Abstract
A modeling approach to computer simulation of power electronic systems that combines the advantages of the modified nodal analysis (MNA) method and state space analysis techniques is proposed. A power electronic system is split into a linear main circuit and nonlinear modules which are modeled by appropriate nonlinear time-varying voltage source and/or current sources. A state space model of minimum complexity is then generated using MNA, while avoiding the graph theory approach. A power electronic system can be simulated by iteratively solving the state space model thus formulated together with nonlinear modules. Moreover, the effect of switching actions on the coefficient matrix of circuit equations to be solved can be eliminated by the proposed approach. The approach is found to be compatible with the SPICE data structure. It can be used to develop powerful multilevel simulators based on SPICE or to improve SPICE capabilities for power electronic system simulation.<>

This publication has 5 references indexed in Scilit: