF process. Both neutral induced and electron induced components of the LCIF are observed in helium afterglow plasma as the reduced electric field (E/N) is tuned from < 0.1 Td to over 5 Td. Under the discharge conditions presented in this study (640 Torr He), the lower limit of electron density detection is ~ 10^12 e/cm^3. Spatial profiles of the 2(3)S helium metastable and electrons are presented as functions of E/N to demonstrate the spatial resolving capabilities of the LCIF method.
The shock waves generated by the underwater electrical wire explosion continuously attract attentions due to its more and more applications. In this paper, the generation mechanism of shock waves generated by the underwater electrical wire explosion was investigated. A microsecond time-scale pulsed current source is used to drive the electrical explosion of copper wires with a length of 5cm and diameters of 200μm. The energy-storage capacitor was charged to a relatively lower energy so that the energy deposited into the wire is not large enough to fully vaporize the whole wire. The discharge is in the mode of the cut-off current without the plasma formation. The discharge current and the wire voltage were measured with a Rogowski coil of Pearson 101 and a voltage divider of Tektronix P6015A, respectively. A pressure probe of PCB138 from Piezotronics was placed at a position of 100mm away from the wire and used to record the waveform of the shock wave. Two shock waves were recorded: the first and weak shock wave was confirmed to be the contribution from the wire melting, while the second and strong shock wave is the contribution from the wire vaporization. By adjusting the initial stored energy appropriately, ranging from 70J to 160J, the time interval of shock waves generated by melting and generated by vaporization can be changed conveniently. Therefore, the process of the shock wave by the vaporization overtaking the shock wave by the melting can be observed.
Hydrophobicity is the key factor for silicone rubbers to fabricate insulators under polluted and wet conditions. The influence of contamination on the hydrophobic properties of plasma-treated silicone rubbers (SR) was investigated in this paper. The hydrophobicity properties of silicone rubber were determined by the static contact angles between the droplet of distilled water and the horizontal surface of the silicone rubbers. Atmospheric pressure plasma jet was applied to treat silicone rubbers with three surface conditions: clean surface, surface polluted by solid layer method and polluted surface using plasma jet treatment as pretreatment.
The results showed that clean SR lost its hydrophobicity rapidly under plasma jet and subsequent hydrophobicity recovery was slower. In contrast, polluted SR surface changed from hydrophobic state to hydrophilic state with contact angle of 120° after 1 min treatment. The polluted surface could keep its hydrophobicity after 200 hours. The silicone rubber under the third condition also became hydrophobic after plasma jet treatment but the process was slower than polluted surface without pretreatment.
The surface was examined under scanning electron microscopy (SEM), X-ray photoelectron spectroscopy (XPS), energy dispersive X-ray (EDS) spectroscopy and fourier transform infrared (FTIR) spectroscopy. SEM and XPS showed clean SR surface became rough while organic surface was destroyed gradually and inorganic layer appeared with plasma jet treatment. EDS and FTIR showed that organic groups appeared on the polluted SR surface and made it hydrophobic.
On your own
TAE Technologies field-reversed configuration (FRC) type, C2-W (Norman), fusion experiment is operating since April 2017. The Active Electrodes are the main mechanism of plasma edge active stabilization. They are located in the Inner and Outer Diverters of C2-W (Norman).
To provide a current to plasma via these Active Electrodes, two 5kV 5kA pulsed Electrode Power Supplies (EPSU) with a deep voltage regulation capability have been designed and manufactured. The energy required for a long pulse of tens of milliseconds is stored in distributed film capacitor banks.
As the Active Electrodes are in contact with the plasma the Power Supply load is rarely predictable and fast changing during a pulse. To manage this stochastic and highly dynamic load, a high frequency switching technique combined with a multi-level hysteresis control has been implemented in EPSU control system. The EPSU is a voltage controlled system, but a pseudo current control mode loop is included to limit the Active Electrodes current.
The EPSU is very flexible system as it can feed either the Inner or Outer Active Electrodes in both “ground referenced” and “floating” mode and switch between electrodes either during a pulse or between operations. A cabinet with fast thyristor and motorized switches provide all the required configurations.
A fast protection system by means of a thyristor crowbar is included in EPSU.
Eagle Harbor Technologies, Inc. is developing a series stack of solid-state switches to produce a single high voltage switch that can be operated at over 35 kV for magnetron driving applications. During the Phase I program, EHT developed a 15-kV high voltage switch module with isolated power gate drive that could switch 300 A at switching frequencies up to 500 kHz for 10 ms bursts. Robust switching was demonstrated for both IGBTs and SiC MOSFETs. Now in the Phase II program, EHT is developing a higher voltage version for driving a pulsed magnetron at the Lithium Tokamak Experiment at Princeton Plasma Physics Laboratory. This pulsed magnetron driver will produce high voltage, low ripple waveforms. EHT will present experimental testing results for the new high voltage switching modules and system designs for a pulsed magnetron driver.
A novel paralled topology of all solid-state Marx generator using metal–oxide–semiconductor- field-effect transistors (MOSFETs) has been developed for high current with steep edges. This kind of Marx generator has a stacked-stage structure from the bottom to the top with low voltage at the bottom and high voltage at the top. Each stage is combined by many modules in a single board, and one of the modules is also a stage in one Marx generator, all modules in one board with equal potential are connected by the same reference node, likely to the linear transformer driver(LTD). This special connected node makes it possible for the Marx generator outputing high stable pulsed current. All the MOSFETs are driven synchronously through magetic rings. The lack of synchronization of the driven signals and switch fault which may cause unbalance of current and overcurrent problems are studied in this paper, and some methods have been carried for current balancing and overcurrent protection in this new topology. For experimental demonstration, we constructed a twenty-stage Marx modulator in series（eight modules paralled in one stage） using 600-v MOSFETs to operate at 500-v dc input voltage, which output pulses with max voltage of ~-10kv, max current of ~400A(depending on the load), rise time of ~30ns, fall time of ~50ns, controlled pulse width from 500ns to 5us. The whole Marx modulator(without the driver board) has a compact structure with a length of 21cm, a width of 21cm and a height of 24cm.
Pulsed power generation is required by various types of industrial applications, including food and water sterilization, particle acceleration, low-temperature plasma, or Tokamak vertical stabilization, etc. The voltage level of pulsed power generators ranges from a few kV to hundreds of kV. With the past development of semiconductor devices, solid-state pulse generation has been growing due to longer life span, higher voltage rise rate, and more flexibility in pulse shape. Based on the requirements of power rating and operation frequency, different types of semiconductor devices can be selected as switching units [1]. Silicon IGBTs offer a good balance between power capacity and switching frequency. Furthermore, recent development in wide bandgap devices, such as silicon carbide(SiC) IGBTs, has expanded limits in voltage capacity and switching speed considerably, presenting more potential for solid-state pulse generators. To fully exploit the merits of Si and SiC IGBTs, series connection of the devices is usually needed for high-voltage applications since the voltage levels of available commercial Si and SiC IGBTs are no higher than 6.5 kV and 1.2 kV, respectively. Proper control must be implemented to achieve balanced voltage sharing among the IGBTs and to prevent device damage or derating.
In the paper, a review of major contemporary series IGBT control methods for both silicon and silicon carbide IGBTs will be provided, including a self-balancing control implemented and verified experimentally by the authors. The methods will be analyzed and benchmarked based on their performance, loss, speed, and estimated cost. The paper will also compare series-connected devices to a single high rating device in the above-mentioned aspects. The paper is aimed to serve as a guideline for benchmarking as well as selecting series IGBT control methods for high-voltage pulsed power generations with solid-state units.
REFERENCES
[1] W. Jiang, K. Yatsui, K. Takayama, M. Akemoto, E. Nakamura, N. Shimizu, A. Tokuchi, S. Rukin, V. Tarasenko, and A. Panchenko, “Compact solid state-switched pulsed power and its applications,” Proceedings of IEEE, pp. 1180–1196, Jul. 2004.
Magnetic switch pulse-sharpening systems rely on the hysteresis (B-H curve) properties of the material and are affected by the residual state of the previous pulse conducted through the ferromagnetic core in rep-rate applications. Prior work has used an external solenoid field (B-field) coil to reset the core material to discrete values of axial magnetic domain bias. Large pulsed current (100s to 1000 amperes) were required that complicate modulation of the injected solenoid fields. Quantifiable B-H manipulation of the magnetic switch core remanence has been demonstrated with an inductively isolated bipolar pulse modulator. User programmable adjustments are made with lower currents and the volt-second (V-s) product of the reset pulse delivered to the HV output rod that supply the main (sharpened) drive pulse to the load. Series injection of current through the pulse sharpening toroids permit the system to reset core materials in the B-theta direction with lower applied currents. Initial results were performed on a 300 kV pulser at up to 50 Hz. The series injection reset modulator has demonstrated reliable operation of ferrite B-H bias at 50 volts delivering 8 amps using a bipolar stacked MOSFET switch.
Abstract: Power transformer is one of the most important electrical equipment in power system and its reliability is very important to power systems. Therefore, the transformer condition monitoring and fault diagnosis have been paid a lot of attention to by the researchers. The mechanical fault diagnosis method based on the vibration signals of transformer has been widely studied because the measurement system has no direct electrical connection with the transformer and has strong anti-interference ability. The traditional vibration signal analysis method generally analyzes the mixed signals on the surface of transformer oil tank, and can’t effectively evaluate the mechanical state of winding and core. This paper presents a transformer vibration signal separation method based on BP neural network in order to evaluate the mechanical status of winding and core separately. In this paper, the vibration mechanism of transformer winding and core is analyzed and the theory of BP neural network is introduced. Transformer was applied no-load test, steady state short-circuit test and load test. The vibration generated in the no-load test is only produced by the core. The vibration generated in the steady state short-circuit test is only generated by the winding. Using no-load test vibration signals as output layer and using load test vibration signals as input layer, the core BP neural network was built. And using steady state short-circuit test vibration signals as output layer and using load test vibration signals as input layer, the winding BP neural network was built. The core waveform similarity coefficient is 0.813 and the winding waveform similarity coefficient is 0.834, which provide an important technical means for effectively evaluating the mechanical status of winding and core separately.
Keywords: Transformer; BP neural network; Winding; Core; Separation of the vibration signal
Silicon (Si) is the most widely used in power electronic devices. However, due to its limitations regarding blocking voltage and switching frequency, wide band gap (WBG) materials have been under extensive research. Especially, Silicon carbide (SiC) device is expected to replace the Si device in many applications. However, current commercial SiC device ratings still has much room for improvement. In order to achieve high voltage and power, series connection of SiC devices can be considered. A major challenge in connecting devices in series is the voltage unbalance issue among the devices. To balance the voltages, a simple RC snubber circuit can be used [1]. The main issue with the snubber circuit is that the total power loss is higher than without snubber circuit. The active gate drive (AGD) control is another promising alternative. It controls the gate current to adjust dv/dt and therefore can achieve transient voltage balancing. The existing gate current control method for series SiC MOSFETs is limited to up to two devices [2]. To connect multiple SiC devices in series, a snubber circuit is needed to be used with the gate drive [3]. Considering a drawback of using snubber circuit, a AGD control method for multiple series connected devices without snubber circuit is therefore proposed. In this paper, an AGD control for multiple series connected SiC MOSFETs will be presented. It is a closed loop control method and can deal with more than two series connected devices without snubber circuit. The gate current of each device will be controlled by monitoring the unbalanced voltages. The simulation results of the proposed circuit and controller will be presented in the full paper.
[1] K. Vechalapu and S. Bhattacharya, "Performance comparison of 10 kV-15 kV high voltage SiC modules and high voltage switch using series connected 1.7 kV LV SiC MOSFET devices," 2016 IEEE Energy Conversion Congress and Exposition (ECCE), Milwaukee, WI, 2016, pp. 1-8.
[2] Z. Zhang, F. Wang, L. M. Tolbert and B. J. Blalock, "Active Gate Driver for Crosstalk Suppression of SiC Devices in a Phase-Leg Configuration," in IEEE Transactions on Power Electronics, vol. 29, no. 4, pp. 1986-1997, April 2014.
[3] S. Hazra, K. Vechalapu, S. Madhusoodhanan, S. Bhattacharya and K. Hatua, "Gate driver design considerations for silicon carbide MOSFETs including series connected devices," 2017 IEEE Energy Conversion Congress and Exposition (ECCE), Cincinnati, OH, 2017, pp. 1402-1409.
In the Traditional series-connected MOSFETs circuit, each of MOSFET requires a separate external drive circuit, and every drive circuit must be isolated from each other magnetically or optically. This results in increased circuit complexity. Furthermore, the differences of drivers in transmission delay will cause overvoltage and overheating on semiconductor devices, which will threaten the safety of the devices and the entire circuit. In order to decrease the circuit complexity of the series-connected MOSFETs module, and improve the working stability, turn-on speed and portability of it, this paper has designed a new series-connected MOSFETs structure which only requires one single external gate driver to control it. The module consists of ten 1200V SiC MOSFETs. The function of the driver is to trigger and turn off the first MOSFET, and the rest of the MOSFETs are turned on and off by the trigger capacitors. The capacitances of trigger capacitors have significant influence on synchronization of the module, which have been researched through experiments and simulations. After the synchronization was improved, performance parameters of the module can reach to the following points: repetitive frequency over 10kHz, blocking voltage over 8kV, on-state current 20A, and rise time 15ns. The module is lightweight, reliable and easy to use, and has good application in the repetitive frequency pulse power system.
Exploding films have potential applications as fast-opening switches, current interrupters, and ignitors of explosive materials. The exploding film phenomenon is a process in which a high-voltage capacitive discharge is passed through a thin layer of metallized particles on the surface of a dielectric film. Heat generated from the increase in current forces the metallized particles from a solid to a liquid state during an initial current strike. If the metal particles experience an additional rise in current during this liquid state, a restrike may be initiated causing the particles to be liberated from the substrate itself in what is known as a flashover event. It is theorized that an increase in temperature would allow the metallized particles to reach their liquefied state at a quicker rate, resulting in an overall shorter event duration. It is also theorized that the force from the ambient pressure, in addition to the force contributed by the film’s current-induced magnetic field, would need to be overcome to achieve this particle liberation. This work experimentally investigates these hypotheses by varying pressure and temperature conditions while subjecting an aluminum metallized polypropylene film to a 2 kV capacitive discharge. The comparison of electrical characteristics of the current waveform captured during the flashover event versus the plasma’s physical transformation characteristics is presented.
The inductively coupled radiofrequency plasma source of SPIDER and MITICA experiments for the ITER Neutral Beam Test Facility in Padova is designed with highly engineered device called “driver” responsible for the ignition and sustainment of the plasma in the source and delivering 100kW power at 1 MHz. In order to provide high reliability of the source, such a high operating power requires accurate analyses of the potential issues related to breakdowns occurrence in the driver region due to the high electric field present on its components.
Particular effort is given to support the study of the voltage hold off in radiofrequency regime in Consorzio RFX Padova developing a dedicated experimental arrangement: the High Voltage Radio Frequency Test Facility (HVRFTF). This facility is design to be a simple, accessible and flexible device and it aims in particular at the characterization of the voltage hold off of the RF components used for the plasma sources of SPIDER and MITICA. The facility is composed of a RF power amplifier (presently rated for 300 W but upgrade amplifier in the kW range is foreseen for the future) working at the frequency of 1 MHz, a resonant circuit to provide high RF voltage and a vacuum chamber to host the device under test (DUT). The full ratings of this facility comprises pressure range 10-3 – 105 Pa and a delivered testing voltage of up 17 kV rms (presently 10 kV rms are reached) at the frequency of 1 MHz.
In order to provide an electrical qualification of the DUT in terms of maximum operative voltage it is necessary to test the voltage hold off up to the breakdown event. A set of breakdown events are identified being either operational foreseen DUT breakdowns or possible fault condition on RF components and also studied by numerical simulation of an electrical model of the HVRFTF RF circuit. The electrical stresses were found not tolerable by the circuit components, thus a proper passive protection circuit (PPC) based on spark gaps and semiconductor devices is identified and developed considering the ratings of the RF circuit components. The analyses and design of the protection system and of its integration in the HVRFTF RF circuit will be presented and discussed.
As Silicon reaches its theoretical limit in power density capabilities, Gallium Nitride (GaN) High Electron Mobility Transistors (HEMTs) are an ideal option in applications of power electronics due to its wide-bandgap, similar to another proven technology is Silicon Carbide, and high electron mobility due to its special channel. Because of GaN being state of the art technology in power electronics applications, there is a high demand to see if the HEMT semiconductors are reliable in a long-term operation in high power switch-mode conditions. Information on overcurrent capability about GaN HEMTs is not well established thus a demand to investigate the devices exists. The high voltage offering from GaN Systems’, the GaN HEMT GS66508P, was tested in pulsed overcurrent operations in higher ambient temperatures to establish the performance and to observe any operational changes. The device rated at 650 V and 30 A continuous. The goal of this research is to see if the device characteristics change under accelerated reliability testing in different ambient temperatures and to analyze the device degradation that occurs. The device was tested in a RLC ring circuit to minimize series inductance for fast risetime current pulses, and after every pulse set was characterized on an Agilent B1500A semiconductor parameter analyzer.
