Sensitivity to simulation method for power electronic circuits

A host of simulation tools are available for the analysis of power system transients. Independent of the particular tool employed, excellent accuracy and simulation speed can be expected when linear circuits are analyzed. Although power electronic devices are nonlinear, in the realm of power system simulation, individual switches, diodes and thyristors are typically modelled as piecewise linear components which are either in a state of conduction or nonconduction. Consequently, it is more appropriate to consider these devices as linear elements which change the topology of a circuit as they switch from periods of conduction to nonconduction or vice versa. If other electrical components within the system are linear, then the entire system including power electronics may be considered linear, but time varying. Taking this perspective, it is easily seen that the ability of a simulation tool to accurately and efficiently model a power electronic circuit does not depend on its ability to accurately model the circuit behaviour during periods of consistent device conduction or nonconduction. During such periods the circuit is fully linear and nearly any solution method is adequate. Accuracy and efficiency of a solution technique is therefore measured only by a method´s ability to correctly identify the precise times at which a power electronic device switches between conduction and nonconduction intervals and to appropriately re-initialize all system state variables