Mercurial > repos > public > sbplib
view +noname/testCfl.m @ 1037:2d7ba44340d0 feature/burgers1d
Pass scheme specific parameters as cell array. This will enabale constructDiffOps to be more general. In addition, allow for schemes returning function handles as diffOps, which is currently how non-linear schemes such as Burgers1d are implemented.
author | Vidar Stiernström <vidar.stiernstrom@it.uu.se> |
---|---|
date | Fri, 18 Jan 2019 09:02:02 +0100 |
parents | 7f6f04bfc007 |
children |
line wrap: on
line source
% noname.testCfl(discr, timestepper_method, T, alpha0, tol,threshold, silentFlag) % Example: % noname.testCfl(Discr(100,4), 'rk4', 1, [0, 1]) function testCfl(discr, timestepper_method, T, alpha0, tol,threshold, silentFlag) default_arg('tol',0.00005); default_arg('threshold',1e2); default_arg('silentFlag', false); % TODO: % Set threshold from the initial conditions of the pde? % Take a set number of steps instead of evolving to a certain time? % Stop evolving when it has blown up? testAlpha = getAlphaTester(discr, T, threshold, silentFlag, timestepper_method); % Make sure that the upper bound is not working ok = testAlpha(alpha0(2)); if ok % Upper bound too large! error('The upper bound on alpha is stable!') end % Make sure that the lower bound is ok if alpha0(1) ~= 0 ok = testAlpha(alpha0(1)); if ~ok error('The lower bound on alpha is unstable!'); end end if silentFlag rsInterval = util.ReplaceableString(''); end % Use bisection to find sharp estimate while( (alpha0(2)-alpha0(1))/alpha0(1) > tol) alpha = mean(alpha0); if ~silentFlag fprintf('[%.3e,%.3e]: ', alpha0(1), alpha0(2)); else rsInterval.update('[%.3e,%.3e]: ', alpha0(1), alpha0(2)); end [ok, n_step, maxVal] = testAlpha(alpha); if ok alpha0(1) = alpha; stability = 'STABLE'; else alpha0(2) = alpha; stability = 'UNSTABLE'; end if ~silentFlag fprintf('a = %.3e, n_step=%d %8s max = %.2e\n', alpha, n_step, stability, maxVal); end end if silentFlag rsInterval = util.ReplaceableString(''); end fprintf('T = %-3d dof = %-4d order = %d: clf = %.4e\n',T, discr.size(), discr.order, alpha0(1)); end function f = getAlphaTester(discr, T, threshold, silentFlag, timestepper_method) % Returns true if cfl was ok function [ok, n_step, maxVal] = testAlpha(alpha) ts = discr.getTimestepper(struct('method', timestepper_method, 'cfl', alpha)); warning('off','all') ts.evolve(T,true); warning('on','all') [v,t] = ts.getV(); maxVal = max(v); if isnan(maxVal) || maxVal == Inf || abs(maxVal) > threshold ok = false; else ok = true; end n_step = ts.n; end f = @testAlpha; end