The LHC Beam Dumping System (LBDS) is a safety critical system ensuring extraction and safe deposition of LHC beams on their respective dump blocks. It consists of 15 extraction and 10 dilution generators per beam and generates up to 1 MA total current at 7 TeV beam energy.
Both extraction and dilution generators employ similar stacks with 10 series connected GTO like thyristors as the main switches. To ensure safe operation of the GTO stack, the individual GTOs within the stack have voltage-sharing resistors and are continuously monitored with the goal to detect failure of an individual GTO. The monitoring system compares the full stack to the lowest potential GTO voltage and generates an alarm/interlock in case of abnormal behaviour. The accuracy of the voltage surveillance system can be compromised by individual GTO leakage current variations if the latter is too high, compared to the current in the voltage sharing resistors. In order to minimise this risk the Automated GTO Measurement and Stack Selection System (AMSSYS) was developed. It is based on simultaneous leakage current measurements of 5 GTOs at multiple (6) voltage levels within our dynamic range (100 V – 3 kV). AMSSYS selects automatically a set of 10 GTOs with best fitting parameters for one stack with very similar behaviour over the whole voltage range at our operational temperature. This minimizes forward blocking voltage (VDS) variations within the stack. The risk of a false detection of a GTO failure was significantly reduced, the sensitivity and reliability of the LBDS stack monitoring was improved. The turn on time of the individual GTOs is dependent on VDS. AMSSYS optimizes the turn on time spread by minimizing VDS variations. The development and design of AMSSYS is outlined. Different test programs for the GTOs are presented and compared.
This paper presents an investigation of trigger circuit for xenon flash lamps driver from the view point of breakdown voltage, high-voltage isolation and reliability. A comparative study of pulse trigger and dc trigger circuits is provided, focusing particularly on the circuit structure of each driver for high voltage isolation. Two kinds of trigger circuits for xenon flash lamp driver are developed with 2.5-kW simmer circuit for maintaining xenon lamp ionization in order to experimentally compare the characteristics of each circuit. The structure of each trigger circuit uses the series trigger method to obtain the advantages of the isolation structure. The experiments are performed with a xenon flash lamp, and the results of the comparison tests are discussed. In addition the results prove that the developed each trigger circuit can be effectively used for xenon flash lamp drivers.
During the past four years, the SIAME laboratory (University of Pau) and the EFFITECH company have combined their skills to design and construct two high different modulators. The name of this project funded by the French Ministry for Defense is “AGIR” which is the French acronym for “Architecture for rectangular pulse generation”.
The first structure (AGIR1) implements a series/parallel combination of resonant modules, optimized for operation in pulse mode and high repetition frequency (200kHz) to reach peak powers up to 10MW. The voltage can be adjusted from 1 to 100kV and the pulse width can be adjusted up to 100μs with a deviation of less than 5%. It is based on half H bridge with SiC components and a pulse transformer with Vitroperm Core.
The second one (AGIR2) is based on a multi-primary pulse transformer powered by four synchronized Blumlein generators. It allows reaching peak powers of several hundred MW. To achieve these specifications, spark gas switches are used and their triggered synchronisation is optimized to allow the complete transfer through the Metglass core transformer. The output performances are a maximum voltage of 300kV, a pulse duration of 600ns with a deviation of 7.5%.
This paper is a global presentation of these pulse modulators.
High voltage DC (HVDC) power supplies require a discharging circuit at their output terminals to dissipate the energy stored in the output filter capacitors when the unit is turned off. This helps to improve the operator safety while connecting and disconnecting loads to the HVDC power supply. There are varying safety standards covering this aspect with most of them requiring that the output voltage be discharged to less than 60 V in 1 sec after the unit is turned off. A high voltage discharge circuit was developed with series connected MOSFETs and discharge resistors. A reliable triggering method without transformer coupling is implemented by inserting a capacitor between gate terminals of series connected MOSFETs. The developed circuit has been tested with a 4000 V DC power supply and the results are presented. Various parameters that limit the end discharge voltage from going below 60 V are discussed. These include gate to source voltage of each MOSFET, value of the coupling capacitance, value of load discharge resistor and the turn on dv/dt of the first MOSFET in the stack.
With the development of energy Internet, the proportion of modular multilevel converter in the system further increases. High-voltage multilevel stepped waves become a typical waveform stressed on the related insulation. In order to study the insulation characteristics of equipment under multilevel stepped wave stress, it is necessary to develop a high-voltage multilevel step-wave voltage source. The regnant single-phase MMC topology in the existing topology is analyzed, while the topology output voltage waveform is biased due to the polarity effect of the load discharge, which does not meet the stability requirements of the voltage source. A new topology structure using positive and negative half-bridge submodules is proposed, which requires fewer switching devices, and is easy to expand to multi-level high voltage, with stable output waveform and flexible frequency and amplitude. The working principle of this topology and the nearest level modulation strategy are introduced. The control system is realized by STM32F407ZGT6. A multistage staircase voltage source prototype is built to verify the feasibility of the proposed topology.
During tests in the frame of commissioning semiconductor based Marx generators short circuit of the load needs to be considered. Hence, a test-circuit comprising an electrolytic load in parallel to a switched inductance has been developed in order to test the over-current protection capabilities of the generator. Thereby, a triggered spark gap switch is used to shorten the output of the Marx generator via the inductance. The test setup has been designed for a voltage range between 4 kV and 150 kV. A set of electrolytic resistors allows for a variation of the load impedance. The resistors have been designed specifically to absorb the energy of a pulse train up to 1 MJ. The current rise under short-circuit conditions is adjusted by means of the inductance. The triggered spark gap enables switching with a defined delay and sufficiently low jitter. It comprises a trigatron-type trigger circuit and adjustable spherical electrodes. The contribution describes the design of resistors and spark gap switch and highlights the design-specialties for the intended application.
This work reports some experimental results of an isolated LC resonant DC-DC converter to generate high voltage regulated. The converter uses a single stage and it is based on a full bridge structure using MOSFETs as switching devices. The output stage of the DC-DC converter is built using series-connected lower voltage modules. The proposed topology is based on the current-fed push-pull DC-DC converter operating and controlled with PWM modulation, active clamping and zero voltage switching. The control technique PWM-Phase-Shift together with a proportional and integral controller is used to in the converter. A high voltage transformer is used to step up AC voltage and its intrinsic capacitance and leakage inductance are utilized to obtain soft-switching zero voltage and zero current switching providing loss reduction, improving efficiency and increasing the power density. The theoretical equations of the circuit operations are studied in detail and an expression for average current in load is presented. The theoretical converter efficiency operating at the nominal output power is almost 90%. The controller proposed is being experimentally used control the output voltage of a 1.8 kW prototype, fed into 400V, which biases a pulsed TWT with voltages of 400V for the grid electrode, 8 kV in the one stage depressed collector and 24 kV in the cathode.
Traditionally, large-amplitude, fast-rising currents and magnetic fields has been measured with electro-magnetic probes such as Rogowski coils or B-dot probes. Such probes are observed to work satisfactorily for many experimental configurations but the probe to digitizer signal is affected by cabling and cabling elements. Measurements must frequently be made in the presence of significant electromagnetic interference imposing unacceptable levels of noise on the probe signals. Furthermore, probe measurements on high voltage electrodes may be problematic if the probes are not sufficiently isolated. An alternative method for measuring currents and magnetic fields involves using the Faraday effect on linearly polarized light propagating in single mode fibers.
Probes utilizing the Faraday effect have been used for many years. Their operation, whereby the magnetic field strength is proportional to the number of probe output “fringes”, is relatively immune to signal cable attenuation losses. Fibers are dielectrics and their electrical insulation reduces breakdown problems near high voltage electrodes. The probe calibration is a material property making in-situ calibrations unnecessary. Previously, the Faraday probe setup required an optical engineer to assemble and align the numerous discreet optical elements (i.e. beam expander, splitter, polarizers and focusing optics). This was time consuming work requiring realignment whenever the assembly was moved. Due to tele-communication advancements, a robust compact Faraday effect optical assembly with fixed alignments is now available at low cost.
Also, due to these advancements, measurements at many different wavelengths are now possible. Theory predicts the Faraday probe sensitivity is inversely proportional to laser wavelength, thus probes of varying sensitivities can be constructed. The authors previously presented the theory and operation of this type probe at 2017 IEEE-PPC London. However, this paper details four Faraday probes optimized for wavelengths of 450 nm, 532 nm, 632 nm & 850 nm and now includes probe calibration efforts.
In our continued effort to reduce the size, weight, and power (SWAP) of high density capacitors we have purchased and tested two experimental capacitor designs from two different vendors. These are critical components for energy storage for a variety of applications. These are paper/foil capacitors with high discharge current capability which are used in Marx bank configurations, and metalized-film energy-storage capacitors which are much more compact but have limited current capability. Tests have included lifetime testing at expected operational current and voltage, testing of some versions at low temperature, and high-current/low-voltage testing of metalized film capacitors, looking for evidence to predict capacitor life time at full voltage. We describe and summarize tests of both types of these capacitors from both vendors.
The High Energy Storage Ring (HESR) for Antiprotons is going to be built at FAIR in Darmstadt on the extended GSI campus.
Charged particle (including protons and antiprotons) of 13 Tm magnetic rigidity will be injected into this synchrotron and storage ring. The injection system of the HESR ring is based on 4 UHV 360 mm long ferrite kickers, each kicker having to generate à 25 T.mm integral field, during 500 ns, with rise time and fall time lower than 220 ns. Each kicker is supplied by a 4000A / 40 kV pulser, based on Blumlein topology, with semi-conductor switches. A prototype of the pulser, using water lines instead of conventional coaxial cables, has been developed to feed the UHV kicker. Electric and magnetic measurements are presented, as well as magnetic transient modelling.
The FAIR accelerator chain comprises a 68 MeV proton injector (pLINAC). Seven identical high power RF systems can supply the RFQ, CCH and CH accelerating structures with up to 2.5 MW. The klystrons require a cathode voltage pulse of -115 kV, 54 A with a flat-top droop of less than 1% over 360 µs. A Klystron modulator has been developed at GSI to meet this demand and two prototype systems are currently under construction in-house.
The selected topology is a capacitor discharge type. A switch-mode current source charges a capacitor bank of 5.4 mF to a programmable voltage of nominally 3.8 kV. A solid state switch assembly generates the output pulse which is stepped up by a ratio 1:33 pulse transformer to the level required by the klystron. The moderate pulse width allows delivering satisfactory flat-top performance without the need for dedicated droop compensation circuitry.
Commissioning of the prototypes is anticipated for the end of 2018 with full availability for operation of the pLINAC RF Test Stand in 2019. Series production of additional seven units will follow.
Spark gap closing switches feature superior switching characteristics in switching speed, current handling and voltage hold-off capability. Unfortunately these switches suffer from electrode erosion and the switching medium has to recover after each pulse, limiting the life-time and repetition rate of the switch. The recovery can be improved by purging the gap to remove heated and ionized gas and to cool the electrodes. The performance of the spark gap system of our pulsed corona demonstrator will be described on the aspect of erosion. The operation is a continuous train of short pulses at high repetition rate. The durations of each pulse is 100 ns, the peak current 3 kA, and repetition rate up to 800 Hz. The experimental erosion results for multiple electrode materials are presented. An analytical and a numerical thermal erosion model is proposed to estimate electrode evaporation rates. Electrode erosion during continuous operation was determined for six materials after 65 million shots. The investigated materials rank, at decreasing quality of performance, as follows: copper > brass > copper/tungsten > stainless steel > tungsten > aluminium. The copper cathode didn’t show any erosion and gained some weight. Detailed SEM images are provided of all tested electrode materials. The developed analytical and a numerical models are both based on the heat equation. The 1D analytical model assumes a static spark radius, and estimates evaporation at a boiling surface. The 2D (rz-plane) FDM (Finite Difference Method) numerical thermal model features a more realistic spark radius which expands according to the trajectory of a shockwave. Time and space resolved heating of the electrode can be simulated, including evaporation and melting. The power input of the models is the product of the measured discharge current and estimated Veff (effective fall voltage) near the electrode. Comparison between modeled and experimental data give insight in the erosion behavior of the tested materials.
Low ON-resistance high voltage analog switch ICs are mainly used for ultra sound imaging system or PCB bare board E-checker system. Especially, PCB E-checker test adjacent PADs open or short state under high voltage over 200V aging condition using Kelvin 4 terminal measurement. To operate switch over 200V, gate oxide thickness should be thicken to guarantee breakdown voltage between gate and source. In this work, Conventional 250V SOI process Lateral High voltage MOSFET (LDMOS) with 20V Gate-Oxide breakdown is used for low ON-resistance analog switch.
To implement bi-directional analog switch, Solid State Relay (SSR) structure was adopted and the level shifting includes the circuits to clamp the gate voltage using a diode and a capacitor. Using HVCMOS technology, this device combines high voltage bilateral DMOS switches and low power CMOS logic to provide efficient control of high voltage analog signals. We use high voltage BCD-SOI process to produce very efficient isolation characteristic and it can be high channel density for greater performance in less space. The proposed structure adopting current boosting short-pulse level shifter and it is suited for low power consumption and high speed switching. The chip size if the developed IC could be reduced 40% compared with the conventional one.
Dielectric elastomer actuators (DEA) are designed to actuate under high voltage stress. The voltage required to actuate is typically at the limit of the compliant material, in this case 3M VHB 4910. As a result, repeated electrical stressing of the material will lead to the occurrence of partial discharges (PD) at the voltage required for actuation. Previous work has shown that repeated large magnitude PD’s precede breakdown, which will be accompanied by the associated ultrasonic signature generated by the large magnitude PD’s. This compliant material also known as a soft dielectric has basis for potential use in novel high voltage applications where traditional solid dielectrics are not a good fit. This paper examines the potential detection of ultrasonic waves and hence PD’s in a soft dielectric immersed in castor oil, using multiple thin film PVDF sensors. This proposed detection method will be used combined with a traditional PD Detection scheme for verification, to confirm actual occurrence and magnitude of the PD.
Neutral beam injection (NBI) is important in fusion science experiments for plasma heating, current drive and diagnostics. Currently, there limited vendors for these systems, and there are no vendors in the United States, which is a potential challenge for the development of private fusion for defense applications. Eagle Harbor Technologies (EHT), Inc. has completed a Phase I program to develop a new power system for NBI that utilizes the state of the art in solid-state switching. EHT has developed a resonant converter that can be scaled to the power levels required for NBI at small-scale validation platform experiments like the Lithium Tokamak Experiment at Princeton Plasma Physics Laboratory. This power system can be used to modulate the NBI voltages over the course of a plasma shot, which can lead to improved control over the plasma. EHT will present initial modeling used to design this system as well as experimental data showing power system operation at 15 kV and 40 A for 10 ms into a test load. Additionally, testing results from a neutral beam system will also be presented.
Evolutionary designs in the packaging of SiC high frequency MOSFETS by CREE® permit the design of compact DC-DCHV power supplies operating at frequencies >100 kHz. A system that uses a DC input of 300 volts from a NiMH rechargeable battery has been built. The battery chemistry is insufficient to source applicable ampacity to efficiently convert the power to elevated kilovolt (kV) potentials at higher converter switching frequencies. The development team acquired compact 6 mF, 125 volt rated capacitors from the Evans Capacitor Company® to provide prompt current delivery for the MOSFET-switched power electronics. A series/parallel combination of Evans capacitors was assembled and placed as an input filter to the converter. A series fuse and contactor are used to isolate the NiMH prime power from the Evan filter capacitance and the converter. The results of Evans capacitor and CREE® MOSFET component testing and integrated system performance (Vout, power, efficiency) will be reported.
In Wuhan National High Magnetic Field Center (WHMFC), a new DC breaker based on the pulsed electromagnetic forming (EMF) technology is developed. The breaker is intended for current interruption in the Battery power supply for protecting the long pulsed magnet. The breaker consists of pulsed magnet (EMF coil), aluminum tube (the main contact of DC breaker) and supporters. The aluminum tube is broken with an electromagnetic repulsion produced by induced eddy current, which is activated in the results of the pulsed magnet powered by capacitor.
