Choose timezone
Your profile timezone:
Cold test results discusson
→
Europe/Zurich
Powering results (Vincenzo)
- Introduction
- Test plan: reduced ramp rate, 1A/s does not come from requirements
- Training: similar to CCTs, very visible precursor always on rope 8-9-10 (top of the groove)
- This indicates a displacement of the conductor in this area but without knowing the exact length that moves, it is difficult to give an estimation of it.
Action: It could help to have a displacement model from FEA.
Quench propagation (Vincenzo)
- The quench propagation always starts with a base slope (that depends on current) and then increase by an integer value of it with a maximum of 7 times.
- This indicates propagation between wires in a rope. It happens in 30 ms
- The localization of the wire that quench can deduced from the steps of increase.
- No rope-to-rope propagation is observed within 150 ms (correlating with FQPC results).
Varistor (Vincenzo)
- 45 kJ are going to the varistor during a quench at nominal (limit 150 kJ)
- However, it takes a long time to cool down and subsequent quenches decreased the performance of the device.
Splices (Vincenzo)
- Measured at 1.2 nOhm with very small dispersion.
- As expected from test on samples and subscale 2
Acoustic sensors (Vincenzo)
- First trials of this new type of sensors on an actual magnet.
- They work fine and give signal in sync with the precursors not the quenches
Action: FFT and check if frequency is changing with field
Magnetic measurements (Mariano)
- Introduction
- Meas. Procedures (decay for 18 extractions, 120s delay to get all eddy current dumping)
- Good agreement with simulation at the center but lower ramp along the s equivalent to 40 mm deviation.
- Radial at center is a bit away from simulation with more field components
- Field Integral as fine in the center but off on the sides
Action: see if it differs a lot with current. As B is linear with current this would indicate if it goes with forces.
- Transfer funct. is also not trivial to understand : global shape with 60 units increase and oscillations of 16 units amplitudes.
Question: Is transfer function changing during training?
Action: provide a description of the fix point of the magnetometer to check contraction of the magnetometer toward former
Optical fibers (Laura)
- Distributed fibers co-wound with magnet, to use with Rayleigh back scaterring effect
- 4 fibers initially, 2 lost during impregnation, 1 after,… 1 left
- 100m long, 5Hz, 2.6 gage length
- 0 at room T, record cool down, no reset at 4K
- There is an obvious pattern repeating in sync with the winding turns
Action: correlate with deformation model on the conductor
Action: check local effect during inner former quench heater tests (about 20 s long).
Quench simulation (Mariusz)
- Low current, 72A, ok
- 288A, 4.2K, al good, matches eddy current coupling and quench back effect (includes thermal contact between wires, ropes, and former)
- 1.9K, quench back is late. It can be fit if changing thermal connections but then it does not work at 4K. It can be fit by tuning the amount of helium, but this is not fair. Same if we introduce more exotic effect as Kapitza, it is too tuneable and can fit everything.
Comment: it seems the model overestimate the first drop (generation of eddy current?). Can this help to fit the quench back part of the decay ?
However, this drop is correct at 4K and should not be impacted by T.
There are minutes attached to this event.
Show them.
