【目的】牵引整流器与牵引逆变器是列车牵引变流器的重要组成部分，牵引整流器主要由绝缘栅双极晶体管(IGBT)、谐振电路以及直流侧电容等部件组成，其故障频率高于牵引逆变器，且IGBT发生开路故障时不会引发大幅过电流或过电压等明显现象，诊断和保护相对较难。为此，本文提出了一种基于混合逻辑动态(MLD)模型与自适应阈值的牵引整流器IGBT开路故障诊断算法。【方法】首先，在阐述牵引整流器工作原理的基础上，建立了牵引整流器的MLD模型并估计网侧电流。然后，对单管IGBT开路故障时网侧电流实际值与估计值之间的残差进行了分析。最后，设计自适应阈值进行故障诊断，并依据自适应阈值的变化区分IGBT发生开路故障的位置，进而提出基于MLD模型与自适应阈值的牵引整流器IGBT开路故障诊断算法，实现IGBT的开路故障诊断与定位。【结果】基于Matlab/Simulink平台建立MLD模型并引入自适应阈值对所提算法的有效性进行仿真验证。仿真结果表明该故障诊断方法能够在几百微秒内检测到IGBT发生开路故障，并能够准确定位到开路IGBT，且存在外界干扰时，不会产生误诊断。【结论】基于MLD模型与自适应阈值的IGBT开路故障诊断算法能够快速检测并定位到故障IGBT。与传统MLD故障诊断算法相比，所提算法能够有效解决不同IGBT开路时的阈值选择问题，且故障诊断精度高，对外界干扰具有很强的鲁棒性。

[Objective] The traction rectifier and the traction inverter are crucial components of train traction converters. The traction rectifier primarily consists of insulated gate bipolar transistor (IGBT), resonant circuit, DC-side capacitor, and other components. Its fault frequency is higher than that of the traction inverter. An open-circuit fault in an IGBT does not result in significant overcurrent or overvoltage, making diagnosis and protection relatively challenging. Therefore, this paper proposes an open-circuit fault diagnosis algorithm for the traction rectifier IGBT based on a mixed logic dynamic (MLD) model and adaptive thresholds. [Methods] Based on the working principle of the traction rectifier, an MLD model of the traction rectifier was established, and the grid-side current was estimated. The residual difference between the actual and estimated values of the grid-side current during single IGBT open-circuit faults was then analyzed. Finally, adaptive thresholds were designed for fault diagnosis, and the location of the IGBT open-circuit fault was distinguished based on the changes in the adaptive threshold. This led to the development of an open-circuit fault diagnosis algorithm for the traction rectifier IGBT based on the MLD model and adaptive threshold, enabling the diagnosis and localization of IGBT open-circuit faults. [Results] The MLD model was established based on the Matlab/Simulink platform, and the adaptive thresholds were introduced to verify the effectiveness of the proposed algorithm. Simulation results indicated that this fault diagnosis method can detect the open circuit faults of the IGBT within several hundred microseconds, accurately locate the faulty IGBT, and avoid misdiagnosis even in the presence of external interference. [Conclusion] The open-circuit fault diagnosis algorithm for IGBTs based on the MLD model and adaptive thresholds can quickly detect and locate faulty IGBTs. Compared to traditional MLD fault diagnosis algorithms, this proposed method effectively addresses the threshold selection issue for different IGBT open-circuit faults, offers high diagnostic accuracy, and demonstrates strong robustness to external interference.

辽宁省交通科技项目(202318)