First, the aluminum wire electrical explosive DC breaker and the EMF technology are combined. Then, the analytical model based upon the solution of Maxwell is built by the Comsol Multiphysics. And finally, the simulation of the magnetic flux distribution, magnetic force, tube deformation and their interactions is completed. Both simulation and primary experimental results show that the design of the new DC breaker with compact volume and easy maintenance is feasible. In addition to the pulsed high magnetic field facility, the breaker can also be applied to numerical potential industrial fields.
Solid state components have revolutionized high voltage pulsed power systems. The typical device voltage for a solid state device (1-3 kV) is generally much less than the output voltage for a high voltage application. In transformer systems, in particular the impedance varies as the square of the step-up so that the required impedance of the primary system is N^2 lower than the output impedance. This can result in difficult requirements for both the transformer, and the "wires" connecting to the transformer. A mechanical design of the connection which involves striplines can be performed to mitigate this problem. A simpler method using flexible lines would be attractive in many systems as well. In this paper we calculate and measure the inductance and impedance of a variety of configurations based on the ubiquitous “ribbon cable” and it's variants. Both standard cables, standard cables with connectors, and 12 kV cables built by Cicoil and others will be evaluated. Voltage capability and inductance are of particular importance. The RF resistance of ribbon cables in comparison with Litz wire will also be discussed.
For simulation of pulse compression networks the saturation behavior of the pulse compression cores are an important parameter. The data from the supplier does not cover the region we need for the calculation. The challenge is to measure the relative permeability of the material from 1 A/m to 10 kA/m. The approach is threefold. In the range of 1 to 1000 A/m the hysteresis loop is used. In the range from 3 to 3000 A/m a transductor method is chosen. From 0.1 to 10 kA/m the measurements where done with a LC discharge circuit. In this case the ringing frequency is defined by the stray inductivity and the partly saturated core. The results give us a data of saturation behavior of the compression cores. This improves the precision of simulation of pulse compression networks.
The Hermes III accelerator is a 20-MV linear induction accelerator that has been in operation at Sandia National Laboratories since the late 1980’s. Energy is initially stored in the accelerator in ten Marx banks that are discharged into twenty intermediate store capacitors. These intermediate store capacitors are then switched with SF6-insulated high voltage rim-fire switches into eighty parallel pulse forming lines that further condition the pulse before finally delivering it to the twenty induction cavities arrayed along the axis of the machine. Originally, a single 0.9-J KrF laser operating at 248 nm, the output of which was divided into twenty separate beamlets, was used to trigger the rim-fire switches, however as part of a recent upgrade to the accelerator the gas laser system was replaced with a new solid-state laser triggering system. The new system is comprised of 10 flash-lamp pumped, Q-switched Nd:YAG lasers (Tempest 300), each having an energy output of 35~40 mJ at a wavelength of 266 nm. Each laser is responsible for triggering two rim-fire switches. Overall reliability for the accelerator’s operation with these new lasers is increased, and by varying the times at which the individual lasers fire it becomes possible to tailor the shape of the output pulse. The optical layout and other details of this solid-state laser triggering system is presented, along with initial operational data from the HERMES III accelerator using this system.
Solid State Power Emulator Improvements for Power Device Evaluation
Richard Thomas
Army Research Laboratory RDRL-SED-P
Power Conditioning Branch
2800 Powder Mill Rd
Adelphi, MD 20783
Dr. Stephen Bayne
Texas Tech University
Department of Electrical & Computer Engineering Box 43102
Lubbock, TX 79409-3102
The advancement of silicon carbide (SiC) and gallium nitride (GaN) metal oxide semiconductor field effect transistors (MOSFETs) have led to lower switching and conduction losses in high power converters and inverters. Evaluation and characterization of these power MOSFET’s typically requires a costly test circuit comprised of large capacitances, power supplies and elaborate loads to provide and then dissipate the switched power before incorporating these experimental devices in a final circuit where other components of value may be exposed to extreme stresses. The Solid State Power Emulator (SSPE) is a system in which the device under test (DUT) is being exposed to the losses it would be subjected to in a continuous power system. The SSPE allows for evaluation of power devices without wasting the power that would be transferred through the DUT. The SSPE reduces the needed power capability of the supply and load by at least 80% and provides a significant reduction of the energy needed in the evaluation circuit capacitor. The SSPE capabilities will be demonstrated in the evaluation of the Cree third generation SiC MOSFET with a voltage and current rating of 1200V & 50A. The SSPE enables the device to operate in hard and soft cases at various switching frequencies. The power required to perform these types of evaluations with the SSPE and conventional evaluation circuits will be compared and discussed.
Transformer winding vibration frequency response characteristics is the direct reflection of winding mechanical condition. Studying transformer winding vibration frequency response characteristics is very meaningful for winding modal analysis and can be utilized for winding mechanical fault diagnosis. This paper use a single winding to perform frequency response experimentation which mainly consist of harmonic source and control system. The electrical excitation is the constant current which sweeps from 50Hz to 1000Hz. Axial and radial vibration response of different discs are measured respectively. Furthermore, in order to study the influence of various clamping pressures on winding vibration frequency response, a special device is utilized to adjust and measure the clamping pressure on the winding. The result will be illustrated in the full paper.
Transformer vibration is closely related to winding and core condition. Vibration-based transformer condition monitoring and diagnosis technique has been widely used recently, and core vibration occupies a large proportion in the total vibration of operating power transformer. Besides the fluctuant operating voltage, temperature of core silicon steel is considered as the major influential factor on core vibration characteristics. It is necessary to study the influence of temperature on transformer core vibration to improve the accuracy of vibration-based fault diagnosis. In this paper, core vibration characteristics under different temperature was studied. Core temperature was controlled by a temperature chamber. The vibration signals of different positions of core surface, including yokes, limbs and joints, were obtained by piezoelectric vibration sensors under rated excitation voltage, and fast Fourier transform was applied to extract the spectrum and feature parameters which can be used to monitor the mechanical condition. Vibration spectrum under different temperatures were plotted and analyzed. The result shows that with the rising of temperature, the total power of core vibration decreases to the minimum at 48℃ first and then increases. Temperature has a significant influence on the magnetostriction of silicon steel. Vibration amplitude of fundamental frequency increases with the increase of temperature within the working temperature range of the transformer core, and it shows a good linear relationship below 70℃ with growth rates of 0.11-0.37mm∙s^(-2)/℃, which is different from the measuring points on the core surface. The growth rate decreases and shows a saturation tendency of vibration amplitude occurs over 70℃. Temperature-vibration correction curve was proposed to correct the vibration response of transformer core under different temperature and improve the accuracy of vibration-based winding and core condition monitoring and fault diagnosis.
As an increasing number of Silicon Carbide (SiC) devices become commercially available, and as Silicon devices have reached their theoretical power density limits, SiC devices are being utilized in an increasing number of power electronics applications. It is necessary to conduct further reliability testing and analysis on SiC devices, to encourage further adoption of these devices. One parameter that is important to study is the surge current capabilities of SiC diodes, especially for high power converters and high current pulse applications such as Marx generators. In this research, a diode surge current testbed was designed and built to test commercially available 1200 V / 10 A SiC JBS diodes from various manufacturers. The testbed is designed to generate a half sine wave pulse (typical period of 10 ms) with a peak current of 150 A. The testbed current rating can be scaled up as necessary to test higher power devices. The testbed uses an RLC ring down circuit to generate a sine wave with the desired period, and a thyristor as the switch to isolate the device under test following the application of the positive half sine wave pulse. The purpose of this research is to independently verify manufacturer datasheet claims regarding the surge current capabilities of their diodes. JBS diodes from ROHM, STMicro, and CREE/Wolfspeed are tested. The test results are analyzed, and the surge current capabilities compared.
Larger entrance capacitance of ultrahigh-voltage transformers makes it more difficult to generate standard lightning impulses (SLI) in the lightning impulse withstand voltage test (LIWVT), especially for the front time.In order to analyze the feasibility of substituting test impulses with longer front time for SLI in LIWVT, this paper experimentally studied the voltage distribution characteristics of the 1000kV transformer windings subjected to impulses with different front time. The transient voltages of the outmost turn of each disc under impulses with front time of 0.1μs, 1.2μs and 6μs, respectively, were measured with the capacitive voltage sensor.The results showed that, for the innershield-continuous low-voltage winding and the interleaved-continuous-innershield high-voltage winding, whose high voltage leads were respectively located at the top and middle, the maximum voltage distributions were almost independent of the front time; while for the innershield-continuous mid-voltage winding, whose high voltage lead was located at the bottom, the maximum voltage distribution changed as the front time varied. In addition, for all three kinds of windings, the front time had little effect on the frequency components of the transient voltages, but was proportional to the front time of the measured transient voltages, which could significantly influence the dielectric behavior of oil-paper insulation. It could derived from the results that, from the perspective of the transient voltage distribution, longer front time of test impulse was only allowable for windings whose high voltage leads were not located at the bottom. Some discussions were also made on the effect of very fast transient overvoltage (VFTO) on transformer windings.The results in this paper have great referential values to the LIVWT of ultrahigh-voltage transformers and can be utilized to validate circuit models of transformer windings.
One of the intrinsic properties of HV devices is that voltage distribution along their physical structures at transients is largely governed by parasitics. By transients, we rigorously mean a state when dv/dt is not zero, which includes also periodic processes. Examples range from garlands of insulators to HV dividers and their components to machine and transformer windings.
Traditionally, analyses of such structures are made with use of equivalent circuits (EC). Parameters are derived from geometry, and then circuit analysis is executed, either analytically, for simple cases of homogeneous ladder circuits, or numerically. Such analyses have a long history; they are well developed and are fast and powerful. The main problem for numerical analysis is sufficient discretization and faithful derivation of circuit parameters. The latter actually calls for field analysis!
It is possible to analyze the electric field directly by solving Maxwell equations with commercial software packages. Thus, stresses on components and insulation, as well as parasitic parameters for further use, can be determined from first principles.
HV constructions may be “long” compared to characteristic wavelength. Then wave formulations of Maxwell equations need to be invoked. In this paper, we exclude such cases from consideration. We limit analysis to an example of a mixed resistive-capacitive divider driven by a step or a ramp voltage. Modeling is done with Comsol Mutiphysics, which also allows a limited mix of field and circuit analysis. Modeling options for actual resistors and capacitors are described. Conductive problem is solved; both conduction and displacement currents are accounted for. Similar to circuit analysis with EC, it is seen that voltage distribution along the divider can be far from linear; moreover, field along the resistor body can vary greatly. Results of field and EC simulations are compared. Influence of design parameters on the nonuniformity is discussed.
The Sandia National Laboratories Z machine is the world’s most powerful and efficient laboratory radiation source, capable of producing over 2 MJ and 300 TW of x-rays. For primary energy storage, Z uses thirty-six separate Marx banks, each comprised of sixty 2.6 µF capacitors rated for up to 100 kV. The total energy storage on Z is 25 MJ at maximum operating voltage (95 kV). Regular operation of Z requires high reliability of these sub-systems – a failure rate of only 1% would imply a malfunction about every three shots. One way to ensure the reliability of the Marxes is to charge them to high voltage and trigger them into a dummy load to ‘certify’ them before installation on Z.
The Marx Test Bed is the system that allows Z operations to conduct this testing. After many years, we have upgraded the Test Bed with new charging supplies and an updated gas handling and control system. In addition to making the Test Bed itself easier to maintain, the upgrades will allow for more operational flexibility going forward. A further consideration during the upgrade is to provide the Z facility with a more general high voltage testing environment – opening up the opportunity to conduct small-scale high voltage testing on components besides Marxes. Here we present an overview of the Test Bed system, as well as a discussion of the protection and filtering required for the HVAC and HVDC portions of the new charging circuit.
PIN diodes are fabricated using high thermal budgets to form deep P+ anode and N+ cathode regions by either diffusion or ion-implantation. These doping methods come with significant difficulties the in fabrication process such as: doping fluctuation, doping activation and costly equipment. To overcome these issues related to fabrication and doping profile, a new approach called charge plasma (CP) has been developed to design and fabricate a device without involving any of these doping process. A high voltage (breakdown voltage 210 V) PIN diode has been designed and simulated using this concept which eliminates the restrictions noted. The charge plasma approach involves putting intrinsic silicon in contact with different metals. Due to the work function differences between metals and silicon electrons and holes will be induced in the intrinsic silicon. In the case of the PIN diode, the cathode metal electrode will have a work function (φm,C) less than silicon (φSi) and the anode metal electrode will have a work function (φm,A) greater than silicon (φSi), which create “n” and “p” regions respectively. These regions filled with electron and hole plasmas are similar to doped “n” (cathode) and “p” (anode) regions in conventional PIN diodes. Two-dimensional Silvaco Atlas device simulation has been used to evaluate the performance of a high voltage charge plasma PIN (CP-PIN) diode and compare its carrier concentration profile, forward and reverse characteristics, temperature dependency and switching properties with a conventional (doped) PIN diode of similar dimensions.
Explosive Emission Cathodes (EEC), used for the generation of relativistic electron beams, require short rise-time high voltage pulses in order to reduce the extraction of off-energy electrons. To this end a risetime sharpening circuit has been developed at the Los Alamos National Laboratory. The circuit consists of a ~7.8nF water filled peaking capacitor with an integrated self-breakdown switch designed to operate down to -300kV. This unit is intended to reduce the rise time of a 4-stage type E PFN Marx Generator and was used to study the operational characteristics of a planar carbon fiber velvet cathode with respect to varying voltage turn-on time. Simulations of the peaking circuit in situ show a reduction in voltage risetime from over 100ns to roughly 20ns. This paper details the simulation, design, and testing of the peaking circuit.
A classical insulation problem occurring in many pulsed power and high voltage
applications is to feed high-voltage-carrying conductors through a conductive wall
at low (or even ground) potential. An ongoing development for a kicker
application utilizes a pulse forming line (PFL) with a charging voltage of up to
80kV. Since both cable ends must be connected within two oil tanks, customized
high voltage feedthroughs are required.
Due to impedance requirements, a special coaxial cable (CPP20) had been chosen
for this application. In order to generate an output pulse of about 7kA, the PFL
will be switched by a thyratron into a matched load. For insulation reasons, both
the thyratron and load will be situated within grounded oil tanks.
A special high voltage coaxial feedthrough has been developed to feed the inner
conductor of the CPP20 PFL into the oil tanks. The feedthrough is designed to
either be fully immersed in oil or to tolerate a certain air distance to the oil level
in the event of a vertical installation. Due to mechanical constraints, this
feedthrough must be as short as possible.
Several high voltage DC and pulse tests have been performed to find a
compromise between dimension constraints and insulation capabilities, in order
to assess the design. This poster will outline the results of this study, and give
details about the optimum design developed to achieve best performance.
Dry-type air-core reactor used as AC filter reactors is one of the main noise sources in HVDC converter stations. The research on the vibration characteristics of the dry-type air-core reactor is of great significance for the design of noise reduction measures. Based on the assumption of linear vibration system, IEC 60076-6:2007 suggests to use sum frequency, difference frequency and double frequency of the loading current frequency as the decomposition of the total excitation, to measure the acoustic sound level of dry-type air-core reactor. To verify the assumption, this paper aims at the nonlinearity of the dry-type air-core reactor vibration system. The radical vibration of the tested reactor was firstly measured by a laser Doppler vibrometer. The measurement results show that: under single-frequency current, except for the double frequency, the vibration spectrum also contains quadruple frequency and the sixth harmonic frequency components; while under multiple- frequency current, other frequency components also exist besides the sum, difference and double components of the current frequency. The ultra harmonic phenomenon is a typical feature of nonlinear systems. Then the vibration equation was also theoretically established. Considering the coupling of magnetic field and vibration, the perturbation method is used to solve the vibration equation. The theoretical analysis results are consistent with the experiments. It is concluded that the nonlinearity of the vibration system of dry-type air-core reactors is not negligible, and the main cause of the nonlinearity is the coupling of the magnetic field and the motion of the reactor.
The transformer winding will flow a larger impact current in the event of an external short circuit, causing strong vibration, resulting in winding loose and deformation. The loss of short-circuit withstand capability of winding is not a one-time process, but accumulated over the years. As the mechanical condition of winding deteriorates, the safety and stability of transformers are affected. Therefore, it is of great significance to carry out the research on fault diagnosis under short-circuit impact. Different from the vibration signal analysis method using single point or some isolated measuring points which can only reflect local information, the sound signal reflects the overall mechanical condition of the equipment. Short-circuit tests were carried on a 110kV power transformer until the winding impedance changed more than 1%, of which phase A 33 times, phase B 8 times, phase C 16 times. Two microphones were used to record the sound on both sides of bushing outlet so that we can determine which side the fault occurred. According to the characteristics of the sound signal in frequency domain, two fault parameters like odd-even harmonics ratio and sound entropy were proposed. It is found that the entropy increases when the winding loosens, while it decreases when the deformation occurs. The change rate of entropy can reflect the degree of deterioration. The odd-even harmonics ratio significantly increases with a change exceeding 50% in the case of a mechanical failure. According to the proposed fault parameters, a diagnosis on winding failure through the impulsive sound of the transformer is proposed. The impedance measurements and the final results of overhaul validated the accuracy of this new method, what is more, the sensitivity is higher than that of impedance measurements. In addition, this method can carry out on-line testing, while the impedance measurements can only detect offline.
