HIL Scada vs TyphoonSim

Question

I'd like to know why the same system, when simulated in VHIL, behaves differently in the HIL SCADA scope and in Typhoon SIM? Especially since both systems are literally identical!

Answer 1

It could be that your model contains some dynamics that are faster than the simulation step chosen for VHIL can track. TyphoonSim is an offline simulation tool, and it can adjust the step to capture all dynamics. On the other hand, VHIL (as an emulator of a HIL device) uses a fixed simulation step, and if you didn't set it properly, then you might experience different results. Anyway, you need to be a bit more specific in describing the difference, so that the cause can be identified.

Comments

For example, I test controllers for grid-connected inverters with a 100-us execution rate, and in typhoonSim it works well—low harmonic content—but when I run the same system on VHIL, the controller becomes unstable. And yet it’s literally the same system.

---

Still, there is not enough information to be certain about the cause. Do you have core coupling elements in the circuit? Can you share your model so someone can take a look?

---

My system does not use coupling elements. Unfortunately, I am unable to share the file. The system is a grid-connected inverter that uses a C function for grid synchronization via a Kalman filter and for adaptive control that regulates the injected current. The inverter uses internal modulation with an FSW of 10 kHz; all execution rates are configured with a Ts of 100 μs.
The filter is an LCL filter, where I use a step function to vary the grid inductance via a variable inductor.
In TyphoonSim, I use the robust trapezoidal DAE solver with a step size of 1e-6, and in VHIL, I use the bilinear solver with a step size of 1e-6.
When you say “configure correctly,” which part are you referring to?
And do you think that increasing the fixed step size in VHIL could solve the problem?

---

I see. That variable inductor could be the problem for VHIL/real-time simulation. It's under the hood implementation differs for these two different platforms. In TyphoonSim we can simply change inductance value, reinitialize solver and continue since it is variable step offline simulation. In real-time (as well as in VHIL that is emulator of the real device) variable inductor is implemented as controlled current source in order to make it possible to run in real-time. Hence, depending on parameters of the entire circuit where such variable inductor is involved, circuit can become unstable due to variable inductor.

I would suggest you to run one simulation with regular inductors with some value from your test case. If that works, problem is identified and you can start looking for solution.

One of them would be to reduce simulation step for VHIL. You mentioned that in TyphoonSim you are using max sim step 1e-6. But that is maximal simulation step allowed for solver. If needed solver can reduce it to order of picoseconds for example, because it is variable step solver. On the other hand, VHIL uses discrete step and you selected 1e-6. In general, smaller step should give you better results, so I would advise reducing simulation step for VHIL (real-time).