add cfl check

[sbryngelson]

No description provided.

[Copilot]

Pull Request Overview

This PR adds comprehensive CFL (Courant-Friedrichs-Lewy) condition checking and adaptive time step control to the CFD simulation to improve numerical stability and prevent simulation failures.

• Implements CFL condition monitoring with both convective and viscous checks using per-node local mesh spacing
• Adds adaptive time step reduction when CFL thresholds are exceeded to maintain stability
• Introduces comprehensive solution validation to detect NaN/Inf values and velocity field collapse early

Reviewed Changes

Copilot reviewed 6 out of 8 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
src/visualize_final.m	Updates visualization to handle final velocity vector format and improves performance by storing computed statistics
src/stability_diagnostics.m	Adds new CFL checking functions and improves node classification with better error handling
src/setup_environment.m	Adds support for time-based simulation control as alternative to step-based approach
src/check_solution_validity.m	New comprehensive solution validator that checks for NaN/Inf values and velocity field collapse
simulate.m	Major updates adding adaptive time step control, CFL monitoring, and comprehensive instability detection
config.m	Adds configuration parameters for adaptive time step control and CFL monitoring features

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

The percentile calculation is incorrect. When prc is a vector (e.g., [90, 99]), k will also be a vector, but indexing cfl_sorted(k) with a vector of indices may not give the expected percentile values. The calculation should handle each percentile individually or use a proper percentile function.

Suggested change

	% percentile computation without toolbox
	cfl_sorted = sort(cfl);
	k = max(1, min(length(cfl_sorted), round((prc / 100) * length(cfl_sorted))));
	cfl_prc_vals = cfl_sorted(k);
	% Use MATLAB's prctile function for accurate percentile calculation
	cfl_prc_vals = prctile(cfl, prc);

[Copilot]

[nitpick] The variable bad_idx is initialized but may not be used if no instability is detected. Consider initializing it closer to where it's first needed or using a more meaningful default value like [] to indicate no bad index found.

Suggested change

bad_idx = 1;

[Copilot]

[nitpick] The error message concatenation using array notation with ... continuation is less readable than using sprintf or fprintf style formatting. Consider using error('SIMULATION FAILED: Complete velocity field collapse detected before final validation at step %d.\nAll U and V components are exactly zero. This indicates the solution was corrupted after the last time step.', j) for better maintainability.

Suggested change

	error(['SIMULATION FAILED: Complete velocity field collapse detected before final validation at step %d.\n' ...
	'All U and V components are exactly zero. This indicates the solution was corrupted after the last time step.'], j);
	elseif vel_magnitude_max_final == 0
	error(['SIMULATION FAILED: Velocity magnitude collapse detected before final validation at step %d.\n' ...
	'All |V| components are exactly zero. This indicates the solution was corrupted after the last time step.'], j);
	elseif u_max_final < 1e-15 && v_max_final < 1e-15
	error(['SIMULATION FAILED: Near-complete velocity collapse detected before final validation at step %d.\n' ...
	'All velocity components are below machine precision (< 1e-15). This indicates the solution was corrupted ' ...
	'after the last time step.'], j);
	error('SIMULATION FAILED: Complete velocity field collapse detected before final validation at step %d.\nAll U and V components are exactly zero. This indicates the solution was corrupted after the last time step.', j);
	elseif vel_magnitude_max_final == 0
	error('SIMULATION FAILED: Velocity magnitude collapse detected before final validation at step %d.\nAll |V| components are exactly zero. This indicates the solution was corrupted after the last time step.', j);
	elseif u_max_final < 1e-15 && v_max_final < 1e-15
	error('SIMULATION FAILED: Near-complete velocity collapse detected before final validation at step %d.\nAll velocity components are below machine precision (< 1e-15). This indicates the solution was corrupted after the last time step.', j);

[Copilot]

This sprintf call is problematic when prc and cfl_prc_vals are vectors. The format string 'P%d=%.3g ' expects scalar values, but the input is a matrix from concatenating reshaped vectors. This will likely produce unexpected output or an error. Consider using a loop to format each percentile-value pair individually.

Suggested change

	tag = sprintf('P%d=%.3g ', [reshape(prc, 1, []); reshape(cfl_prc_vals, 1, [])]);
	tag = strtrim(tag);
	tag = strtrim(strjoin(arrayfun(@(p, v) sprintf('P%d=%.3g', p, v), prc, cfl_prc_vals, 'UniformOutput', false), ' '));