A recent shot series on the Saturn accelerator was done to both better diagnose machine and load operating parameters, and to examine whether more current to Saturn's ring-diode radiation load could increase radiation output. This work is being done in preparation for a potential upgrade to the machine. Critical to this experiment was current measurement in the vacuum magnetically-insulated transmission lines (MITLs) near the load that is not possible with normal shot configurations. Therefore, the load region was configured with only one cathode with its upper and lower anodes, thus driving two parallel MITLs. This configuration allowed installing calibrated current monitors in the MITLs shortly upstream of the load. To examine current scaling the pulse power was configured so that either 12 (1/3 of the machine) or 18 (1/2 of the machine) pulse-forming lines were connected to this single cathode. Shots were done with either a short-circuit or a ring-diode radiation load for both configurations. A full-machine 36-line model was constructed for these configurations and compared to current and voltage measurements within the machine and at the load. Results of these comparisons and validation of the model will be presented.
For over a decade, velocimetry based techniques have been used to infer the electrical current delivered to dynamic materials properties experiments on pulsed power drivers such as the Z Machine.[1] Though originally developed for planar load geometries, in recent years inferring the current delivered to cylindrical coaxial loads has become a valuable diagnostic tool for numerous platforms including Magnetic Liner Inertial Fusion (MagLIF) targets. The process for determining a load current from velocimetry data is colloquially referred to as performing an “unfold.” Various uncertainties that can affect the accuracy of the unfolded current are discussed and accounted for, including errors in the velocimetry measurement itself, sensitivity of the velocimetry to low pressure fluctuations, and the influence of the computational material model chosen. A convenient error relation is developed that indicates for currents above 10 MA the uncertainty is 1-3%; this makes velocimetry determined currents the most accurate high current diagnostic known for pulsed power drivers.
[1] R. W. Lemke, M. D. Knudson, D. E. Bliss, K. Cochrane, J.-P. Davis, A. A. Giunta, H. C. Harjes, and S. A. Slutz, “Magnetically accelerated, ultrahigh velocity flyer plates for shock wave experiments,” Journal of Applied Physics 98, 073530 (2005), http://dx.doi.org/10.1063/1.2084316.
The Plasma, Pulsed Power, and Microwave Laboratory at the University of Michigan (UM) is home to two large pulsed-power drivers, the Michigan Electron Long Beam Accelerator (MELBA) and the Michigan Accelerator for Inductive Z-pinch Experiments (MAIZE). MELBA is a Marx-Abramyan generator capable of generating a 10 kA electron beam at -1 MV for up to 1 µs; this accelerator is currently configured to produce -300 kV and is used for relativistic magnetron, high-power microwave (HPM) generator research. MAIZE is a 3-m-diameter, 40-brick, single-cavity Linear Transformer Driver (LTD) that supplies a fast electrical pulse (0–1 MA in 200 ns) for high energy-density physics (HEDP) research. UM is also constructing a third pulsed-power facility (BLUE) consisting of four, 1.25 m diameter, 10-brick LTD cavities. These four cavities were previously part of Sandia’s 21-cavity Ursa Minor facility, and can be stacked together to increase load voltage.
Recent HPM developments include: a multi-frequency Recirculating Planar Magnetron (RPM) capable of simultaneous production of 1 and 2 GHz signals at the 10's of MW level, a Harmonic RPM utilizing a dual-frequency slow-wave structure to enable low-Q operation, and a novel crossed-field recirculating planar amplifier operating at ~13 dB gain at ~3 GHz.
Pulsed power and HEDP efforts include collaborations with Sandia National Labs on cylindrical foil experiments to investigate the formation and mitigation of the electrothermal and magneto Rayleigh-Taylor instabilities, development of an improved laser-entrance window for pressurized gas targets (such as MagLIF capsules), and development of deuterium pinch neutron sources. In support of these experiments, several diagnostics (multi-frame XUV camera, CW laser, visible and UV multi-frame laser backlighter, visible and UV spectroscopy, differential B-dot probes, high-current Rogowski coil) and driver upgrades (switches, capacitors, improved power feed) are being deployed.
* Research funded by AFOSR #FA9550-15-1-0097, ONR # N00014-13-1-0566/N00014-16-1-2353, DoE #DE-SC0012328, NNSA #DE-NA0003764, Sandia National Labs, and the Michigan Memorial Phoenix Project.
Historically, the design of magnetically insulated vacuum transmission lines (MITLs) was based on successful prior MITL designs, experience & intuition, and simple theory. E&M Particle-in-Cell (PIC) codes were often used more to validate a MITL design than as a primary, iterative design tool. We have used the Screamer circuit code to optimize MITL design based on the assumptions that a constant impedance MITL was desired and that abrupt changes to MITL impedance are to be avoided. We show that the optimum design of an MITL depends strongly on the electrical parameters of the driver and the details of the load. In the case of short circuit (z-pinch) loads, the ideal MITL profiles deviate from an ideal, constant geometric impedance. The presence of inductances inside a transmission line cause the actual impedance of a MITL to be lower than the geometric impedance. This can be a large effect at radii of 15-2 cm and huge effect at smaller radii. Finally, the impedance of a MITL is impacted by vacuum electron flow. We have adopted Mendel's Z-Flow model to obtain an estimate of the "Flow Impedance" of a vacuum transmission line at all locations in the line. We present MITL designs that include details of the load parameters and the effect of vacuum electron flow. Nearly constant-impedance MITLs are designed for drivers with voltages 1-1.5 MV and total currents of ~ 7.5-10 MA.
This study serves to describe three-dimensional particle-in-cell (PIC) simulations of a tunable reflex-triode virtual cathode oscillator (vircator). Experimental data from the compact hard-tube reflex-triode vircator developed at Texas Tech University (TTU) is used to validate simulated results. The vircator developed at TTU is capable of burst-mode operation at pulse repetition rates (PRFs) up to 100 Hz for a period of one second. The vircator is driven by a pulse energy of 158 J, and 600 kV (open circuit) pulse forming network (PFN) based Marx generator1. The vircator is comprised of a bimodal, carbon fiber cathode and pyrolytic graphite anode, with the ability to quickly change the distance between the anode-cathode (A-K) gap, backwall distance, and bottom plate distance between experiments. The PIC simulations have been performed using CST PIC Solver, by Dassault Systemes. The models detail virtual cathode formation and the subsequent extraction of radiated microwave power for a variety of cavity geometries. A working three-dimensional, relativistic, electromagnetic, particle-in-cell model of a vircator allows for quick, predictive results relative to building an experimental setup. The model is used to determine the necessary driving voltages, A-K gap distances, and cathode current densities to extract microwave radiation at a desired. Simulated results aid in identifying mode contributions. Voltage, current, and microwave data are presented and compared against experimental results at different operating conditions.
Helical designs have previously been selected for high power pulsed antennas due to their ability to radiate a high power circularly polarized signal over a relatively wide bandwidth. However, conventional techniques to increase the gain, including both increasing the number of turns of the helix and forming arrays of helical elements, greatly increase the size and weight of the antenna. While several techniques have been developed to reduce the size of helical elements, significantly decreasing the distance between turns of the helix is constrained by the resulting loss of circular polarization and significant reduction of the bandwidth. NanoElectromagnetics LLC has developed techniques to load helical antenna elements with high dielectric constant, high dielectric strength composites to enable significant reduction of the helix turn spacing and overall length while maintaining circular polarization, wide bandwidth, and high power capabilities. With the shorter helical elements, the element spacing in a four-element array has also been reduced while minimizing side lobes. The U.S. Army Armament Research, Development and Engineering Center (ARDEC) and NanoElectromagnetics LLC have leveraged these techniques to design, develop, and validate a high power helical antenna array with significant Size, Weight, and Power (SWaP) improvements over conventional helical antenna arrays, resulting in approximately an order of magnitude reduction in the volume and weight of a four-element array. This work will outline the limitations on conventional helical antenna designs and present modeling and experimental results of a high power pulsed antenna array.
*Work supported by US Army Research, Development and Engineering Center (ARDEC) Directed Energy and Non-Lethal Branch
Nanosecond-pulse Discharges: Mechanism, Characteristics and Applications in IEE CAS
LT-Spice is a powerful simulation language that is specifically optimized for modeling Switch Mode Power conversion. It is not limited to small numbers of nodes and freely available. We are presenting several examples of simulations for popular electronic power factor correction circuits that improve the input power factor of AC Power Supplies by active wave-shaping of the AC input current and the associated avoidance of harmonics.
There is growing interest, from both industry and academia, for a DC power supply capable of outputting 2MW at 1MV. Ideally, the geographic footprint of such a power supply would be as small as possible, meaning that the design will need a high power density. If a fault, such as a short circuit, were to occur in a system that stored a lot of energy the results could be catastrophic. For this reason, the amount of stored energy should be kept to a minimum. The combination of these factors means that many conventional technologies, such as Cockroft-Walton Generators, are ill-suited for the purpose.
A number of other techniques were explored, leading to a more thorough investigation of two possible technologies. Namely the High Frequency Cascade Transformer (HFCT) and the Insulated Core Transformer (ICT). Each were simulated to model the flux loss across them. Physical experiments were also carried out, on a practical test rig, to verify the results of the simulations. Methods for limiting flux loss within both designs will be detailed in the full paper.
The Lockheed Martin Compact Fusion Reactor (CFR) Program endeavors to quickly develop a compact, 100 MWe class fusion power plant with favorable commercial economics and military utility. The CFR uses a diamagnetic, high beta, magnetically encapsulated, linear ring cusp plasma confinement scheme. One advantage of the CFR is how the size lends itself to an iterative, rapid prototyping design cycle, economically demonstrating key physics results while successively increasing in size and engineering complexity towards a Q > 1 device. Experiments began on a tabletop and have grown to the order of a few cubic meters in volume. As the scale of each experiment has grown, so too has the demand for pulsed power systems for plasma experimentation. The CFR pulsed power systems will be reviewed, including power supplies designed for plasma generation, plasma heating, and driving of confinement coil systems. Power supply topologies include IGBT switched DC/DC converters, resonant inverters, and pulse forming networks. A high-level overview of each existing system and their applications on the CFR will be presented, and projections for future pulsed power system development will be discussed.
Due to global concerns over climate change and the depletion of fossil fuels, the worldwide interest in renewable energy sources, such as photovoltaic (PV) and wind, is increased significantly for electricity generation. Grid-connected PV and wind energy conversion systems with battery energy storage are considered to be the fastest developing clean energy technologies because of their increased power capacity and improved efficiency. These systems commonly employ DC-DC bidirectional buck-boost converters, which play a vital role in controlling, storing, and transferring electrical energy between two sources in both directions. However, these converters suffer from large semiconductor loses in silicon (Si) power devices, which are operationally limited due to their intrinsic material properties. Wide bandgap (WBG) power devices, especially gallium nitride (GaN), provide superior advantages with their tremendous operating capabilities and reduced conduction and switching losses. The benefit of replacing all Si devices in a bidirectional buck-boost converter with GaN devices is not well-defined. The main objective of this paper is to investigate the impact of using cascode GaN-FET devices on the converter’s switching performance and energy efficiency. Si-based and GaN-based converters are designed and compared under identical operating conditions of blocking voltages, switching frequencies, and working temperatures to evaluate the converter performance. The switching behavior of the Si and GaN devices is examined through the double-pulse test (DPT), taking different gate resistance values and switch currents into account. The total power loss in the two converters is computed to determine their efficiency over a wide range of switching speeds, input voltages, and output power levels. The outcomes reveal considerable benefits of emerging GaN semiconductor technology in the bidirectional buck-boost converter, leading to significant improvements in switching performance and energy efficiency.
We are reporting on models for a group of inverters that can feed real and reactive power into a utility grid in Grid-Tied mode and is able to transition smoothly transition to islanded mode. In grid tied mode, the inverters are operating in D-Q mode and inject controllable amounts of real and reactive power into the grid. In islanded mode the inverters are grid forming and share power using droop control. We are presenting MatLAB-Simulink models and results of the simulations including the transitions.
The LCLS-II project requires a beam spreader to distribute the 929 kHz electron beam between two undulators and a dump. Three kicker sections and a septum are required to divert beam for each undulator. When the kickers are not pulsed, beam proceeds to the dump. Each kicker magnet is approximately one meter in length and composed of multiple sections. Each section consists of a ferrite and a capacitor to form a lumped element transmission line. The output magnetic field is a <250 ns base-to-base pulse with an amplitude of 4 mT. This paper presents the unique challenges of designing a transmission line kicker and driver with >500 kHz pulse rate. Provided experimental data illustrates pulse-to-pulse repeatability of approximately 100 ppm, transverse field quality, and field-settling time. Additional discussion includes thermal considerations and circuit protection. Adaptation of the kicker system for a low field long flattop pulse is also considered.
The SNS High Voltage Converter Modulators power 92 klystrons at up to 135 kV and 120 A at pulse widths of up to 1.35 ms and 60 Hz to accelerate the H- beam through the linac to 1 GeV. Until recently, pulse droop was tolerated since the low-level RF system has sufficient control margin to compensate. However, with pending upgrades to the accelerator complex and concerns over elevated voltage levels on klystron lifetime, pulse flattening is critical to the future of the SNS.
Early attempts to compensate for pulse droop in the HVCM systems were unsuccessful due to excessive switching losses in the IGBTs in the inverter section. After the addition of a new HVCM control system with enhanced timing features, pulse flattening has been enabled during an 18-month campaign. Operating without droop has resulted in significant improvements in RF performance. In the paper, these improvements and their impact on accelerator performance will be addressed. Modulator operational modes to eliminate droop will also be presented.
*This material is based upon work supported by the US Department of Energy, Office of Science, Office of Basic Energy Sciences under contract number DE-AC05-00OR22725.
The U.S. Naval Research Laboratory (NRL) has designed, constructed, and tested a twelve-bank system designed to be powered exclusively through Li-ion batteries and discharged into a water-based resistive load. An NRL-designed rapid capacitor charger is used to modulate the ≈630 V open-circuit voltage on the battery bank to charge the capacitor to a maximum value of 5 kV. The charger uses a full-bridge IGBT with a modified pulse pattern to avoid saturation of the transformer and is controlled with a microcontroller on a custom circuit board. A series of experiments has been performed to evaluate the performance of the batteries used in the pulsed power system. These experiments simulated both a single-cell and a parallel-cell system at both high discharge and low discharge rates. The experiments show a reduction of life of approximately 75% when the cells are operated in parallel versus as a single cell system. The results of the lifetime study and operational testing will be discussed, as well as an analysis of a post-mortem battery study when end-of-life has been reached on batteries operated at the high discharge levels necessary for pulsed-power systems.
In this contribution we present our findings on efficient ozone generation for environmental applications using a flexible solid-state Marx generator with a DBD plasma reactor. The flexibility of the Marx generator allows for a parametric study over various parameters, such as voltage amplitude, pulse duration, pulse repetition rate and repeated bursts of pulses. We measured ozone concentrations and electrically characterized the plasma for all these parameters. From these measurements we obtained energy efficiencies and ozone yields. The results show high ozone yields (up to 80 g/kWh) and that this yield significantly decreases when the Marx generator is operated in burst mode. We also compare the results to a much faster spark-gap switched nanosecond pulse source with 200 picosecond rise time. The nanosecond pulses from this source produces plasma that is twice as energy efficient as the Marx generator with respect to ozone production. However, the advantage of the Marx generator is that it is much more flexible, compact, and above all: solid-state (and therefore has a significantly longer life time).
Sandia National Laboratories is researching an inertial confinement fusion concept named MagLIF [1] – Magnetized Liner Inertial Fusion. MagLIF utilizes Sandia’s Z Machine to radially compress a small cylindrical volume of pre-magnetized and laser-preheated deuterium fuel. These initial conditions appreciably relax the radial convergence requirements to realize fusion-relevant fuel conditions on the Z accelerator. The configuration of these experiments imposes unique design criteria on the external electromagnetic coils that are used to diffuse magnetic field into the Z target and bulky power flow conductors. Since 2013, several coil designs have been fielded to magnetize the ~75cm3 target region to 10-15T while achieving field uniformity within 1% in the fuel. These Helmholtz-like coil pairs require an extension of the Z vacuum transmission lines, raising total system inductance and limiting peak current to below 18MA. The MagLIF team will study scaling of fusion yield with increased drive current, magnetic field, and deposited laser energy. Planned experimental campaigns require parallel design efforts for Z-Machine power flow hardware and electromagnets to enable 15 – 20T fuel magnetization while simultaneously delivering 18-20 MA machine current. Simultaneous operation of 20 – 25T with 20 - 22MA is also in development. We present the design of a new coil that achieves these field levels in the reduced-inductance Z feed geometry. In this configuration, a single solenoid’s magnetic field external to the bore is used to magnetize the target. Inefficient coupling alongside high field strength requirements generate internal pressures greatly exceeding the yield strength of copper conductors, necessitating internal reinforcement. We discuss the coil design, experimental results, and efforts to understand lifetime. We will also introduce a conceptual design for a new coil pair that will achieve 25-T in a split coil topology compatible with the 20-22MA platform.
