Setting number of iterations, time step, and convergence criteria

[jundonglin]

Hello, how to set the number of iterations, iteration step length and convergence criteria, I can not find the relevant methods to set these on the official website

[jundonglin]

I found the example in the script file, I think you should give a guide on the official website, the location of the instance is a little hidden for beginners.

[jundonglin]

The code version in the example is too old, and I still can't find the corresponding method for the solver setup

[ma-sadeghi]

I assume you're referring to the number of iterations in the nonlinear solver, if that's the case, here's a code snippet from OpenPNM that explains it:

OpenPNM/openpnm/algorithms/_reactive_transport.py

Lines 21 to 46 in 5e2404c

	class ReactiveTransportSettings:
	r"""
	Parameters
	----------
	%(TransportSettings.parameters)s
	sources : list
	List of source terms that have been added
	relaxation_factor : float (default = 1.0)
	A relaxation factor to control under-relaxation for the quantity
	solving for. Factor approaching 0 leads to improved stability but
	slower simulation. Factor approaching 1 gives fast simulation but
	may be unstable.
	newton_maxiter : int
	Maximum number of iterations allowed for the nonlinear solver to
	converge.
	f_rtol : float
	Relative tolerance for the solution residual
	x_rtol : float
	Relative tolerance for the solution vector
	"""
	relaxation_factor = 1.0
	newton_maxiter = 5000
	f_rtol = 1e-6
	x_rtol = 1e-6

Basically, once you create your algorithm (ReactiveTransport for instance), you can modify these parameters by accessing the algorithm's settings attribute:

alg.settings.newton_maxiter = 100

[ma-sadeghi]

Upon re-reading the title, it seems that you might be referring to the transient solver? If so, find the corresponding Settings class, and you should be able to find relevant parameters there.

[jundonglin]

I found it, thank you!