Which input variables might TESPy alter during solving?

[JanKochanski]

Hi @fwitte,

we discussed recently that previous TESPy versions changed the UA-values during solving, is there anything like this for other variables as well? I just found that the heat output from my heatpump model alternates quite significantly, even though I provide a constant value of 22 MW as an input. You can see in the picture below that the Q_DH alternates a lot, with spikes up to 38MW (plotted behind eta_S).
Just for explanation of the other variables in the graph:

• Eta_S values below 0.8 are valid from my compressor map, if I am outside the compressor map I set eta_s to .85 to be able to solve but also distinguish invalid results (just for explanation).
• PR_Comp_HP is the pressure ratio over the compressor stage
• Comp_HP_Volumetric_flow is the volumetric flow rate over the compressor stage
• T_SW_warm is the heat source temperature

Please ignore the region of invalid results around time step 2000 - I need to adapt the compressor map input data.

I am currently using the most recent TESPy version from pip (v0.9.16.post2).

Hope you have some idea what could cause these fluctuations, thanks in advance for your feedback 🙂.
Best, Jan

[fwitte]

Do you track the status of your model in each simulation? Only nw.status==0 is a valid solution.

[fwitte]

So either it is a bug or an issue in your setup

[JanKochanski]

I am currently rerunning the dataset, and I am now tracking both nw.converged and nw.status. It seems that if converged I get status==0, If not converged it either status 2 or 99. I will recheck my setup! Thanks for the ultra quick reply!

[fwitte]

Try to sort the simulations in a way so the change of the inputs is minimal between two simulations, that can improve robustness. If you need we can meet another time

[JanKochanski]

Hi @fwitte ,
It seems that the problem came from that I calculated the mass flow from the heat demand to be able to use this as an input to one connection instead of two heat exchangers. After switching to Q as a direct input I get clean results. Thanks for your help!
Best, Jan