One of the problems of powerful pulse technology is the use of transformer inductive storages (TINs) (at the time of energy storage ~ 1-3sec) to generate high-power nanosecond pulses (~ 1012 -1015 Watt and more) [1, 2].
One of possible solutions was offered for consideration in articles [3-5]. The essence of this proposal is to use the well-known and studied phenomenon of the skin-effect to form pulses of the specified duration and power.
The indicated articles done the rationales:
- parameters of the high-resistivity layer covering the primary winding of TIN (thickness, resistivity, etc., depending on the load resistance);
- time parameters of the auxiliary opening switch (plasma opening switch (POS) and semi-conductive opening switch (SOS);
that allow make the best use of the skin effect as opening switch (SEOS).
In this paper, the effect of electrical properties of SEOS materials on the formation of a pulse, as well as the conditions for their cooling to stabilize the parameters of the pulse on the load, are considered. The purpose of this work is to show the possibility of generating a series of powerful nanosecond pulses by TIN-based generators.
Reference
1. "Energy storages" /Ed. D. But, M., Energy, 1991,400p.
2. "Opening switches" /Ed. A. Guenther, M. Kristiansen, and T. Martin PP, NY, 314p.
3. O.G. Egorov "Skin-Effect as the Opening Switch Phenomenon in Combination with the Semiconductor Opening Switch in High Power Generators Based on Inductive Storage" //Book abstract IEEE PMHVC 2016, San Francisco , CA, USA, p.54.
4. O.G. Egorov “Pulsed power generators based on inductive storage and skin effect opening switch (energy correlation and technical application)” // Proc. IEEE PPC, Brighton, UK, 2017, (to be published).
On your own
As the key equipment of power plants and substations, Transformers carry heavy responsibility of power transmission, distribution and transformation of voltage levels. Recent years, with the construction of UHV projects, the highest voltage level in China has increased to more than 1000kV and has formed several UHV networks, enhancing the reliability and efficiency of power supply. However, this also put more stringent requirements on the operation and maintenance of transformers. When a short-circuit fault occurs, the current flowing through transformer windings will be bigger than before, exerting larger electromagnetic forces on transformer windings, resulting in the collapse of the transformer structure. In that case, it is of great meaning and imperative to calculate short-circuit forces precisely.
Different from traditional analytical method, the finite element method can accurately solve the magnetic field distribution in all parts of the windings, and plays an important role in the transformer design process. In order to calculate precise electromagnetic forces, this paper reviews the analytical method of electromagnetic force calculation firstly. And then, the three-dimensional and two-dimensional simulation are carried out by FEM. Because the short-circuit impedance is determined by the leakage flux in circumferential direction, the short-circuit impedance of the three-dimensional model which is more real than two-dimensional model is compared with the measured one to determine the model accuracy. The steady state analysis of the two-dimensional model are compared with the three-dimensional model to verify the accuracy of the magnetic field distribution in some specific location of the windings. Finally, the transient analysis of the two-dimensional model is carried out to find out the variation laws of short-circuit electromagnetic forces applied to the windings with time. The short-circuit electromagnetic forces obtained through the above method is accurate and quickly. And it can provide the necessary reference for the transformer design.
In the framework of the LHC Injector Upgrade (LIU) project, the Proton Synchrotron Booster (PSB) at CERN will be upgraded to receive 160 MeV H– beam from the new linear accelerator LINAC4. During the injection process the two electrons of the H– ions are removed with a stripping foil and the resulting protons are injected into the ring on a machine orbit locally bumped with four individually-pulsed ferrite-cored kicker magnets. The local bump is shifted during the injection process in order to fill the machine aperture (‘phase space painting’) and thereby produce high intensity proton beams. The magnets, two of 36µH inductance and two of 320µH inductance, will be excited by piecewise linearly decreasing currents with a maximum current of 400A for the 36µH magnets. The waveforms will comprise four different programmable slopes, changeable from pulse to pulse in the range of 8μs to 150μs. The four kicker waveforms must be well synchronized and deviate less than 0.5% from a reference waveform. Each pulse generator contains four stages of pre-charged capacitors that one after another are switched to the magnet to generate the current with the required slopes; an additional stage with a power amplifier allows fine control of the slope linearity. The switching stages are connected in series like in a Marx generator but are controlled individually. Special capacitor chargers were developed featuring a floating output, a discharge function, a digital PI-controller and a PROFINET interface. The performance of the first operational prototype is presented and compared to theoretical calculations.
Recent experiments at Sandia National Laboratories have been focused on understanding the physics of power flow in the magnetically insulated transmission lines (MITLs) at the Z Pulsed Power Facility. One outstanding question is determining time-dependent current loss for both ions and electrons in the Z inner MITLs. To answer this question, we have developed a novel velocimetry based current loss detector that is fabricated from multiple materials, e.g. aluminum and gold. Using the fact that different metals, such as Al and Au, have different charged particle stopping powers and equations-of-state, the velocity response of an Al flyer, which is located on the inside of the inner MITLs, can be different from a flyer of identical thickness that is a combined substrate of Al and Au. By analyzing the differences in the velocity responses of these flyers, which is measured using Photonic Doppler Velocimetry (PDV), one can infer time-dependent loss currents of electrons and ions in the inner MITLs. In this presentation, we describe the physics of this detector, and show recent results of this detector from Z experiments.
Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.
Utilization of nanosecond pulsed electric fields (nsPEFs) has been widely studied as a novel approach to induce cellular stress for biological and medical applications. Endoplasmic reticulum (ER) stress has become to be considered as a cause of diseases as diabetes mellitus, Parkinson's disease, and so on. The objective of this study is to activate the ER stress response, which is a built-in function in cell to avoid ER stress, by applying pulsed high electric fields. Eventually, we aim to establish therapy and prevention methods for the diseases. When we previously applied nsPEFs on skeletal muscle of mice, a little of ER stress response was induced. Therefore, nsPEFs were applied on suspended solution of cultured cells to investigate adequate condition to activate stress response in this study. The nsPEFs of 14-ns and 70-ns pulse-width were applied on suspended solution of HeLa and MEF cells. The applied pulse electric field strength was changed with the charging voltage of Blumlein pulse forming network and selection of a cuvette with electrodes. The number of applied 14-ns and 70-ns PEFs changed from 10 to 30 and from 10 to 500, respectively. It is well known that eIF2α phosphorylation is induced by various cellular stress, especially ER stress. Western blotting was used to analyze the nsPEFs-induced eIF2α phosphorylation. The cell viability after nsPEFs exposure was measured with WST-8 assay (water soluble tetrazolium salts). It was confirmed that 14-ns and 70-ns pulses activated ER stress response. The 14-ns pulses activated a little response. The 70-ns pulses activated stress response as large as thapsigargin treatment (enzyme inhibitor inducing ER stress). In addition, the viable cells decreased with increase of the number of applied pulses by 70-ns pulse application. However, it was confirmed that 70-ns pulses could activate sufficiently large stress response even when most cells survived.
The high stability flat-top pulsed magnetic field has both high field strength and high field stability simultaneously. That is necessary for some high precision scientific experiment, for example, nuclear magnetic resonance (NMR) and heat specific measurement, whose resolution of experimental data is closely related to both field strength and field stability.
In this paper, an innovative method for generating high stability flat-top pulsed magnetic field base on battery power supply is proposed. In this system, a new modified active filter including a bypass circuit and a divider resistor is developed. The bypass circuit is in parallel with the magnet to form a parallel branch, with which the divider resistor is in series. Different from traditional application, the bypass circuit consisting of IGBTs works as a controllable constant current source rather than in the switch model. The current of bypass circuit can be adjusted by the driving voltage of IGBTs to achieve the continuously regulation of the voltage division ratio between divider resistor and magnet, then the current of magnet maintain stability and high stability flat-top pulsed magnetic field can be obtained. In order to generate the 40T/100ms flat-top pulsed magnetic field with the ripple of 100ppm, three parallel FZ3600R17HP4 IGBTs and 5mΩ divider resistor are used, the driving voltage of IGBTs are controlled by a high speed digital PID system with DSP28335 and the feedback signal is the magnetic field measured by pick-up coil.
The designed system can provide high-stability flat-top pulsed magnetic field and offer better experimental physical environment for NMR and heat specific measurement.
Nonlinear transmission lines (NLTLs) have been studied for application in aerospace radars, telecommunication, and medical devices. The radiofrequency (RF) pulses generated by the NLTLs can be radiated by antennas connected at the output of the lines. There have been relatively few articles that show experimental results regarding the transmission and receiving of RF signal generated by NLTLs employing dipole antennas, in special with nonlinear lumped capacitive lines. In this paper, the results of simulations and measurements performed on a 30-section capacitive NLTL using low-voltage varactor diodes (BB809) as nonlinear elements are presented. The pulsed RF signal around 40 MHz was generated by a low voltage pulsed DC signal at the input of the line. The NLTL was characterized using time- and frequency-domains´ analysis of the pulsed RF signals measured on a resistive load connected at the line output and on transmitting and receiving half-wave dipole antennas.
*Work Supported by US Air Force Office of Scientific Research under contract no. FA9550-18-1-0111.
For surface modification in industry, a dielectric barrier discharge (DBD) under atmospheric pressure with air as carrier gas is feasible and cost efficient, see [1]. However, a stable and reproducible plasma is required for industrial applications. To fulfil the industry specifications, the plasma ignition shall be analysed under the following aspects: pulse form, edge steepness, pulse width, pulse repetition rate, transferred power, electrode area and time resolved optical properties.
The in house rectangle pulse generator (DC-330 kHz, ±8 kV) based on a series circuit of MOSFETs in push/pull mode will be compared to the sine pulse generator “Minipuls” from GBS electronics. It will also be analysed whether the voltage overshoot depends on the oscilloscope sample rate. Measurements have already shown that the pulse form has an influence on the ignition behaviour (ignition duration: rectangle pulse ~300 ns, sine pulse ~30000 ns). The time resolved optical and current measurements indicate correlations, which are under further investigation.
Acknowledgements: HGS-Hire Graduate School by HIC for FAIR
[1] Atmospheric pressure plasmas: A review, C. Tendero et. al., Spectrochimica Acta Part B: Atomic Spectroscopy, Vol. 61 2-30, Jan 2006
The Virtual Cathode Oscillator (Vircator) is one of the methods to generate high power microwaves. Simplicity not needing external magnetic fields, and high-power capabilities are some of the advantages of Vircators. However, low efficiency is a serious problem. In the Vircator driven by ETIGO-IV generator, it is known that electron beam, 40% is lost at the mesh anode with transmittance of 65%. By using a high transmittance anode, it is shown that the losses at the anode decreased and microwave output improved. In this research, the characteristics of the microwave and losses at the anode were measured with high transmittance (85, 90, 95%) wire/mesh anodes. Increasing transmittance of anode, microwave energy and energy efficiency decreased. At the transmittance of 95%, microwave did not oscillate. In addition, it was confirmed that oscillation time became shorter, and oscillated at multiple frequencies over a wide frequency band due to increasing transmittance of the anode. Microwave energy and energy efficiency were not affected by the difference in anode structure. In the wire anode, it was confirmed that the microwave oscillates gently, then it oscillates rapidly. In the measurement of electron beam losses, it was confirmed that the electron beam, about 20% on the average is lost at the anode, and the losses of the electron beam did not decrease depending on the transmittance of anode.
The Antiproton Decelerator (AD) at CERN uses a magnetic horn as a focusing lens to capture particles generated by shooting a proton beam into a target. It is powered by a capacitor bank discharge circuit generating current pulses with an amplitude of 400kA and a duration of 30µs. As part of the ongoing consolidation of the magnetic horn, a test-bench installation was built with an increased peak current to 500kA to test, validate and stress the performance of the new horn. This has been achieved through the upgrade of an existing system developed in the 80’s by increasing the size of the capacitor bank and the operating voltage from 6kV to 7kV, by the replacement of ignitron pulsed power switches by stacked solid-state switches with an improved triggering system and by the implementation of a passive snubber circuit for protection against reverse blocking voltages. A newly developed control system monitors the state of the entire magnetic horn installation and allows triggering of the switches only if the installation is in a safe operating condition. This paper describes the work done on the pulsed power supply used to power the horn. First test results are presented for operation up to 500kA.
The transport of electron beams through moderate pressure gases is a challenging problem that involves beam evolution due to self fields and plasma generation through complex chemical pathways. The modeling of such experiments is made even more difficult due to low repetition rate, shot to shot variation, and large geometries. This work reports on the development and characterization of a field emission-based electron beam system designed specifically for model validation. The source is driven either by a self-matched transmission line switched through a spark gap or by a solid state pulser. In both cases the nominal pulsewidth is on the order of 10 ns with a beam voltage of 10 kV and a cathode (carbon velvet) current density of approximately 200 A/cm$^2$. Current and voltage diagnostics, beam imaging, and energy spectra are presented, followed by a discussion of future work.
Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.
There is a significant effort to measure current losses on the Z machine at different points along the magnetically-insulated transmission line (MITL). The convolute region is one of the key locations where models, measurements, and simulations say much of this current loss occurs. Positive ions comprise one component of this current loss in addition to electrons and negative ions crossing the A-K gap. We have developed a diagnostic to detect protons that cross the A-K gap directly across from the convolute anode posts. We call this diagnostic CIDZ for Cathode Ion Detector on Z. The first iteration of CIDZ included radiochromic films (RCF) that measures the time-integrated ion fluence. On the first “Power Flow” Z shot series, we successfully fielded this design and measured an ion signature. Subsequent calibration of the RCF films suggests that the ion fluence was around 1x1012 protons/cm2. There was strong evidence of partial magnetic insulation of the ion beam. The second iteration of CIDZ attempted to include faraday cups to collect a time-resolved measurement equivalent in addition to nearby RFC film with slots to allow for more ions to impact the film. Unfortunately this design suffered from heavy debris impact and a large external cable pickup due to impact from free-electrons outside of CIDZ. A refined design will include additional cable shielding.
Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA-0003525.
In a high power vacuum insulated transmission line, a critical current must be reached such that magnetic insulation is obtained. Before this magnetic insulation, the threshold for vacuum field emission is reached leading to emission of electrons from the cathode. Additionally, the large conduction currents lead to Joule heating of the electrodes and subsequent sublimation of any impurities on the electrode surfaces. The combination of these processes may lead to plasma formation and loss of insulation. In this work, electron and ion flow patterns are analyzed for a 30-degree wedge of a full 31-cm convolute section used on the Z-accelerator. An electromagnetic, particle-in-cell (PIC) code is used to couple the transient wave phenomena with charged particle physics. Emission of neutral water from the electrode surface is included and allowed to interact with the charged species. Load currents are calculated with and without charged species to estimate a current loss metric. This three-dimensional example of the convolute requires cell sizes on the order of 10$^{-6}$ m near the electrode surfaces due to anticipated plasma densities on the order of 10$^{14}$ cm$^{-3}$ - 10$^{16}$ cm$^{-3}$. The limitations of using an explicit PIC code on modern computing hardware are examined as this example stresses computational requirements.
Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525. This work was supported by the Laboratory Directed Research and Development (LDRD) program, Project No. 209240 at Sandia National Laboratories.
This paper presents the design and characterization of a compact, battery-powered, high-voltage impulse generator intended to produce nanosecond pulses for UWB applications. This system was developed in-house at the Army Research Laboratory and it consists of a +14.8-V to -30-kV charger and a -30-kV to +240-kV Marx generator circuit. The output of the 240-kV, 10-J Marx generator is used to drive various types of antennas.
With the continuing push to minimize size, weight, power and cost (SWaP-C) boundaries for mobile platforms, it becomes a challenging task to contain >200 kV in a small form factor. This paper explores some of the challenges we faced in the laboratory in designing for a compact form, and it provides preliminary results for the impulse generator’s system performance.
