基于D-S证据融合算法的逆变器故障诊断方法研究

Research on Inverter Fault Diagnosis Method Based on D-S Evidence Fusion Algorithm

光伏逆变器是光伏发电系统的重要环节,针对存在大量不确定性因素的逆变器故障诊断问题,深入研究逆变器的故障表现特征及故障模式,对比分析诸如小波分析、证据理论、神经网络、多级向量机等智能诊断方法 ,提出将D-S证据融合方法应用于逆变器的故障诊断,解决了存在交叉数据、误判率较高和诊断速度慢的逆变器故障模式识别问题,并通过计算找出不和谐焦元,重新分配信度,解决证据体间可能出现的冲突问题。实例分析表明,基于D-S证据融合算法的逆变器故障诊断方法具有良好的准确性和灵敏性。

Photovoltaic inverter is an important segment of the photovoltaic power generation systems, for the inverter fault diagnosis problems that have many uncertainty factors, In-depth research fault performance characteristics and mode of inverter, analyzing contrastively intelligent diagnosis method such as wavelet analysis, evidence theory, neural network and multistage vector machine, a D-S evidence fusion method applies to fault diagnosis of inverter is proposed, to solve inverter fault mode recognition such as cross data, higher miscarriage rate and slow diagnosis speed, and through calculation to find out disharmonious focal elements, then redistribute the reliability and solve the problem of the possible conflict between evidences. The instance analysis shows that inverter fault diagnosis method based on D-S evidence fusion algorithm has good accuracy and sensitivity.