Different simulation behavior if simulation is called from within function vs. from script

Hello,

I have a simulink model which uses a variety of parameters from the workspace. If I execute the exact same piece of code (initialization of parameters+ run simulation) from within a function vs. from a script the simulation results are (slightly) different which leads to the simulation running into a stopping condition if called from the script, but finishing smoothly if called from within the function.

The two different pieces of code are like this:

params_val = goodRuns_all(ind_choosen,:);

%% Version 1 from script
parameters = params_val;
init_script;

MP = setParams(MP, 1, parameters(8),  parameters(10),  parameters(11), parameters(12), parameters(13));
simulation_length = 30;
options = simset('SrcWorkspace','current');
warning('off','all')
simOut = sim('model_name', [],options);
warning('on','all')

%% Version 2 with function call
simOut_func = runSim(params_val);

%% Function declaration
function simOut = runSim(parameters)
init_script;


MP = setParams(MP, 1, parameters(8),  parameters(10),  parameters(11), parameters(12), parameters(13));
simulation_length = 30; 
options = simset('SrcWorkspace','current'); 
warning('off','all')
simOut = sim('model_name', [],options);
warning('on','all')

end

Also, if I first execute the code from the script and then (without a clear) run the function, also the simulation called from the function runs into a stopping condition whereas if I directly call the function it does not.

What could be the reason for the changed behavior?

[Steve Miller]

I think we would need to see the model in order to diagnose the problem.

--Steve

Answer 1

[Steve Miller]

I believe the 'SrcWorkspace','current' is different if you call from within a function or a script.

See this answer.

https://www.mathworks.com/matlabcentral/answers/94215-how-can-i-set-the-source-workspace-to-caller-when-using-a-the-sim-command-from-a-function-in-simul

--Steve

[Elsa Bunz]

Sorry for not responding and thanks a lot for your answer and time!

In the meantime I continued to search for the reason and actually the simulation results don't seem to be different but the difference comes from a delay before my stopping condition.

This is how that part (and the settings of the delay) of my model looks like:

The scope for the call from the script looks as follows:

whereas for the function call it is like this:

Therefore, I guess somehow the sampling time seems to be affected causing this effect? Something in my script call seems to influence this, because if I first run the script and then the function without a clear inbetween then also the function version runs into the stopping condition. I didn't find where exactly it comes from, but I just solved the problem by removing the delay.

However, if you have any hint on where this could come from I'd be very happy to hear it!