A cancer treatment by an ultra-short pulse high electric field is one of new biological applications. This work focuses on the development of a high power nanosecond pulsed electromagnetic wave generator for the cancer treatment, and effect on a cancer cell by the developed device. We have developed a burst pulse generator that outputs the multiple pulses continuously by a NLTL (Nonlinear Transmission Line) using magnetic switches [1]. The NLTL makes the pulse train by delaying the propagation of the pulse through the magnetic switch of each ladder. The number of pulses in the pulse train can be varied by the number of the units of the magnetic switch. In this study, we have investigated whether applying a burst pulse to the cancer cells is more effective than applying a single pulse to the cancer cells under the same conditions. The condition of the burst pulse is the electric field strength of 80 kV/cm, the frequency of 130 MHz, the repetition rate of 1 pps and the number of LC stages of 5. The condition of the single pulse is also same electric field strength and frequency. However, for same applying power, the number of repetitions is 5 pps. Here, we have used the yeast that is simulated the cancer cell and have used the PI staining method in order to confirm how many the cancer cells dead by applying the pulses. The result has shown applying the burst pulse to the cancer cells is possible to give lethal impact than that of the single pulse.
[1] Keita Yasu, Yasushi Minamitani, Ken Nukaga, “Development of High Power Burst Pulse Generator Based on Magnetic Switch for Bioelectrics,” Proc. of 2016 International Power Modulator and High Voltage Conference, pp. 392-396, (2016)
The LHC Injector Upgrade project aims at increasing the performance of the LHC injectors and includes the replacement of Linac2 by Linac4 as injector to the Proton Synchrotron Booster (PSB). In order to distribute the 160 MeV beam from Linac4 to the four rings of the PSB, a new distributor system has been built. The required five pulse generators have a pulse length ranging from 20 µs to 620 µs with a rise time of less than 2 µs and a maximum flattop ripple of ±1 %. Four generators distribute the beam to the four vertically stacked PSB rings whilst the fifth generator dumps the beam tail. The basic generator hardware consists of a Pulse Forming Network (PFN), two series connected IGBT switches, an optical trigger interface and a Fast Interlocks Detection System (FIDS) for hardware protection. During the iterative development phase, solutions to fulfil the required performance relating to rise time, jitter, long-term stability and reliability have been implemented and will be presented. The different failure modes of the generator have been studied, identified and mitigated, mostly by means of the FIDS. The FIDS is integrated into a larger control system infrastructure via an Ethernet backbone and industrial bus (Profinet). The industrial bus manages all PLC controllers for slow supervision of ancillary systems and remote control systems from the operating room. The process of the final tuning and the reached performance are outlined. This paper presents the development phases of the generator, the hardware protection system, the final design choices and its performance.
Ozone has been used for several applications as sterilization and deodorization because of the strong oxidizing power with low environmental load. The efficient and high-concentrated ozone-generation technology has been required. Ozone generation using nanosecond pulsed power has a feature of high efficiency but not high concentration. We have adopted a thin reactor in the ozone generation to improve the ozone concentration. The thin outer cylindrical electrode selectively utilizes high-density streamer discharges near the inner wire electrode in coaxial reactor. The ozone concentration increased but then decreased with increase in the diameter. In this study, the dependence of the state of streamer discharges in the reactor on the generated ozone was discussed. The relationship between distribution of the propagating voltage pulse in the reactor and the ozone produced by the streamer discharges was investigated when three coaxial cylindrical reactors were directly connected. The ozone concentration produced in the intermediate reactors was lower than the other two reactors. Although the voltage pulse propagated with decay in the reactors, the pulse was reflected and superimposed at the open-circuit end of reactor. Because the voltage distribution applied on the intermediate reactor would become lower, the streamer discharges in the intermediate reactor would be weak or partial. Because a part of energy of the supplied pulse was only consumed in the reactor in this experimental system, relatively long and large wave tail appeared even when a voltage pulse of 1.5 ns width was applied on the reactor. Therefore, secondary streamers and/or spark discharges, which should be inefficient for discharge-chemical reaction, might occur in reactor. We have considered effect of the tail on streamer discharges with a simulated electrodes as a reactor. Here, it was considered by using the ozone producing reactor. Effective ozone production and design of the reactor was discussed based on the results.
Cancer cell populations contain proliferating, quiescent, and dead cells that drive tumor growth and cancer metastasis based on the surrounding microenvironment [1]. Many existing cancer treatments are highly cytotoxic and predominantly target the proliferating cells, which drive cancer development [2]. This begs the question about whether electric pulses (EPs), which are under investigation for cancer treatment [3], preferentially target proliferating or quiescent cells. To address this, we apply a mathematical model of cancer cell population dynamics based on coupled differential equations representing proliferating and quiescent cell growth [1] to experiments applying EPs of various pulse durations, electric fields, and pulse number to Jurkat cells. EPs above a specific energy threshold severely retarded cell growth such that it did approach the standard “S-curve” exhibited by untreated cells, despite replenishing the cell growth media daily. Fitting experimental data to the mathematical model [1] showed that that the proliferating cell population decreased with a concomitant increase in quiescent cell death rate as we applied additional EPs. This caused the quiescent cells to consume additional nutrients, further reducing the proliferating cell population and driving down the final steady state cell population for potential treatments. The implications of these results for tuning EPs for applications ranging from cancer treatment to cell growth stimulation for wound healing will be discussed.
[1] A. L. Garner, Y. Y. Lau, D. W. Jordan, M. D. Uhler, and R. M. Gilgenbach, “Implications of a simple mathematical model to cancer cell population dynamics,” Cell Prolif., vol. 39, pp. 15-28, 2006.
[2] C. Sawyers, “Targeted cancer therapy,” Nature, vol. 432, no. 7015, pp. 294–297, 2004.
[3] R. Nuccitelli, R. Wood, M. Kreis, B. Athos, J. Huynh, K. Lui, P. Nuccitelli, and E. H. Epstein Jr, “First-in-human trial of nanoelectroablation therapy for basal cell carcinoma: Proof of method,” Exp. Dermatol., vol. 23, pp. 135-137, 2014.
Virtual cathode oscillator is a class of high power microwave source. The virtual cathode oscillator uses oscillation of virtual cathode formed behind the anode to generate high power microwave. In spite of its low efficiencies, the virtual cathode oscillator has been studied widely because of its simple structure and its availability of high voltage. To increase the efficiencies, various type of virtual cathode oscillator have been studied such as axial virtual cathode oscillator, reflex triode virtual cathode oscillator, and coaxial virtual cathode oscillator. Most experiments on the virtual cathode oscillator have been made using high voltage pulses from several kilovolts level to megavolts level. Because the virtual cathode oscillator is operated using high voltage pulses, it is hard to construct compact high power microwave system.
The purpose of this work is to design a compact high power microwave system and analyze the electrical characteristics of the system. Axial virtual cathode oscillator is used as a high power microwave source and analyzed using relatively low-level voltage pulses. As a pulsed power generator, Marx generator applies 150-kV voltage pulses with pulse width of 200-ns. Aluminium cathode and stainless-steel mesh anode are used as a vacuum diode. The AK-gap is set to 1.2-cm. This paper presents experimental results of the compact virtual cathode oscillator and the electrical characteristics of the system when it is operated in the low voltage level circumstances.
We describe the experimental demonstration of a linear actuator and cylindrical motor which utilizes the controlled turn-less motor (CTM) principle discussed elsewhere in this conference. The technology combines low voltage high current inverters with turn-less electro mechanics and high power density batteries and micro channel coolers resulting in extremely high force density and power density electro mechanical devices.
A key element of the development is the design and construction of the extremely compact inverters and their integration into complete electro-mechanical devises such as motors and actuators. The complete motor control system which includes the MOSFETs, H-bridge, gate drivers, MCUs, encoder and micro channel cooling will be described and the experimental results will be discussed.
The gas temperature within a pin-to-pin air gap discharge created by nanosecond-pulsed high-frequency voltage pulses was studied via emission spectroscopy of the N2(C-B) state. The temperature evolution was studied over a burst series of 10 pulses with the frequency varying from 1 kHz to 250 kHz. Emission from each individual discharge was studied to determine the temperature associated with each pulse. The results show that within a burst of repetitive pulses, the effects of the previous pulse caused a significant temperature increase at the ignition of the following pulse if the time between pulses was less than 0.1 ms. This indicated that the discharges were “coupled” for frequencies above 10 kHz and temperatures continued to rise throughout the burst series. At even higher frequencies, the rapid temperature rise within the burst of pulses was found to be several thousand Kelvin over a few microseconds. This type of rapid heating could be very advantageous in controlling ignition probability in a combustion system. In this work, the energy deposition is examined and related to the temperature evolution. The nanosecond-pulsed discharge coupling theory is also discussed in term of plasma parameters.
Two pulsed power systems have been upgraded for the g-2 experiment at Fermilab. The Pbar Lithium Lens supply previously ran with a half sine pulsed current of 75 kA peak, 400 us duration and a repetition rate of 0.45 pps. For the g-2 experiment, the peak current was reduced to 25 kA, but the repetition rate was increased to an average of 12 pps. Furthermore, the pulses come in a burst of 8 with 10 ms between each of 8 pulses and then a delay until the next burst. The charging rate has still gone up by a factor of 20 due to the burst speed. A major challenge for the upgrade was to charge the capacitor bank while keeping the power line loading and charging supply cost to a reasonable level. This paper will discuss how those issues were solved and results from the operational system.
During the last two decades, HPM (High Power Microwaves) effects on systems such as electronic devices and sometimes biological samples have been studied in CEA/Gramat. These studies led to actual designs of several HPM generators, each corresponding to a specific technology or experimental area. This paper presents two high-sized electromagnetic simulators named Hyperion and MELUSINE.
The “Hyperion” facility is a building comprised of a large testing area (18x18x18 m) and a RF power generation room. It uses two large reflectors to radiate a plane wave in that testing area. Targets are positioned on a 50-ton swing bridge, allowing testing on heavy objects. A wide variety of sources is available to generate field levels between 10 kV/m and 1 MV/m. As an example, several relativistic magnetrons were associated with a Tesla generator, providing electromagnetic power up to several hundreds of MW in L band.
MELUSINE is a 100 m long, 6 m radius half-cylinder semi-anechoic tunnel. This facility is used for the study of 100 MW to 1 GW class HPM relativistic generators above 1 GHz. As an example, the ultra-compact generator CLAIRE has been deployed and used in Melusine. CLAIRE is low-volume (0.3 m3) source capable of producing powers up to several hundreds of MW in X band.
Classic and specific metrology equipment is deployed in these electromagnetic simulators, such as field sensors, infrared thin film EMIR, power couplers. Also, specific measurement technique such as calibration or new optoelectronic field measurement methods is presented.
This paper describes a high-speed imaging setup for capturing exploding detonators with short exposure time cameras. With three, commercially available, high intensity, pulsed xenon light sources (> 107 candela intensity) satisfactory image quality was achieved for a minimum exposure time. Beyond that the combined flash lamp output was too dim at desired shorter exposure times. As such, an alternative lamp system was pursued that would provide the higher required light intensity output.
Two types of lamp arrays were designed and tested. A large lamp array comprised of a few high energy flash lamps and small lamp array comprised of many low-energy flash lamps. The large lamp array is intended for multiple shot use and is placed behind a protective sheet of polycarbonate far from the detonator. The small lamp array with low cost flash lamps is intended for one time use and will be placed within much closer to the detonator. As such, these small lamps were driven beyond the manufacturer’s specified energy limit to maximize the light intensity output. Multiple six stage, Rayleigh line PFNs were constructed to deliver the lamp energy. The PFN is modeled using LTSpice to verify circuit operation. Experimental measurements on the voltage and current were collected and compared to simulated results. A photodiode is used to measure relative light intensity from the different light sources. Finally, images are shown of the exploding detonator at various frame rates.
Ablation outcome assessment is critical for confocal therapy. Irreversible electroporation (IRE) is a new emerging tumor treatment in recent years. The variation of the impedance has been proved the ability to reflect the treatment outcome of IRE. However, nowadays the impedance is mostly measured by the treatment electrodes which are used to apply pulses. The information obtained by that is limited to reflect the lethal results. Here, we constructed a four-electrode system. Two of four electrodes in the diagonal are used to apply high voltage pulses, resulting in the IRE in potato tuber. Before and after treatment, the impedance between different electrode-pairs is measured by a commercial impedance analyzer. The impedance from the electrodes which are employed to deliver pulses or not is compared. The results show that the impedance change from the electrodes which are not used to apply pulses have less deviation and give more information on the ablation boundary. This information from the electrodes which are not used to apply pulses that time is significant for the ablation outcome detection, especially in the clinical trials with multi-electrode.
Nowadays Nonlinear Transmission Lines-NLTL have been studied for RF generation to be applied in different systems of communications such as satellite, military and biomedical applications. NLTL is built by using a network composed by inductors and capacitors, where at least one of these components need to have a nonlinear behavior. One limiting factor for applications is in the peak power range, and a way to increasing this is improving voltage modulation depth -VMD, which is main goal of this work. In this work, it is simulated and built three configurations of NLTLs using ceramic capacitors to improve VMD and consequently the RF power generated at the output. As shown by the results the cross-link configuration is the best method due to the highest VMD generated.
To understand effects of processing and surface chemistry on nanodielectric material properties, we study interaction of molecular species used in synthesis, functionalization, and suspension of BaTiO3 particles in solution. We investigate interaction of water with BaTiO3 using density functional theory (DFT) applied to slab configurations. Slabs meant to represent the environment of a nanoparticle surface are constructed of sufficient size to approach convergence for defect and surface-reacted molecules while remaining computationally tractable. We investigate the dissociation/formation of H2O on stoichiometric and O-deficient surfaces, including successive placement of hydrogen to form H-decorated surfaces.
H2O adsorption, H-decoration, and Oxygen vacancies all lead to significant relaxation and notable polarization effects in the surrounding material. We pay particular attention to the oxygen vacancy, as it can significantly alter the electrical properties of the material, and is an expected defect when processing or functionalizing under acidic conditions. Given ligands used to functionalize particles are often acidic, a thorough understanding of this defect is important. Oxygen deficiency can cause particles to have a reduced band gap, or to become metallic. Macroscopic materials made with these particles will show effects such as changes in overall conductivity, breakdown, and dielectric properties. The oxygen vacancy is significantly lower energy a few layers into the slab than at the surface, likely due screening on a charged defect. We discuss this result and the required level of theory for accurate description of formation and selected properties of the oxygen vacancy.
DFT-derived results and interpretation are compared with concurrent experiments using solvated BaTiO3 nanoparticles.
This work was supported by AFOSR FA9550-16DCOR281. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia LLC, a wholly owned subsidiary of Honeywell International Inc. for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525
Linac Coherent Light Source (LCLS) at SLAC has started to run since April, 2009 and is driven by eighty-four klystrons. Each klystron is powered by a line-type modulator with a high power thyratron as a switch. The thyratron life time has become an important concern since its invention. SLAC has accumulated thyratron life data in past long time operation. In this report, we will review the life of thyratron tubes in LCLS and some life extension developments of the thyratron will be presented.
Surface flashover of insulators in vacuum has been investigated for decades, the most widely accepted mechanism for the process is secondary electron emission avalanche (SEEA) and the gases generated from surface discharge play a crucial role. In this paper, the flashover was excited by a microsecond single pulsed generator with a rise time of 500 ns and a full width at half maximum of 8 μs. The gap between the finger electrodes is 1 mm. Optical diagnostic methods were adopt to investigate the luminous properties of polytetrafluoroethylene (PTFE), polymethylmethacrylate (PMMA) and polyamide (PA6). Time evolution of luminescene signal was measured by PMT with wavelength range from 300 nm to 650 nm. The emission light spectrum was also measured to detect the gases generated from surface discharge. The results show that the luminescence signal appears approximately 300 ns before flashover, and flashover occurs approximately 80 ns before light intensity sharply increases. The luminescence signal was also measured when flashover didn’t occur in the lower applied voltage. The result shows that the luminescence signal intensity is weakened with increase of applied pulsed numbers. The emission light spectrum is mainly in range of 400-800 nm, and the C2 peak is distinctly detected in three materials, indicating that the carbon chain of polymeric materials is broken when flashover occurs. H2 and F peaks are also detected in PTFE, and H peak in PMMA and PA6. These may come from the desorption gases or decomposition products in the course of surface flashover.
Electric pulses can effectively treat advanced, inoperable, chemo- and radiation-resistance cancers. This technique, known as electroporation, involves the application of high intensity, short duration (nanosecond (ns)) pulses. Depending upon the energy applied, which is a function of the magnitude, length, and number of pulses, one can potentially induce no membrane change, temporary opening of the cell membrane after which cell survives (reversible electroporation), or permanently open the cell membrane and induce cell death (irreversible electroporation (IRE)). When a suitable electric field is applied to the cancer cells, IRE permanently permeabilizes the cell membranes, resulting in a loss of cell homeostasis [1, 2]. The advantages of IRE are that it has a drug-free, non-thermal mechanism of action that allows focal tissue ablation and requires a very short time of application. IRE also allows ablation of tumors in close proximity to other tissues with no protein denaturation.
