False instability in RMS-type power system simulators with fast current sources or sinks: A case study

Power system dynamic simulation by the 'RMS' (root mean square) methodology greatly simplifies the computational effort required, by retaining a passive AC phasor representation of network voltages and currents, and updating the phasor currents according to the internal dynamics of connected plant. Owing to its high efficiency and scalability it is extensively used for routine stability assessment. Unfortunately, the assumptions underlying RMS simulation break down when the network contains 'fast current sources' such as DC-AC inverters, which are capable of aligning their AC current with a changing AC voltage on time scales shorter than an AC cycle. By means of a very simple case study, it is shown how and why an RMS simulator may misreport a stable system as unstable when fast current sources are present. Specifically, it is demonstrated that the rate of convergence in simulation software such as PSS/E can depend critically on the plant's reactive power operating point, in a somewhat unexpected manner. It is suggested that in response to this 'false instability', practitioners should not abandon the RMS simulation methodology, but should recognise its limitations and where necessary, seek alternative evidence on a plant-specific basis to support study conclusions.