基于快速过采样主成分分析法的光伏阵列故障诊断

针对光伏阵列出现的组件阴影遮挡、短路与断路等故障,提出一种基于快速过采样主成分分析(over-sampling principal component analysis,OS-PCA)算法的光伏阵列故障诊断方法,实现故障检测与故障识别。通过检测各组串电流信号,利用快速OS-PCA算法计算各组串异常度,从而检测出故障串;通过误差补偿对光伏阵列工程模型进行优化,并通过分析故障时阵列工作点状态来识别故障类型。实验表明,该故障诊断方法可有效诊断出多变环境下组件阴影遮挡、短路、断路等故障,此外该方法在计算量以及内存占用上具有较强优势,适用于大型光伏电站的实时监控。

Fault Diagnosis Method of PV Array Based on Fast OS-PCA

Aiming at various types of faults in photovoltaic(PV)array,such as partial shading,short circuit and open circuit,a fault diagnosis method of PV array based on fast OS-PCA(over-sampling principal component analysis)is proposed,able to realize fault detection and identification.By detecting the current signal of each PV string and implementing the fast OS-PCA algorithm,abnormality of each string is obtained so as to detect the string with fault.Furthermore,the PV engineering model is optimized with error compensation,and the fault type is identified by analyzing working point of PV array.Experimental results show that the proposed method can effectively diagnose shading,short circuit and open circuit of PV modules in changing environment.In addition,the method has strong advantage in computation and memory occupation,suitable for real-time monitoring of large PV plants.