Execution rate causes inestability

Question

Good morning,

I am currently working with an averaged model of a DAB in Typhoon Sim. The model works perfectly for an execution rate equal or inferior than 1e-6 (components execution rate <= 1e-6 works perfectly). The problem is that if I try to lower the execution rate it appears that the simulation of the systems turns unstable, causing either solver error (DAE Solver failed at time 1.3360000000e-02 to find a solution. Please make one of the following changes: (i) adjust solver settings: maximum simulation step, relative and/or absolute tolerance values, (ii) provide better initial conditions, or (iii) choose a different DAE solver.) or not desired results (most signals turn unstable and tend to +- inf). 

This particular problem occurs both with closed and open loop, so the control doesn't appear to be the issue.

I need to run the model on a 402 physical HIL, which is unable to run with such small execution rate and I must be able to run the model with a bigger execution rate. I get the same problem with HIL SCADA, if I try to compile the model with an execution rate < 1e-6 the systems turns unstable, otherwise seems to work fine (on vhil scada because as I said the 402 cannot handle this execution rate without CIO an AO warnings with this system).

I've tried to simplify a bit the model but I keep getting the same issue). If anyone has had similar issues or could help me fix the issue any help is appreciated. I'll attach some pictures of the current model that i'm trying tom simulate. 

Here's everything but the control, if the resolution is too bad to appreciate the images, let me know and I'll try to provide more information.

Thank you very much in advance.

Salutations,

 

Answer 1

Hi Genis,

My guess is that the averaged DAB becomes unstable when interacting with the rest of the circuit. As you know, system is expected to be less stable as the discretization step increases. Is your model unstable if you put fixed voltage sources at the input/output of the DAB?

Cheers

Comments

Hello Jovan,
 
Thanks for your quick answer, there is a constant volatge source both on the input and output of the circuit. The unstability appears even with this elements. Have you ever had similar issues. Any tips would be highly appreciated.

---

At that point I'm afraid it's a debugging effort. Try to capture the data as short as possible. You can probably see the tendency towards instability in only several execution rates. Be wary of fast time constants in the electrical part of your model - maybe your signal processing is too slow for the circuit dynamics.

---

I'd like to share how I solved the issue just so it can be also helpful for others. The problem was mainly the two current meshes you can see in the first image. This two meshes contain contolled sources that depend of each other and the interaction between them can become unstable.

I put just the current meshes separately in a new model while I was debugging and apparently if you raise the execution rate up to a certain point the interaction between them becomes unstable for some reason. I was able to simplify the model a bit, getting rid of the issue.

---

Are there unit delays between the circuit loops (I can't see it correctly on the image)? If so, it seems a classical instability problem similar to ITM couplings. If you are still curious about the issue's cause, you can try searching for interface algorithms for co-simulation (see the paper at the end of the comment).

Basically, there is an impedance rule of thumb to ensure stability.
You can compute equivalent resistances for a capacitor as Rcap=Ts/C, and for an inductor as Rind=L/Ts


ref:
Improve the Stability and the Accuracy of Power Hardware-in-the-Loop Simulation by Selecting Appropriate Interface Algorithms