As the core device of power electronic equipment, IGBT (Insulated Gate Bipolar Transistor) is related to the reliability of the system, so its online health monitoring is particularly important. Due to the large amount of data and high sampling rate for IGBT online health monitoring, data transmission has become a major problem, the method of effectively reconstructing the original signal has become one of the current research hotspots. Based on the theory of compressed sensing, the acquisition of IGBT gate waveforms is studied in this paper. First, the theoretical principle of compressed sensing is expounded and analyzed, and the evaluation criteria for signal reconstruction performance are given. Then, from the theory and simulation, the sparseness of IGBT gate drive signals under different sparse bases and different measurement matrices are discussed and analyzed. Next, aiming at the shortcomings of the stagewise weak orthogonal matching pursuit algorithm, a SWOMP algorithm based on the backtracking strategy of different Sigmoid functions is proposed. Simulation and experimental results show that the improved SWOMP4 algorithm has the best reconstruction effect. Finally, the SWOMP4 algorithm is applied to the process of IGBT gate drive signal reconstruction.
This paper mainly collects and reconstructs the IGBT gate drive signal through the improved SWOMP4 algorithm. The designed algorithm has high reconstruction accuracy and can be used in the IGBT module online health status system.