Line Loss Assessment Method Based on Scene Clustering Method
Bai Tai, Jiaju Wang, Chen Liu, Ran Zhang, Xue Li-si, Xiaolu Sun · 2023
In order to minimize unreasonable power loss and make effective and reasonable use of electric energy, it is essential to effectively evaluate the line loss state of the transformer area and provide guidance for the subsequent line loss control. This paper introduces a method of grid line loss assessment based on scene clustering, which evaluates the state of grid line loss through the probability density function after scene clustering. Firstly, the multi-dimensional monitoring data is obtained by preprocessing the data collected by the monitoring device. Then, in order to conduct scene clustering for the power grid, this paper uses 3D K-means clustering method to cluster the power data, and divide it to obtain multiple line-loss state scenes. The data with the highest correlation with the line loss rate is obtained as the label of maximum correlation of line loss in this situation by calculating the mutual information between the line loss rate data and other monitoring data in various circumstances. The multiplier approach is then used to solve the model after fitting several probability models to create a general line loss probability model. Ultimately, the power grid's general line loss probability model is developed under various situations, allowing for the evaluation of the line loss condition and serving as a guide for managing the line loss moving forward. A province's measured data are analyzed and evaluated in order to determine whether the proposed model is valid.