The University of British Columbia

State variable approach is often used for modelling power-electronic circuits at the system level with controls. In the approach considered herein, the minimal state-space representation of the overall system is generated from the circuit branch data and updated for each new topology of the system. The network partitioning into optimal sets of tree and link branches is achieved using the spanning tree algorithms. For system with variable parameters recomputing the entire state equation to reflect the parameter changes may be prohibitively expensive. In this paper, a computationally effective algorithm for updating the required terms based on topological information is presented.