Risk Assessment of Distribution Network Operation Involving Fusion of Electrical Time Series Data and Image Data

Risk assessment can quantify distribution network (DN) operation risk and is a necessary part of DN security analysis. Recently, UAVs have been widely promoted in the power inspection and provide staff with abundant power inspection images. In this paper, we developed a risk assessment method for DN involving the fusion of electrical time series data and image data. First, the collected time series data from PMUs is pre-processing by using PCA; Then, the features of this data are extracted by K-Nearest Neighbors (KNN) and further calculate the values of the three developed risk indicators; Meanwhile, combine the Mask Region-based Convolutional Neural Network (Mask R-CNN) with the images from UAV to determine whether there is any hidden danger in DN; Finally, the final DN operation risk value is obtained by fusing the risk sub-indicators with the operation hazard identification results through weights.