Replace x_dot with x_prev

Main reason is the numerical difference between the actual previous time step solution and the computed one as x - \dot{x}\cdot\Delta t. The difference is negligible for smooth functions but not for if-conditions or Heaviside functions. Actual error case is THM-Freezing where the freezing point might be missed due to the error; the code relies on the x^{t-1} being exactly equal to the solution of previous time step.

For review: Changes are organized in three parts: NumLib and common ProcessLib, then processes with implementation changes, finally the processes with just renames. The changes in the processes are separated in rename-only part and actual FEM implementation changes. I suggest commit-wise review.

1. Feature description was added to the changelog
2. Tests covering your feature were added? Hopefully the existing tests do cover all changes.
3. Any new feature or behavior change was documented?