Why is simulation significantly faster when using lookup table with filtered output?

[matija]

I have a battery model which is described using a lookup table as in the attached models. If there is a low pass filter added to the output of the 2D lookup table simulation speed is significantly improved, and without it, it’s much slower. How do you explain this? If you add another time constant to a model simulation should be more complicated and slower, so how does the filtering output of the lookup table influence simulation speed? In the given models there is a simplified lookup table that describes the battery cell voltage as a linear function of the state of charge and current V(SoC, I), but when I put real data of a battery the result in speed simulation is the same regarding the filter.

with_filter.plecs (28.4 KB)

without_filter.plecs (27.6 KB)

[Kris_Eberle]

As per the simulation warning message in the model without the low pass filter (see the attached image), you have an algebraic loop in the system and this forces the PLECS solver to take a more iterative, computationally-intensive approach.

And the low pass filter simply injects the necessary delay to break the algebraic loop, and hence the model with it runs expectedly faster.