Digital Twin Health Monitoring of Five-Level ANPC Power Converter based on Estimation of Semiconductor On-State Resistance

In this paper, a novel digital twin approach is introduced for the health monitoring of five-level active neutral point clamped (5L-ANPC) power converters, which have been applied in many safety-critical high-power applications. By establishing a real-time interactive digital replica (i.e., digital twin), various parameters are measured to monitor the inverter's health condition, enabling improved reliability and lower downtime cost. To address challenges posed by multiple series-connected semiconductor switches in the 5L-ANPC converter where different switches have different junction temperatures, the on-state resistance of multiple switching devices is estimated. This estimation is achieved by using particle swarm optimization (PSO) and sensed signals. The sampled data and the PSO cost function are modified to minimize estimation errors, improving the precision of health monitoring. The robustness and effectiveness of the proposed method are demonstrated by considering various switching devices. As the first-ever digital twin approach developed for the health monitoring of multilevel power converters, the proposed technique significantly enhances the reliability of multilevel converters for high-power safety-critical applications.

I. Introduction

Digital twin (DT) modeling is a concept that creates a virtual representation of a physical system on a digital platform (e.g., system microprocessors, cloud or edge data centers), replicating and predicting the performance of a physical system (i.e., physical twin). As shown in Fig. 1 using a power electronic converter as an example, this interactive virtual DT model is continuously updated with real-time data acquired from the corresponding real-world physical systems, so the DT model can provide high-fidelity health monitoring information to the real converter system.