Thermal Balance Oriented Model Predictive Control for 100 kHz Large-Signal Parallel Inverters With Optimized Pulse Patterns

In sound navigation and ranging, the parallel inverters are required to drive electroacoustic transducers to emit 100 kHz acoustic waves. Optimized pulse patterns (OPPs) can achieve the wide bandwidth of the parallel inverters with high switching frequency utilization. However, the junction temperatures of parallel inverters by using OPPs are imbalanced at a low carrier ratio. The inductance mismatches exacerbate the reliability issue. Therefore, this article proposes a thermal balance oriented model predictive control (TBOMPC) for 100 kHz Large-Signal Parallel Inverters with OPPs. First, the principle and the reliability of OPPs are analyzed. Then, the mathematical model is constructed for the inductance parameters. The initial switching states of TBOMPC are designed to identify the inductance parameters. The proposed TBOMPC takes account of the inductance mismatches and redistributes the switching states of OPPs, which can effectively balance the junction temperatures of power devices at a low carrier ratio. Finally, the effectiveness of the proposed method is verified by the experimental results. The maximum junction temperature differences of TBOMPC are less than 1.5°C under inductance mismatches.