Old multi harmonic code working again + some additions.
Some performance issues are still there:
- size of stack in Cal_WholeQuantity is to be modified if the number of harmonics is higher than 40 (NBR_MAX_HARMONIC);
- assemble is very slow if -prealloc is not properly set;
- for being able of increasing the size of the stack a bit more, I've added MAX_STACK_SIZE0 and fix its value to 2 => no tensors by default in the MH case;
- real version of PetSC must still be used => feature not available in the pre-packaged version!
Working cases:
- (nonlinear) ferromagnetic core with three windings (compumag2001, ieeetrmag 2002);
- permanent magnet synchronous machine (icem 2002, compel 2003)
- 3-phase 5-limb transformer with and without rectification (epnc 2002);
- induction machine (cefc2002, ieeetrmag 2003)
- (nonlinear) homogenisation of windings (cefc 2016, ieeetrmag 2017, compumag 2017)