Add aggregation and breakage to crystallization module

[jbreue16]

There are multiple branches with parts of the functionality that we want, i.e. sole aggregation, sole breakage, sole PBM and all of them combined. However, the code is not modular.
The respective branches are on @WFlynnZ fork of CADET.
We need to figure out a neat modularization and then incorporate the code from the aforementioned branches, which we will do on this branch here.

Todo:

• rebase on master after PR Crystallization reaction module #112 is merged
• unify PBM, aggregation and fragmentation codes
• Put all parts in the same _bin loop
  • add abstraction layer PBMKernel
  • put pbm, fragmentation and aggregation in the same bin loop
  • update Jacobian accordingly
  • Figure out a more efficient containerization of the parameters? (see ReconstructionParams and JacobianParams, which have too many/unused fields depending on which reconstruction is used)
  • use bitflags to determine the crystallization mode
• Update the test data:
  • Update cadet-verification to vary the reconstruction schme used in crystallization tests (currently only upwind is tested)
  • update corresponding json and h5 files to cadet-core test data
• add tests for all crystallization modes
• add documentation for aggregation and fragmentation
• Can we improve performance by storing repeatedly calculated values?

[WFlynnZ]

Things to consider when modularizing the current code @jbreue16 :

1. Aggregation, Breakage and PBM (Population Balance with Mass Balance) should be able to be solved independently, and combined with two or three together.
2. Because Aggregation and Breakage are mass-conserving, the mass balance equation (and also PBM) is not needed when solving them independently.
3. When solving the PBM, two additional components are added: solute concentration is the first component and the solubility is the last component. These two components are absent in the Aggregation and Breakage code. When changing code, need to make sure the residual and Jacobian are also adapted.

Possible combinations are:

1. Aggregation
2. Breakage
3. PBM
4. Aggregation, Breakage
5. Aggregation, PBM
6. Breakage, PBM
7. Aggregation, Breakage, PBM

Only 1, 2, and 4 don't have the two additional components.

[jbreue16]

@WFlynnZ I found that the Jacobian tests fail, if we set the growth_scheme_order to 2, 3 or 4, which also happens for the code on master .. do you have an idea why that is, or is there a bug in the Jacobian?

[WFlynnZ]

@jbreue16 I remember we troubleshooted this a while ago, if I recall correctly, there was an issue with the CSTR module, which was fixed? I checked the PBM Jacobians a number of times now,,, it should be hard to find more mistakes in these higher order schemes, maybe this is related to the tolerances? Some entries need large tolerances because the absolute values are big.