Changing Simulink Resolution from Seconds to Minutes causes problems

So I have a thermodynamic cycle that I am modeling. I wanted to set it up so that the run time was in minutes rather than seconds, so I just set gains in front of every integral block and controller (the integrals were taking in kJ/s, so multiplying by 60s makes them kJ/min. With this the model runs much faster, but it also has picked up some 'noise' in the signals. As they reach steady state they do not just flatten out as they did in the seconds resolution version--they oscillate back and forth (on a small scale, but noticeably). Is this a numerical problem caused by changing the effective resolution of the model?? Any help would be appreciated. Thanks,

Sarah

Answer 1

[Doug Eastman]

Hi Sarah,

Without seeing the model it is difficult to know exactly what the issue is, but in general changing the unit from seconds to minutes should not introduce instabilities like you are seeing. The problem may have to do with how you converted the model. What you probably want to do is go through the model and convert all the block parameters that have time units into minutes (for example, add a '*60' in the mask if it was /s initially). That should do it. You don't need to put gains in front of the integrators and controllers in that case.