Add new solver options to phasor dynamics application

[Steven-Roberts]

Description

I think these options will be needed for the fixed versus adaptive time step comparisons.

Proposed changes

This adds JSON options for tolerances and a fixed time step

Checklist

• All tests pass.
• Code compiles cleanly with flags -Wall -Wpedantic -Wconversion -Wextra.
• The new code follows GridKit™ style guidelines.
• There are unit tests for the new code.
• The new code is documented.
• The feature branch is rebased with respect to the target branch.
• I have updated CHANGELOG.md to reflect the changes in this PR. If this is a minor PR that is part of a larger fix already included in the file, state so.

Further comments

I noticed the existing dt JSON option has a few issues. It is documented as solver time step but it seems to be the output/monitoring time step. Perhaps it should be rename monitor_dt and the new fixed_dt should be called dt. Also with this rounding, the monitoring is not guaranteed to occur every dt simulation time.

GridKit/application/PhasorDynamics/DynamicSimulation.cpp

Line 48 in c099ac0

int nout = static_cast<int>(std::round((event.time - curr_time) / dt));

Any suggestions on how to handle this @pelesh?

[pelesh]

Yes, dt is output monitoring step for adaptive step solvers.

[pelesh]

Also with this rounding, the monitoring is not guaranteed to occur every dt simulation time.

I believe original idea was to have something like this:

for (i = 0; i < nout; ++i) {
  ...
  t += dt;
}

It was originally intended for nout to be an approximate number of outputs and output steps to be precisely of size dt.

[pelesh]

I like the idea of dt being the fixed time step and dt_moinitor to be output time step. Even if we use fixed time step, which is typically a quarter of a cycle, we probably don't want to output values at each step but say every 10th or so.

CC @lukelowry @abirchfield

[lukelowry]

(related note) For validation, we need dt to be a quarter cycle so we have comparable samples against PowerWorld.

The large cases produce a lot of monitored output, @pelesh is right in his suggestion that dt_monitor ~10x the fixed time step dt for practical use. These are the same in existing solvers, so the distinction is usually moot. But we should make the distinction here

I like the naming dt_monitor, good suggestion. To prevent confusion, would dt_fixed be clearer than dt?