Numerical Error Analysis in a PWM-Driven Real-Time Simulator: a Buck Converter Case Study

Real-Time Simulators (RTS) are a growing alternative to mitigate the inherent challenges of design, prototyping, and fault detection of power electronics circuits driven by PWM signals. These challenges are common to electromobility applications, such as EV chargers and onboard electrical systems. Despite the widespread use of RTS, critical issues remain unaddressed in the literature, particularly concerning the accuracy of these models and the timing mismatches they encounter. This research seeks to help bridge this gap by modeling a buck converter, a simple yet fundamental converter, and developing an RTS based on its model. Through experimental validation of the RTS and exhaustive simulations of equivalent models, the study evaluated the performance parameters of the RTS, including the execution period, timing drift, and duty cycles, while measuring the errors arising from modeling inaccuracies and time-shift discrepancies. The analysis discerns low-from high-frequency errors, with findings indicating a close relationship between model error and duty cycle, suggesting that analytical compensation within the model is possible. Furthermore, while oversampling and latency were found to have minimal impact on low-frequency errors, high-frequency errors were influenced by latency, although with negligible consequence. In general, this study offers insight into improving PWM-driven RTS, highlighting the imperative of improving the accuracy of modeling to address low-frequency errors in power electronics applications, with benefits extending across industries, including electromobility.