This study evaluates the efficacy of IRE for various ns pulse parameters on human, triple negative breast cancer cells, MDA-MB-231. Specifically, we applied different numbers (8, 20, and 40) of 35kV/cm, 60ns pulses to assess the impact of delivering different energy levels and assessed the viabilities were 24 and 72h after treatment. The viabilities varied from 71% to 33% following 8 pulses (untreated control at 100%), 28% to 8% after 20 pulses and 21% to 7% after 40 pulses. These results indicate the sustainable effect of a single application of electrical pulses, even after 72h. This suggests that it is possible to tune cell death and proliferation control to achieve complete tumor ablation by appropriately tuning the electrical parameters.
[1] B. Rubinsky, G. Onik, and P. Mikus, “Irreversible electroporation: a new ablation modality – Clinical implications”, Technology in Cancer Research and Treatment, Vol. 6, No. 1, pp. 37-48, Feb 2007.
[2] B. Rubinsky, “Irreversible electroporation in medicine,” Technology in Cancer Research and Treatment, Vol. 6, No. 4, pp. 255-259, August 2007.
Eagle Harbor Technologies, Inc. (EHT) is developing a low-cost, fully solid-state architecture for the pulsed RF heating systems and diagnostics at fusion science experiments. EHT has constructed and tested a high voltage inductive adder to drive gyromagnetic and diode-based nonlinear transmission lines (NLTLs). The inductive adder is capable of 35 kV output with a 10-ns rise-time into 50 Ohm loads. During this program EHT has experimented with the development of diode-based lines as a method to produce high power pulsed RF at frequencies from 0.1 to 3 GHz. Here the concept is to utilize the system to demonstrate pulsed plasma heating with an inductive adder high power RF source. EHT will present results of a new diode based NLTL and show the design of the experimental setup for plasma heating.
At the Frankfurt Neutron Source (FRANZ), it is intended to generate a neutron beam with energies up to 300 keV by a pulsed proton beam with the Li7(p,n) reaction. To prebunch the proton beam for FRANZ, a 6 kV chopper with a repetition rate of 256 kHz and a pulse width of 100 ns has been designed and build. The current pulse generator bases on transformers and therefore forms a clipped and damped sine oscillation. The output amplitude is asymmetric between the positive and negative electrode and the voltage conversion depends upon frequency and primary voltage. As a sine half wave has no plateau, the protons occupy a larger phase space range, which could be cropped by the RFQ acceptance. In order to achieve a higher accelerator transmission, a square pulse would be desirable. A transformer less pulse generator would also generate a more predictable output pulse amplitude, due to the hard coupling of input and output voltage. The recent progress in the field of semiconductor technology made it possible to construct a MOSFET based generator. We present the status of a new designed pulse generator using market available semiconductors for pulsing the FRANZ chopper with a rectangle pulse.
Several prototype inductive adders (IAs), for use with kicker systems, each with very different specifications, are under construction at CERN. Historically pulse generators for kicker systems use thyratrons and either a Pulse Forming Network (PFN) or a Pulse Forming Line (PFL). The IA has several advantages compared to these conventional pulse generators, including: the use of semiconducting switches instead of thyratrons; the modularity, scalability and ease of including redundancy; and the possibility to actively reduce the ripple to improve pulse quality. In addition, semiconductor switches that can both turn-on and turn-off allow the PFN or PFL to be replaced with a capacitor bank. The IA is a very interesting option for older kicker systems, especially for those where spare PFL is no longer available: the kicker systems of the proton synchrotron (PS) at CERN were built in the 1970s and are thyratron based PFLs. Due to increased need for maintenance and difficulties to source various components, the replacement of the PS kicker pulse generators by a modern technology is a very attractive option. Hence studies have started to design and build a prototype IA for use in various kicker systems in the PS. Thanks to its modular design the design of a prototype IA developed for the Future Circulator Collider (FCC) injection system can be readily modified to reach the required output voltage and impedance values for the PS systems. In particular the IA is being considered for three PS kicker systems: the proton extraction kicker systems (KFA4 and KFA71/79) and the proton injection kicker system (KFA45) with a pulse length of 2.1 and 2.4 µs respectively. This paper gives a short introduction to IA technology and discusses the challenges for an IA prototype to meet the PS pulse requirements.
The advanced beam-driven C2-W field-reversed configuration experiment [1] utilizes two, merging compact toroid’s, formed and accelerated by two theta-pinch formation systems. Two new pulse power systems (Bias and MR), configured in parallel were designed and built to drive current in 34 (17 pairs) formation coils. The total stored energy in 38 Bias power units is about 1.4 MJ and provide the initial current rise time of about 100 us. The second circuit of the pulse power system – Main Reversal (MR) – consists of 136 power units which generate current pulse with the rise time about 5 us. The total stored energy in the MR power units is ~650kJ. Each of these pulsed power units utilize both a proven technology, using existing pulsed power capacitors and ignitron switches and a relatively new pseudo spark switch technology. The system design and characteristics as well as initial results of its test and full operation will be presented.
[1] M.W. Binderbauer et al., AIP Conf. Proc. 1721, 030003 (2016).
Diversified Technologies, Inc. has developed a PEF (Pulsed Electric Field) system for processing fruits and vegetables by softening their tissue at the cellular level. This process, which applies short (microsecond scale), high voltage pulses to the product in a water bath / slurry, makes slicing, dicing, peeling, drying, and juicing easier, and saves substantial energy in these downstream processes. PEF processing can be applied to non-thermal pasteurization of liquids, such as juices, or tissue modification of fruits and vegetables, which lowers their processing cost by actually making more of the cell contents accessible. Utilizing microsecond 1-5 kV/cm high voltage pulses to perforate cell membranes, this system can prepare tons-per-hour of whole fruits and vegetables for downstream processing.
The PEF system consists of a switching power supply, capacitor, solid state hard switch, and controls, all packaged in a NEMA-4 enclosure for operation in a food processing plant. The PEF system operates at 15 kV, and up to 100 kW average power. DTI has built numerous PEF systems, for different applications and specifications, since 2000.
This paper will describe PEF System design, applications, and trade-offs in the specification and construction of PEF system.
Nonlinear transmission lines (NLTLs) have been used with great success to generate high power radio frequency (RF). Generally, their operation consists of a lumped line based on the nonlinear behavior of the LC section components, capacitors or inductors, as a function of the applied voltage or current, respectively. However, considering high power signals, the employment of ceramic capacitors in capacitive lumped lines is restricted to frequencies around 80 MHz, since at high voltages their parasitic impedances limit the NTLTL maximum operation frequency. On the other hand, the use of low-voltage variable capacitance diodes has enabled the operation of NLTLs at higher frequencies (250 MHz). In addition, with the advent of high-voltage silicon carbide (SiC) Schottky diodes, it is expected that NLTLs can operate at higher power and frequencies. This paper presents the construction of a nonlinear transmission line based on SiC Schottky diodes to generate RF at high-frequency. Its working principle is presented, where the voltage dependence of the diode capacitance is modeled. The experimental and simulation results were also compared and discussed. Generation of oscillations at a frequency of about 200 MHz was obtained.
*Work supported by US Air Force Office of Scientific Research under contract no. FA9550-18-1-0111.
RF measurements, using previously published techniques, were made to determine the transverse impedance of a pillbox. The pillbox had near realistic physical dimensions of a cell being designed for a conceptual next generation multi-pulse linear accelerator. Details of the measurement techniques as well as comparison to analytical and simulated results will be presented.
UHVDC transmission system with hierarchical connection mode is a new trend. In this mode, the inverter station adopts the way of layered access to 1000 kV and 500 kV AC power grid, which can realize the optimization of UHVDC power transmission, improve the voltage support ability of the receiving end AC system and guide the reasonable distribution of power flow.
Insulation coordination is a fundamental problem to be solved when building UHVDC transmission system. The requirement is to achieve an acceptable insulation failure rate with minimal insulation investment. The key issue is to determine the voltage level taken during the withstand voltage test of the insulation structure. Thus it is necessary to study the generation and distribution of over-voltage of UHVDC transmission system with hierarchical connection mode.
In this paper, an electronic transient model of ultra-high voltage DC transmission system with hierarchical connection mode is set up by EMTDC/PSCAD simulation software based on ±1100 kV Zhundong-East China project. Basing on the simulation platform, this paper studies the mechanism and distribution of switching over-voltage caused by typical faults, such as grounding fault of AC bus, opening fault of neutral line and opening fault of metallic return line. The over-voltage suppression measures are proposed for protection start delay and fast ground switch operation delay according to the simulation results obtained. The research results provides a method of optimizing the configuration of arresters. The conclusions of this paper can be applied for the insulation coordination and fault analysis of the following UHVDC transmission project with hierarchical connection mode.
A fast travelling wave kicker operating with 80 MHz repetition rates is required for the new PIP-II accelerator at Fermilab. We present a technique to drive simultaneously four series-connected enhancement mode GaN-on-silicon power transistors by means of microwave photonics techniques. These four transistors are arranged into a high voltage and high repetition rate switch. Using multiple transistors in series is required to share switching losses. Using photonic signal distribution system is required to achieve precise synchronization between transistors. We demonstrate 600 V arbitrary pulse generation into a 200 Ohm load with 2 ns rise/fall time. The arbitrary pulse widths can be adjusted from 4 ns to essentially DC.
Spatially dispersive nonlinear transmission lines (NLTLs) have attracted interest as frequency tunable, wide-band, high power microwave (HPM) sources. The characteristics of these sources need to be further evaluated and understood to optimize their design. This paper presents the experimental performance of a spatially dispersive NLTL composed of Nickel-Zinc (NiZn) ferrites possessing a range of loss tangents, initial permeabilities, and dimensions. The NLTL’s performance is presented for each ferrite across a range of operational input voltage levels and ferrite bias conditions. Results show output frequency tunable from 0.81 to 1.39 GHz and peak RF powers in excess of 100 MW. In addition, higher peak powers were observed for ferrites with higher initial permeabilites versus lower loss tangents for the same dimensions. In addition to the experimental evaluation, a COMSOL model of the NLTL has been developed as a further method of gaining insight into the transient operation of this NLTL system architecture.
In the framework of collaborations between the SIAME Laboratory of Pau University and the French Alternatives Energies and Atomic Energy Commission (CEA) on Marcoule Center, we are working on the improvement of an electrostatic precipitator (ESP) for gas cleaning.
The background of the present work was established during the past ten years. Several successive studies allowed improving the treatment efficiency of electrostatic precipitation (ESP). Improvements are based on the replacement of the classic wire active electrode by an optimized emissive conception and the use of a combined voltage (HV pulses superimposed on DC voltage).
The presented paper is dedicated to the last new results due to a novel geometric arrangement. We are going to detail the evolution of treatment efficiency in our reduced scale ESP as a function of 2 parameters:
-The first one is electrical one and based on the voltage type (DC or hybrid), the level and the repetition rate.
- The second one is based on size particles.
The behaviour of filtration efficiency is analysed over long operating times. The development of back corona, which affects the efficiency, acts first of all on the fine particles. This influence of particle size on filtration efficiency is more pronounced under direct voltage than under pulsed/direct voltage.
Antibiotic resistance mechanisms render current antibiotics ineffective, necessitating higher concentrations of existing drugs, the development of new drugs, or the discovery of a method or mechanism to counter them [1]. Combining nanosecond electric pulses with tobramycin and rifampicin enhanced the cytotoxic effect for gram negative Escherichia coli, gram positive Staphylococcus aureus, and a Methicillin Resistant strain of Staphylococcus aureus. Specifically, we observed that individually applying 1000 pulses of 300 ns duration for a 20 kV/cm electric field or clinical doses between 2 and 20 ug/mL of tobramycin did not induce appreciable cell death; however, combining the pulse parameters and tobramycin resulted in more than a 3-log inactivation. Over a range of pulse and drug parameters, we observed a consistent 3-log synergy, which increased to 7-log reduction when combining tobramycin and rifampicin. This indicates the great advantage of combining electric pulses with a single drug, but indicates the dramatic improvement that can arise from combining multiple antibiotics [2] with electric pulses. This synergy arose on timescales of 15 minutes, compared to timescales of hours to days for antibiotics alone to achieve similar bacterial kill off.
[1] S. Reardon, “Antibiotic resistance sweeping developing world: bacteria are increasingly dodging extermination as drug availability outpaces regulation,” Nature, vol. 509, no. 7499, pp. 141–143, 2014.
[2] L. P. Kotra, J. Haddad, and S. Mobashery, “Aminoglycosides: perspectives on mechanisms of action and resistance and strategies to counter resistance,” Antimicrob. Agents Chemother., vol. 44, no. 12, pp. 3249–3256, 2000.
Recently, improvements of drainage water treatment technology has been required. The streamer discharges generated by nanosecond pulsed power was applied to treat surfactant solution as a persistent chemical. Radicals with a high redox potential, such as especially hydroxyl radical, is expected for degradative treatment of the persistent chemical. The effects on production of hydroxyl radical and the treatment of the surfactant solution was studied, by changing the ozone concentration in the treatment reactor when oxygen gas was used as a surrounding gas. The treatment experiment of anionic surfactant was carried out with a coaxial cylindrical reactor by changing the flow rate of oxygen gas. The voltage pulses of 1.5-ns pulse-width were applied on a solution of 1 L of purified water dissolved with 0.1 g of sodium dodecyl benzene sulfonate (DBS) as an anionic surfactant for 80 minutes. The concentration of anionic surfactant was evaluated with a methylene blue visual colorimetric method. The concentration of the surfactant decreased with treatment time. The absorbance at 225 nm, where was a peak absorbance of DBS, decreased with treatment time, and then, treatment of the anionic surfactant was confirmed with the two methods. On the other hand, the absorbance at 260 nm of the solution increased with treatment time. Because ozone in water had a peak absorbance at 260 nm, it implied that ozone was dissolved in the solution and hydroxyl radical was produced as a result. The ozone concentration ejected from the treatment reactor decreased with increase of flow rate of oxygen gas. On the other hand, treatment speed of surfactant was not significantly changed with the flow rate. The electrical conductivity significantly increased from 26.3 μS/cm to 126.1 μS/cm in the treatment. The change in conductivity might affect the generation of streamer discharges on the solution surface.
Partial discharge (PD) is a common and detrimental phenomenon that can cause damage, and eventual breakdown for insulation systems. Currently there are several standards for PD detection and measurement at atmospheric pressure, with power frequency (up to 400 Hz) excitation voltage. In a gaseous medium, such as air, the partial discharge inception voltage (PDIV) will generally decrease with decreasing pressure, at power frequencies. However, similar standards for PD measurement under impulse excitation are not well established. The lack of literature on PDIV behavior under high dv/dt impulse voltage excitation, combined with the increasing usage of inverter-fed-drive motors in aircraft, makes the topic very relevant.
This paper presents experimental results and associated analysis for several samples of aircraft motor-related components in air at various pressures, either with 60 Hz ac voltage or with impulse voltage excitation. In the initial testing, PDIV values for both 60 Hz and fixed-risetime impulse excitation decreased with decreasing pressure, which is expected. However for some tests, especially at relatively low pressure conditions, PDIV values have been observed to show an abnormal trend. Several statements on appropriate testing procedures for PD testing under low pressure conditions will be put forth according to the test experiences.
Due to the limitation of the test samples and equipment, the effect of the impulse voltage rise rate (dv/dt) was not established. However, there will be additional discussion on research in progress to reproduce realistic voltage impulses similar to those seen by aircraft motors in inverter-fed-drive applications; this will enable further investigations on the effect of impulse voltage rise rate.
As the power demands increase for airplanes, heavy duty electric vehicles and data centers, their dc bus voltages are expected to increase. While this reduces cable weight and improves system efficiency, electric arcs become an increasingly dangerous fault condition. Arcs fall under two categories—in series or in parallel with a load. Parallel arcs release large amounts of energy as high and low voltage electrodes are almost shorted, whereas series arcs act as an inserted impedance. Both parallel and series arcs can cause electromagnetic interference, insulation damage and fires. Parallel arcs cause large load current surges, which can be detected and handled with overcurrent protection. Series arcs decrease the load current, making them more difficult to detect than their parallel arc counterpart. Various detection methods and series arc models have been studied in the past. Although some models reflect the underlying phenomena, few explain how the occurrence and severity of series arcs are linked with load and source impedance of the circuit. Based on a large group of test data with RC loads and a fixed loop inductance, this paper derives a generic waveform and formalizes a two-stage energy mechanism to describe transients of series dc arc. These provide analytical insights into the effects of load capacitance. It is found that load capacitance plays a major role in arc severity, limiting the rate at which arc voltage rises, which is a critical factor on whether a sustained arc can be formed. The findings in this paper can be extended to constant power loads and guide future designs of loads with series arc prevention in dc networks.
