基于MLD模型的神经网络三电平逆变器开路故障诊断研究

发布时间：2018-01-22 16:39

本文关键词： NPC三电平逆变器 MLD模型神经网络故障诊断　出处：《新疆大学》2017年硕士论文　论文类型：学位论文

【摘要】：本文对处于风电系统并网核心部位的并网逆变器进行研究分析,从三电平逆变器结构拓扑及其工作原理出发,以二极管钳位型(NPC)三电平逆变器为对象,深入分析各桥路IGBT导通逻辑与输出端电压、电流之间的相互关系。针对同一相上桥臂(下桥臂)处于中间部位的IGBT单管故障和上桥臂(下桥臂)两个IGBT同时故障时两种故障模式的输出电压、电流波形相同情况;同时针对驱动失效或器件自身断路两种原因导致的开路故障输出特征相同情况,提出一种对IGBT故障进行准确定位并判定导致其故障发生原因的方法。构建三电平逆变器的MLD(Mixed Logical Dynamic)模型,并且在Simulink中,对其进行有效性验证。为了能准确诊断出故障并定位失效驱动部位,对逆变器的MLD模型表达式进行等效变换,得到逆变器的DMLD(Diagnosis Mixed Logical Dynamic)模型,使其正常工作时输出电压电流与仿真相同,故障情况下输出与仿真不同,用来有效定位失效脉冲。其次,以DMLD模型和实际仿真模型二者输出的线电流残差为故障信号源,以残差最大值Max、最小值Min、Add=Max+Min和App=Max-Min作为IGBT触发脉冲失效的特征表述。利用FFT(Fast Fourier Transformation)分析钳位二极管两端钳位电压的谐波含量,以1次、2次、4次和6次谐波含量作为器件断路诊断的有效特征。并将两种特征重新组合,生成新的同时包含触发失效状态和器件状态的特征集。以三层BP神经网络,识别故障特征集,结果表明本文提出的故障诊断办法真实有效。
[Abstract]:In this paper, the grid-connected inverter in the core of wind power system is studied and analyzed. Based on the structure topology and working principle of three-level inverter, the diode clamped NPC-three-level inverter is taken as the object. The IGBT conduction logic and output voltage of each bridge are analyzed in depth. The relationship between current. The output voltage of two fault modes when the upper arm (lower arm) is in the middle part of IGBT fault and the upper arm (lower arm) of two IGBT faults simultaneously. The current waveform is the same; At the same time, the output characteristics of the open-circuit fault caused by the driving failure or the device's own broken circuit are the same. This paper presents a method to locate IGBT fault accurately and determine the cause of the fault. The MLD(Mixed Logical dynamic of Three-level Inverter is constructed. Model. In order to accurately diagnose the fault and locate the failure driving position, the MLD model expression of the inverter is equivalent transformed. The DMLD(Diagnosis Mixed Logical dynamic model of the inverter is obtained, and the output voltage and current of the inverter are the same as those of the simulation. The fault output is different from the simulation, which is used to locate the failure pulse effectively. Secondly, the line current residuals of the DMLD model and the actual simulation model are used as the fault signal source, and the maximum residual Max is used as the fault signal source. Minimum Min. Add=Max Min and App=Max-Min are used as characteristic expressions of IGBT trigger pulse failure. The harmonic content of clamping voltage at both ends of clamping diode is analyzed by Fast Fourier transformation. The harmonic content of the first and second times and the sixth harmonic are taken as the effective characteristics of the device open circuit diagnosis, and the two characteristics are recombined. A new feature set including trigger failure state and device state is generated and the fault feature set is identified by using three-layer BP neural network. The results show that the fault diagnosis method proposed in this paper is true and effective.
【学位授予单位】：新疆大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM464

【参考文献】