Abnormity Detection Method of Intelligent Electricity Consumption for Nontechnical Loss
Nontechnical loss has been a significant factor influencing the profits of electric power companies and the power quality.A method of nontechnical loss detection based on customer types is proposed in this paper,including offline parameter optimization and online detection.Once the training data are determined,the real time abnormal behaviors would be tested out.With the updating of historical data,the detection accuracy gets higher till the optimal point.In this paper,support vector machine is adopted to detect the nontechnical loss and genetic algorithm is used to derive the optimal parameters for different users to improve user's nontechnical loss detection accuracy.Finally,in the case study,the real-time detection system assesses the classification accuracy and fault detection accuracy,verifying the stability of this method in periodic updating and advantages in reducing the cost of electric power companies.
nontechnical lossreal time detectionparameter optimization
NETL
NSTL
万方数据
维普