It has been ascertained already, without doubts, that feeding rotating machines by power electronics drives can cause accelerated degradation and premature breakdown of electrical insulation. This is due to various factors, but the predominant one is the inception of partial discharges, PD, which are not expected to occur in Type I (organic) insulation, or the abnormal growth of PD (in amplitude and/or repetition rate) in Type II (mixed organic/inorganic) insulation. IEC standards, specifically 60034-18 41 and 60034-18-42 describe the proper approach to perform design qualification and type tests on Type I and Type II insulation, respectively. One of the most delicate issue of these standards is that, for reasons of simplicity, cost and availability on the market of adequate voltage impulse generators, PD and life tests can be carried out by both voltage repetitive impulses and sinusoidal voltage supply, with the only care that the test peak to peak voltage is the same whatever the chosen waveform (thus, implicitly, the peak to peak voltage is considered the predominant aging factor).
This paper has the purpose to speculate about how the type of waveform can affect PD features and life performance, considering not only repetitive impulse and sinusoidal waveforms, but also the real waveforms applied to rotating machine terminals by two, three and five level inverters. The effect of number of inverter levels on life and on PD pattern will be highlighted, and the relevant standards discussed.
In recent years, with the development of higher voltage and larger capacity of electrical equipment, the improvements of electrical insulation performance for various materials are urgent. The present investigations indicate that the characteristics of material surface, such as adhesion, hydrophilicity, charge accumulation, hardness, can be enhanced by surface modification including direct fluorination, nano-doping and low temperature plasma treatment. Among them, low temperature plasma surface modification takes many advantages, such as environment-friendly, controllability and flexible structure. In our group, pulsed plasma was used for polymer and metal surface modification to inhibit micro-discharge in high voltage insulation system. Different plasma discharge devices, such as dielectric barrier discharge (DBD), atmospheric-pressure plasma jet (APPJ) or diffuse discharge were developed for surface modification. Both SiOx, TiO2 and SiOx-TiO2 composite thin film with nanostructures were deposited on polymer and metal surface. The film thickness, composition and morphology can be easily controlled by changing pulsed plasma parameters. The experiments can be divided into two parts: 1) By introducing SiOx on polymer surface, the surface charge dissipation rate was accelerated and the surface flashover voltage was increased. Further studies showed that SiOx film with Si-OH group intruded shallow traps on the material surface. 2) The SiOx and TiO2 thin film covered on copper surface can decrease the electric filed distortion and inhibit micro-discharge. Furthermore, the deposited film was used for metal particle activity suppression in DC GIL. The results showed that, after film deposition (any of them), the lifting voltage of metal particle and the lifting time delay were enhanced. The lift voltage and time delay enhancement degree were TiO2-SiOx composite≈TiO2>SiOx. In summary, the pulsed plasma surface modification technology developed by our group has been proven a feasible method for high voltage insulation performance improvement.
Pulsed plasma jet with the advantages of compact size, low coat, abundant reactive species and flexibility have been a hot spot in recent years. The application of pulsed plasma including surface modification, medical application and environmental protection, etc. Especially, by grouping individual plasma jets together, large area and abnormal shape subjects can be easily treated. In this study, a coaxial designed 3-D multiple plasma jet array driven by nanosecond pulse power supple are well studied. The plasma jet array is composed with six individual plasma jets coaxial arranged and focused toward the center, which is ideal for cylindrical subject treatment. The electrical characteristics, optical emissions, spatial-temporal evolution and interaction of plasma jet array with and without subject are investigated. With applied voltage of 8 kV, only one plasma jet is ignited and all of the plasma jets are ignited when applied voltage increased to 12 kV. Further increase of voltage turns to arc discharge. The dynamic of ionization front of plasma jet array behaves differently with and without cylindrical subject: The ionization front moves in a faster speed on metal subject and seems to be more disperse on insulating surface. Plasma bullet reflection effect is observed when touching with the metal surface. More details will be shown and discussed in the paper.
This work is supported by National Natural Science Foundation of China（51507169,11611530681, National Basic Research Program of China (973 Program)（2014CB239505-3）
Guan-Jun Zhang, Run-Dong Zhou, Guang-Yu Sun, Bao-Hong Guo, Bai-Peng Song, Zheng-Shi Chang, Hai-Bao Mu
State Key Laboratory of Electrical Insulation and Power Equipment,
School of Electrical Engineering, Xi'an Jiaotong University, Xi'an, Shaanxi 710049, China
Surface flashover across solid insulation in vacuum is a great limitation of many electrical and electronic systems, since it typically takes place on the surface region of insulating material at applied electric stress much lower than the bulk breakdown strength of the material. This paper experimentally studies the initiation and development process of flashover across solid insulation under pulsed voltage in vacuum. It is observed that discharge initiation is greatly affected by local electric field at both cathode triple junction (CTJ) and anode triple junction (ATJ). The flashover phenomena transfer from cathode-initiated to anode-initiated as the electric field at ATJ is increased, i.e., flashover is generally initiated from CTJ, and when the field at ATJ is too much stronger than that at CTJ, the flashover would be initiated from ATJ. However, the ATJ discharge is restricted and hardly directly develops toward cathode, and the final breakdown happens only when CTJ discharge is triggered by the ATJ discharge. The flashover characteristic results suggests that the withstand voltage of anode-initiated flashover is higher than that of cathode-initiated flashover. However, a too strong field at ATJ would lead to the decreasing of flashover voltage. A competitive model is proposed to describe the relationship between the flashover initiation and development and the different electric field configuration. It is practically useful to greatly improve insulation withstand performance by optimizing local electric field, i.e., appropriately increasing the ATJ field and reducing the CTJ field to make flashover occur in a critical state between the cathode-initiated and anode-initiated discharge.
The plasma electrons bombarding a plasma-facing wall surface can induce secondary electron emission (SEE) from the wall. A strong SEE can enhance the power losses by reducing the wall sheath potential and thereby increasing the electron flux from the plasma to the wall. The use of the materials with surface roughness and the engineered materials with surface architecture is known to reduce the effective SEE by trapping the secondary electrons. In this work, we demonstrate a 65% reduction of SEE yield using a velvet material consisting of high aspect ratio carbon fibers. The measurements of SEE yield for different velvet samples using the electron beam in vacuum demonstrate the dependence of the SEE yield on the fiber length and the packing density, which is strongly affected by the alignment of long velvet fibers with respect to the electron beam impinging on the velvet sample.
The limitations of bond wire performance under pulsed conditions are of particular interest in the development of high current Photoconductive Semiconductor Switches (PCSS) for pulsed power applications. Literature regarding the limitations of bonded wires under DC (steady state) conditions are readily available. However, the current carrying capability of bond wires under pulsed conditions has not been studied extensively, especially for the case of the very small wires with less than 6.4x10-4 cm2 typically used for semiconductor development.
The current handling capabilities of gold cylindrical wires with radius 0.0254 mm (1 mil) and Aluminum ribbon wires with rectangular cross-section measuring 0.0254 mm x 0.2540 mm (1 mil x 10 mil) are investigated under pulsed conditions. Pulsed widths range from 8 µs to 24 µs with current peak densities covering 2.5 x 106 A/cm2 up to 6.2 x 106 A/cm2. The wires were bonded using a TPT HP05 manual wire bonder and two bonding techniques: ball bonding for cylindrical wire consisting of ultrasonic plus heat and wedge bonding for the ribbon wire consisting of only ultrasonic forces. Two failure modes are characterized: bulk failure and bond failure. Bond failure result in the wire disconnecting at the bond location. Alternatively, a bulk failure typically occurs along the wire itself.
Post mortem SEM imaging was used to determine if the wire was vaporized from excessive ohmic heating or if mechanical stresses played a role in the failure. High speed imaging was also used to capture the time progression of the failures. The integral of current action (ICA) is used to analytically predict the wire current limit for a specific excitation current. These theoretical predictions are compared with the experimental results.
This research was supported by the Air Force Research Laboratory (AFOSR) under contract number FA9550-14-1-0019.
Numerical Studies into the Possibility of ″Lock-On″ in a GaN Photoconductive Switch for High Power Applications
Animesh R. Chowdhury and Ravindra P. Joshi
Department of Electrical and Computer Engineering
Texas Tech University, Lubbock, TX 79409.
“Lock-On” was reported for GaAs and InP Photoconductive Semiconductor Switches (PCSS) over two decades ago [1]. More recently, focus has shifted to the GaN wide bandgap material for potential PCSS applications. Given that GaN is a direct bandgap material like GaAs, it can also be expected to support Lock-On. Potential benefits include an ability to withstand larger voltages and power, very low dark currents, and reduced heating due to larger thermal conductivity. However, Lock-On in GaN has not been observed to date. Here, a fundamental assessment of the potential for Lock-On in GaN is reported based on physics-based numerical modeling.
A one-dimensional time-dependent model is developed and discussed. Field-dependent carrier velocities have been taken from the literature [2]. A drift-diffusion model is used to implement mobile carrier transport. Trapping-detrapping dynamics are included on the basis of rate equations, with trap-to-band impact ionization being an important and separate process. As detailed information regarding traps (their densities and energy levels), and impact ionization rates is not yet fully available, various possibilities are probed and will be reported. Our contribution also includes investigations of hole injection, photon recycling, and field dependent multiphonon trap emission to predict the potential ″Lock-On″ scenarios. For completeness, we will also discuss Lock-On in GaAs based on the model developed with suitable comparisons to experimental data.
References:
1. F. J. Zutavern G. M. Loubriel M. W. O’Malley L. P. Shanwald W. D. Helgerson D. L. McLaughlin, IEEE Transactions on Electron Devices 37, pp. 2472-2477, (1990).
2. U. V. Bhapkar and M. S. Shur, Journal of Applied Physics 82, pp. 1649-1655, (1997).
We report on the development of a 10-kV SiC MOSFET-based switching module, which can be easily scaled-up because of the principle of a serial connection of the SiC MOSFETs. In detail, the switching module consists of two circuit boards, each one featuring five Wolfspeed 1200-V C2M0080120D SiC MOSFETs, that are getting stacked in order to minimize the area of the current loop between high voltage and ground connector and as a result the inductance of the structure. The gate control of each SiC MOSFET is connected to a separate high-speed gate driver circuit, which is being triggered optically. The DC voltage supply of the components of each of the gate driver circuit is battery powered. As an overvoltage protection for the SiC MOSFETs and the optical receivers, the switching module uses transient voltage suppression (TVS) diodes. The turn-on time was experimentally determined to 21 ns at a switching voltage of 10 kV and a drain current of 38.5 A. A first successful test in burst-mode with a burst of 10 cycles at a pulse repetition frequency of 5 MHz was carried out. Planned usage of gate-boosting techniques for a further speeding-up of the turn-on behavior will be discussed.
The controlled turn-less structure motor (CTM) described in other papers of this conference requires arrays of high current low voltage inverters occupying a width of ~6mm and distributed around the periphery of a motor.
The recently developed class of semiconductor switches which operate at the 10 to 50Volts range and planar current density in the 3-4kA/cm2 range both in Si and GaN, opens the opportunity to integrate these devices directly with electro-mechanical systems so that each stator conductor in the gap is separately switched controlled in an H-bridge 3- phase manner. Lastly, the integration of high power density batteries and micro channel coolers in a compact configuration results in an overall enhanced specific power and performance of motors, generators and actuators.
Separately, the low voltage allows proximities of semiconductor dies on their original wafers, which eliminates the need of individual packaging and allows arrays of high power H-bridges to be integrated with their loads more compactly and economically.
We will discuss the underlying semiconductors characteristics and limitations of both Si and GaN MOSFETs when operating in the low voltage high current regime, and the overall power density limits associated with dense packaging and enhanced cooling using micro channel coolers.
Furthermore, we will discuss the various manufacturing choices for such a motor and tradeoffs. While the present devices are prototypes, and thus their construction is not representative of mass manufacturing, due to the planar topology of the windings, MEMS manufacturing can be used here for etching the conductor grooves and depositing insulator layers and electroplating the copper conductors to form the turn-less structures in one operation.
In the microsecond and sub-microsecond pulsed power area, compared with other traditional gas switches, the IGBT switch has the advantage of short recovery time, high reliability and high power density, etc. The soldered IGBT has high parasitic inductance and failure problem due to the wire bonding and single-side heat dissipation, while the new type of press-pack IGBT could make an effective improvement. In order to compare the performance of IGBTs in pulsed power applications with those two types of packaging styles precisely, taken the packaging parameters of the commercial devices as a reference, the finite element simulated models for the soldered and press-pack modules with the same current rate of 1200A are established, separately. The parasitic inductance abstracting results show that the collector and emitter inductance of the press-pack IGBT are only 2.1% and 14.5% of that of the soldered one. Because the soldered IGBT has more parallel branches, they have shown similar current rise rate. On the repetitive frequency operation modes, the soldered IGBT gets lower average temperature by much larger heat sink, but the press-pack IGBT has higher power density because of the compact structure. In order to verify the characteristics difference of the two types of IGBT applied in the pulsed power applications, the pulsed power platform based on the press-pack IGBT and the soldered IGBT is established. Through the experimental measurement for the pulse waveforms and temperature distributions, the simulated results have been proved to a certain extent. In general, the press-pack IGBT has lower packaging parasitic inductance, double-sided heat dissipation, higher power density, easier to be pressed into stacks and short circuit failure mode, etc., which makes it have more advantages in the pulsed power applications. The discussion and analysis based on the models of commercial devices give us the information in detail and quantitatively.
High power photoconductive switch (PCSS) is an attractive device in pulse power systems. For the wide bandgap (3.26 eV), high critical field strength (3 MV/cm), and high thermal conductivity [4.9 W/(cm∙K)], semi-insulating 4H-SiC was considered an ideal material for PCSS. The on-state switch resistance should be as lower as better for high power output, but only about 1 Ohm on-state resistance was achieved so far, which may due to the low carrier mobility of SiC. So, it is important to check the limit of SiC material as substrate for PCSSs. This work provided a concise method for measurement of transient resistance of semi-insulating SiC under laser pulse.
The vertical PCSS with a transparent Al doped ZnO anode, semi-insulating 4H-SiC substrate, and high reflection silver cathode was introduced. A circular aperture was used to change the area of laser illuminating to the transparent anode. Under a 2kV bias and the 532nm laser with power density of 18.2 MW/cm^2, the on-sate resistance of the PCSS using different area of aperture was tested. The on-sate resistance was made of the resistance of SiC under laser excitation (Ron_SiC), and the inductive impedance (RL) in the switch and test circuit. The expression is given as follows, Ron_test = Ron_SiC + RL = ρH/S+ RL [Formula (1)], where ρ is the resistivity of SiC under laser excitation, H is the thickness of SiC substrate, S is the area of the aperture.
The measurement data and the fitting result by Formula (1) show a good linear relationship between Ron-test and 1/S. The resistivity of SiC excitated by 18.2 MW/cm^2 laser was 0.96 Ω∙cm by linear fitting, result in a resistance of 0.34 Ω for 1mm thick SiC under a circular aperture with diameter of 6mm. So, sub-Ohm on-state resistance is possible for SiC PCSS.
Come in and watch a movie. Snacks will be provided.
Abstract: This short course focuses on the basics of power semiconductor devices and its performance under normal and extreme condition. The course will also include the advent of silicon carbide (SiC) in the area of high temperature and high power semiconductor devices and its comparison to silicon (Si) counterpart in terms of device performance under extreme operating condition especially in pulsed power applications. The course will begin with a brief overview of material parameters, physics of power semiconductor devices and blocking voltage capability. Power devices including P-I-N and JBS diodes, Power Metal Oxide Semiconductor Field Effect Transistor (MOSFET), Gate Turn off (GTO) device, comparison of silicon to silicon carbide power devices and power device modeling and simulation will also be discussed during the course.
Instructor: Dr. Stephen B. Bayne received his PhD, MS and BS degrees in Electrical Engineering from Texas Tech University. After completing his Doctoral studies, he joined the Naval Research Lab (NRL) where he was an electronics engineer designing advanced power electronics systems for space power applications. After two and a half years at NRL, Dr. Bayne transferred to the Army Research Lab (ARL) where he was instrumental in developing a high temperature power electronic and pulsed power program. After 8 years at the ARL, Dr. Bayne transitioned over to academia where he is a Full Professor at Texas Tech University. His research interests at Texas Tech are Power Electronics, Power Semiconductor Devices, Pulse Power and Renewable Energy. Dr. Bayne is currently conducting research with a focus on SiC and GaN power semiconductors for pulse and continuous operation.