| The main source of imprecision, in the fixed step and real time simulation of power electronic circuits, comes from the imprecision of the switching in the circuit. These imprecision are of two types. First of all, they come from the switching delay, when the switching does not coincide with a calculation time step. Secondly, they come from an inaccurate evaluation of the initial conditions or the operating point of the circuit, after a switching. Also, the handling of the switching requires, from the simulation, an significant calculation effort since the set of equations of the circuit has to be reevaluated, in order to reflect the new state of the switches.; Prior works have proposed a simple and effective approach to eliminate the impact of the switching delay. This method involves the interpolation of the simulated signals at the precise switching instant and to restart the simulation from this point. Unfortunately, these works did not propose an effective method to evaluate the required initial conditions.; The switching method, proposed here, takes advantage of the interpolation method and adds to it a simple, yet effective, method to evaluate the initial conditions. The new operating point of the circuit, following a switching, is evaluated on the basis of the conservation of energy principle and on the hypothesis that the switching is instantaneous. With that, the energy accumulated in the reactive components do not change during the switching and it can be used to determine the operating point of the circuit.; In addition to the switching method, this work proposes a few technics to optimize the handling of the switching. One of these technics proposes to replace the switches by controllable current sources and, hence, to transfer the burden of updating the set of equations to the reduced set of equations which models the switches.; The switching method is implemented in the two most popular simulation approaches, the nodal approach and the state space approach. Finally, the precision and performance of the method are evaluated using a few well chosen setups. Each of these setups were simulated with both simulation approaches and provided results which met the original objectives of this work. |
