Quench models (moved from #337)

In GitLab by @mkovari on May 25, 2017, 14:41

I have written a new coil quench model.

1. The fits to the integrated heat capacity and resistance functions have been replaced.
2. The resistance of the dump resistor is taken to be constant (I think) in the old model. In fact the resistance will increase as the resistor is heated. This gives a slight beneficial effect by allowing the initial value of the resistance to be reduced, thereby reducing the effective dump time constant without increasing the peak voltage to ground, which occurs at the beginning of the quench. An alternative approach is to adjust the resistance dynamically to keep the voltage constant. I have implemented both models as options:

  !+ad_vars  quench_model /exponential/ : switch for TF coil quench model:<UL>
  !+ad_varc                  <LI> = 'exponential' exponential quench with constant discharge resistor
  !+ad_varc                  <LI> = 'linear' quench with constant voltage</UL>
  !+ad_varc                    Only applies to REBCO magnet at present
  character(len=12) :: quench_model = 'exponential'

3. I have implemented a delay time in the model for quench detection

  !+ad_vars  quench_detection_ef /0.0/ : Electric field at which TF quench is detected and discharge begins (V/m)
  real(kind(1.0D0)) :: quench_detection_ef = 0D0
  !+ad_vars  time1 : Time at which TF quench is detected (s)
  real(kind(1.0D0)) :: time1 = 0D0

    if(quench_detection_ef>1d-10)then
        ! Two-phase quench model is used.
        ! Phase 1
        ! Value of resistivity at which detection voltage is reached
        etamax = quench_detection_ef * croco_cable%copper_fraction / jwptf
        Check the temperature at which this resistivity is reached....

2-phase_quench_model.docx

3. Only the TF coil is considered, not the CS.

   I notice that there is a database of the thermo-physical properties of solid materials at cryogenic temperature at
   https://supermagnet.sourceforge.io/solids.html, including references and Fortran code.

I have implemented a new integration approach with new materials data, in subroutine quench_integral and superconductors.f90.

At the moment this is only effective for the REBCO option, but we should rewrite the other options to use these routines.

@jmorris-uk Could you have a look through the revised and expanded quench model document attached above? Thanks
@jmorris-uk Could you have a look at the code in superconductors.f90, and the relevant parts of global_variables.f90 and sctfcoil.f90?