Set Convergence Parameters for Adaptive Grid method

[andrew-octopus]

My thoughts:

• number of times without change in any grid value: 6, so intuitively there is something like a 1/2^6 chance we are settling to the wrong value.
• number of times without change in threshold_inequalities satisfied: 6, for consistency.

I want to avoid a situation at this stage where we keep getting near-misses.