Choose timezone
Your profile timezone:
Conductor modelling - Update #1
→
Europe/Zurich
CERN
CERN
Participants:
- Baffari Dario
- Bordini Bernardo
- Bottura Luca
- Keijzer Ruben
- Succi Giovanni
- Willering Gerard
Giovanni presented an upgrade of the mono-dimensional model of the cable with a broken strand following input from Bernardo and Dario. The new version of the model is characterized by a different schematization of the Nb3Sn cable joint with the NbTi cable: before entering into the Nb3Sn cable the current flows in an equipotential superconducting cable and then in a purely resistive cable. The quasi-equipotential cable is obtained by setting the inter-strand contact resistance to values as close as possible to 0; the length and the resistivity of the resistive cable are set to obtain a resistance similar to the one generally obtained in NbTi/Nb3Sn Rutherford cables. Despite the parameters of the simplified joints were not optimal (in particular the resistance of the joint was 2/3 orders of magnitude lower) relevant new results were found:
- After the current ramp-up, during the plateau, the current redistributed from the damaged wire to the two adjacent strands starts to diffuse also in other strands. Bernardo pointed out that the amount of current redistributing in other strands and the diffusion time constant could be significantly affected by the resistance of the joint and by how much the equipotential boundary conditions prevail with respect to the condition of uniform current distribution between the cable strands in the joint. Dario proposed a sensitivity study to evaluate the influence of these aspects.
- A negative signal, which is often observed in damaged magnet coils, was observed across the defect in the strands adjacent to broken wire. Luca asked Gerard and Bernardo to prepare a page where we put down all the questions, rising from the different magnet test results, that we would like to answer with the models we are developing.
Ruben showed some preliminary results on the analysis of MQXF coils, highlighting that his results suggest that no negative voltage is possible across the defect because of the voltage trend along the cable strands. Ruben suggested that the difference with respect to Giovanni’s results may be due to the shorter cable length simulated by Giovanni. Bernardo pointed out that the differences might also be imputed to the different boundary conditions while Gerard commented that the two models were not incompatible since they refer to different cables length. Luca agreed that the boundary conditions might have a role in the overall solution for long cable but this needs to be further investigated.
Action points:
- Sensitivity study to assess how low the interstrand resistance has to be to have a quasi-equipotential boundary condition before the joint (Giovanni)
- Sensitivity study to study the influence of the inter-strand resistance on the current redistribution process (Ruben).
- Implementation of a hydraulic element in the THEA model, to reproduce the equivalent joint resistance of about 1 nOhm maintaining a temperature of 1.9K on the joint (Giovanni).
- Assessment of the influence of the cable length and the voltage taps location on the voltage signal across the defect (Giovanni).
- Assessment of the influence of the boundary condition on the long-cable simulation (Ruben).
- Expand to the model to include more strands (current model has 12 strands) (Ruben).
- FEM modelling of the joint to evaluate current redistribution phenomena to be transferred to the 1D model (Dario, Bernardo).
- Prepare a page with all the questions, rising from the different magnet test results, that we would like to answer with the models we are developing (Gerard, Bernardo).
There are minutes attached to this event.
Show them.
