Mixed modeling approach for efficient simulation of single-phase PWM inverters

Abstract

This paper proposes a new modeling approach to obtain accurate and fast simulations of pulse width–modulated (PWM) inverters with minimal loss of information. The idea of this methodology combines the use of precise switched models and fast averaged models. Switched models are used during transient evolution, and averaged models on periodic steady-state (PSS) regimen . In this way, the switching behavior of an inverter can be obtained even in the context of a long-term simulation where switched models cannot be used because of the central processing unit (CPU) requirement. The novel modeling strategy implements an algorithm that, based on the evolution of some variables, detects the transient or periodic steady state condition and automatically performs the commutation between both models. Furthermore, the mentioned algorithm also estimates the averaged model parameters when the precise switched model is used. Consequently, it is not necessary to calculate the average parameters to create the model and accurate results are obtained regardless of the operating point. This paper also reports simulation experiments of a single-phase PWM boost inverter model to clarify the idea of the modeling strategy and shows the simulation time and accuracy advantages of the proposed modeling approach. The Modelica language was used to program the novel modeling strategy and build the boost inverter examples.