Title :
Predication of wireless communication failure in grid metering automation system based on logistic regression model
Author :
Tao Liu ; Shaofeng Wang ; Shaocheng Wu ; Jing Ma ; Yueming Lu
Author_Institution :
Shenzhen Power Supply Bur. Co., Ltd., Beijing Univ. of Posts & Telecommun., Shenzhen, China
Abstract :
Shenzhen power supply bureau (SZPSB) urgently need a series of methods which can analyze the cause of wireless communication fault and predict wireless communication failure, these methods can facilitate operations staff take the initiative to master the signal status, solve the problem of signal fault in time, and increase the terminal online rate. In this paper, we present a model named GCFPM (Gradient descent to iterative Calculate optimal regression coefficient in power-grid communication Failures Prediction Model) which based on Logistic Regression Algorithm (LRA), GCFPM can effectively predict the possibility of communication failure, enhance the level of acquisition terminal operations, increase the terminals online-rate and guarantee the practical effect of the electric power metering automation system (EPMAS).
Keywords :
failure analysis; gradient methods; iterative methods; metering; power grids; power system measurement; radiocommunication; regression analysis; telecommunication network reliability; EPMAS; GCFPM model; LRA; SZPSB; Shenzhen power supply bureau; acquisition terminal operations; electric power metering automation system; gradient descent to iterative calculate optimal regression coefficient in power-grid communication failures prediction model; grid metering automation system; logistic regression model; signal fault problem; terminal online rate; wireless communication failure prediction; Abstracts; Automation; Data models; Educational institutions; Erbium; Logistics; Predictive models; communication failure; fault probability; logistic regression model; power grid;
Conference_Titel :
Electricity Distribution (CICED), 2014 China International Conference on
Conference_Location :
Shenzhen
DOI :
10.1109/CICED.2014.6991837