基于残差的逆变器滤波电路故障检测与分离方法研究

Research on Fault Detection and Separation Method of Inverter Filter Circuit Based on Residual

随着清洁能源的不断普及,逆变器作为电能转换的中枢,其故障的诊断已成为研究的热点。逆变器故障诊断的难点之一是由于开关函数的非线性,导致逆变器难以精确建模。采用状态空间平均法来对逆变器建立无故障与故障下的模型,在两个模型中比较得到残差,通过建立残差评价函数体系,提出了残差信号斜率、残差信号斜率绝对值、残差信号范数及残差信号范数绝对值4个评判指标,对逆变器的故障进行故障分离并判断故障的类型。通过GeckoCIRCUITS搭建仿真电路,获得了滤波电路中电容与电感值降低时指标发生的变化规律并分析不同故障在指标下的差异性,结果表明对故障的检测与分离是可行的。

With the continuous popularization of clean energy,inverter as the power conversion center,its fault diagnosis has become a research hotspot.One of the difficulties of inverter fault diagnosis is that it is difficult to accurately model the inverter because of the nonlinearity of switching function.Using state space average method to establish a trouble-free inverter and fault model,the comparison of the two models get residuals,by establishing a residual evaluation function system,the residual signal slope,slope residual signal absolute value,residual signal norm and residual signal norm absolute value of the four evaluation indexes were put forward,the fault of inverter was separated and the type of fault was judged.The simulation circuit was built by GeckoCIRCUITS,the change law of the index was obtained when the capacitance and inductance of the filter circuit decrease,and the difference of different faults under the index was analyzed.The results show that the method is feasible for fault detection and separation.