Health Analysis Method of Batch Smart Electricity Meter Based on K-Nearest Neighbor Algorithm

Abstract

During the operation period of smart electricity meters, there is great uncertainty for the health index of smart electricity meters in batches. In order to better evaluate the health state of electric energy meters, this paper proposes a method for health analysis of electric energy meters based on the KNN(K-Nearest Neighbor) algorithm and structural equation model. To be specific, the key indicators are selected and the evaluation system is constructed through the KNN algorithm firstly. Then, the influence relationship and weight value between each key index are analyzed and determined by utilizing the structural equation model. Finally, the health of the smart meter is analyzed and predicted based on the weighted average method. This method has realized the transformation from traditional manual diagnosis to intelligent analysis and judgment based on machine learning. The empirical analysis shows that the proposed method could reduce the influence of human subjective factors in the health evaluation model, distinguishing the health degree of different batches accurately, and provide a convenient and accurate evaluation method for the front-line staff.