基于瞬时幅值的光伏系统电流传感器微小故障检测及估计

电流传感器是光伏系统中用于系统控制和状态监测的重要元件,然而受运行环境影响,电流传感器易出现性能退化,影响系统运行安全.为了准确检测和估计出电流传感器微小故障,提出基于瞬时幅值的传感器微小故障检测和估计方法.首先,建立基于瞬时幅值的电流传感器微小故障模型,利用Hilbert变换(HT)估计相电流瞬时幅值将测量的三相正弦电流转换为相互独立的三维直流信号分量;其次,利用快速移动窗主成分分析(FWMPCA)对三维直流信号组成的数据矩阵进行特征提取,获得主元和残差子空间向量的概率密度分布函数;然后,利用Kullback-Leibler(KL)距离定量度量实际运行数据相对于无故障运行数据的微小变化,在此基础上,设置故障检测阈值,构建故障幅值估计模型,实现电流传感器微小故障检测和估计;最后,利用RT-LAB实验平台验证所提方法的有效性.

Incipient fault diagnosis and estimation for current senors of PV system based on instantaneous amplitude

Current sensors are the important component for system control and state monitoring in PV systems. However, due to the influence of operation environment, the current sensor performance degradation often occurs, which affects the operation safety of the system. To detect and estimate the current sensor incipient faults, a novel incipient fault detection and estimation method based on instantaneous amplitude is proposed. Firstly, the incipient fault model of a current sensor based on instantaneous amplitude is established, and the Hilbert transform (HT) algorithm is utilized to estimate instantaneous amplitudes of the three-phase currents, which makes the sinusoidal current signal transformed into the three-dimensional direct current signal. Then the fast moving window principle component analysis (FMWPCA) is used to extract the features of the data matrix composed by the three-dimensional direct current signal, and the probability density distribution functions of principal components and residual subspace vectors are obtained. Subsequently, the Kullback-Leibler (KL) divergence is used to quantitatively measure the small change of actual operation data distribution relative to fault free operation data. Based on above, the fault detection threshold is set and the theoretical model of fault amplitude estimation is constructed to realize the incipient fault detection and estimation of current sensors. Finally, the effectiveness of the proposed method is verified by the RT-LAB platform.