基于神经网络的电力通信预警系统设计及优化  

Design and Optimization for Electric Power Communication Early Warning System Based on Neural Network

作　　者：沈雪红 SHEN Xuehong(School of Communication Engineering,Zhejiang Technical College of Posts and Telecom,Shaoxing 312366,China)

机构地区：[1]浙江邮电职业技术学院通信工程学院,浙江绍兴312366

出　　处：《自动化仪表》2021年第6期48-51,56,共5页Process Automation Instrumentation

摘　　要：为提高电力通信预警系统的预测精度和效率,以SDH电网通信预警系统为基础,通过分析其基本结构和现状,开展了基于人工神经网络的粗糙集算法优化和通信预警系统设计研究。以电网中通信光板故障为研究对象,选取80组样本数据,开展了优化后的通信预警系统预测误差精度对比分析。结果发现:采用粗糙集优化后的电网预警系统预测值误差率分别为4.66%、6.12%、3.37%和1.78%,平均值为3.98%;而传统神经网络预测算法的误差率为7.63%、8.81%、5.22%和6.25%,平均值为6.98%;优化后的预警系统预测精度相比较提升约3.23%。所提出的算法明显提高了电力通信预警系统的运算效率和精确程度,为电力系统预警提供了一定参考和借鉴。In order to improve the prediction accuracy and efficiency of the power communication early warning system,based on the SDH power grid communication early warning system,by analyzing in the basic structure and status,the design and study for artificial neural network-based rough set algorithm optimization and communication early warning system are carried out.Taking the communication light board failure in the power grid as the research object,80 sets of sample data are selected to carry out a comparative analysis of the prediction error accuracy of the optimized communication early warning system.The results show that the error rates of the predicted value of the grid early warning system optimized by rough set are 4.66%,6.12%,3.37%and 1.78%,with an average of 3.98%;the error rates corresponding of the traditional neural network prediction algorithm are 7.63%,8.81%,5.22%and 6.25%,with an average of 6.98%.The prediction accuracy of the optimized early warning system is increased by 3.23%.The proposed algorithm improves the operational efficiency and accuracy of the power communication early warning system,and provides a reference for early warning in power system.

关 键 词：电力 通信 预警 神经网络 粗集度 误差分析 优化 电网 

分 类 号：TH86[机械工程—仪器科学与技术]