How to avoid numerical oscillations in power system with synchronous machines?

Good morning!

My question is apparently very simple. I would like to run a model of a power system which includes synchronous machines in steady-state and obtain constant generators' output frequencies. Since I am interested in running some tests simulating faults, I would like to start off with frequencies as constant as possible.

This can be well achieved if I run a Phasor simulation using the Powergui block, but it cannot be obtained in either a continous or discrete simulation (also Discrete Phasor simulation). I have read this article and asked for suggestions in order to limit oscillations, but all given solutions seem not to be working in my case.

At the moment, I have already tried:

• Increasing the time step of the model (from few tens of us up to 0.5 ms)
• Adding a resistive load to the terminals of each machine whose power is proportional to the time step (as suggested in the model). I have also tried adding a resistive load much bigger than the minimum one obtained from the proportionality with the time step.
• Changing the type of discrete simulation (from ode1 to higher orders)
• Changing the type of solver model for the synchronous machines (using the trapezioidal and Euler "robust" versions)

At the moment I get an oscillating frequency around 60 Hz whose peaks reach up to 60.04 and 59.96 which is clearly not ok for a steady-state case.

How can I improve this result having already tried (apparently) everything suggested? Alternatively, can I place my power system in one block and simulating it using Phasor simulation, and locate my sensor models in another block using a discrete simulation? Could that work?

Thanks a lot for your help.

[PAUL SERNA TORRE]

I am having the same problems. I tried all the methods you suggested, but the numerical oscillations are still there. The only way (I found) to get around this issue is using the simplified synchronous machine.

[PAUL SERNA TORRE]

New update: I was able to run a three-bus network with three synchronous machine pu standard without any frequency oscillation. What I did is to add a very small mechanical power to the one that the power flow determined in one machine. This very small mechanical power I found by trial and error until the frequency settle at 60 Hz for 2 seconds.