基于滑动窗口-小波提升对地电场预测算法  

Prediction algorithm for electric field to ground based on sliding window-lifting wavelet transform

作　　者：顾涛 GU Tao(School of Computing,North China Institute of Science and Technology,Yanjiao 065201,China)

机构地区：[1]华北科技学院计算机学院

出　　处：《计算机工程与设计》2019年第10期2903-2907,2984,共6页Computer Engineering and Design

基　　金：中央高校基本科研业务费专项基金项目(3142015024);河北省物联网监控工程技术研究中心基金项目(3142016020)

摘　　要：为解决10KV架空线路对地电场值的估计问题,提出一种基于滑动窗口-db2小波提升变换对地电场预测算法。将每个滑动窗口宽度取20个采样点,利用3个相邻的滑动窗口的db2小波提升变换低频系数加权值构造出第4个窗口的小波提升变换低频系数,使用第3个滑动窗口小波提升变换高频系数作为第4个窗口的小波提升变换高频系数;将所得到的第4个窗口低频系数、高频系数进行逆小波提升变换,得到第4个窗口的20个数据,并取最后一个数据作为下一个采样点的预测初值;预测偏差取第4个窗口前19个变换数据与第3个窗口所对应的19个测量数据的偏差数学期望求取;最终预测值为预测初值加上修正系数与预测偏差乘积。将仿真结果与其它算法进行比较,比较结果表明,该算法在电场极值点处预测精度高于其它算法,数据分析表明,所预测10KV架空线路对地电场值相对误差满足现场指标要求,预测数据可以作为技术指标和事故分析依据。To solve the problem of estimating the value of electric field to ground for 10 KV overhead lines,an algorithm based on sliding window-db2 lifting wavelet transform was proposed.The width of each sliding window was taken as 20 sampling points,and the low frequency coefficients of lifting wavelet transform in the fourth window were constructed by weighting the low frequency coefficients of db2 lifting wavelet transform in three adjacent sliding windows,and the high frequency coefficients of the third sliding window lifting wavelet transform were used as the high frequency coefficients of the lifting wavelet transform in the fourth window.The low frequency coefficients and high frequency coefficients of the fourth window were transformed through inverse lifting scheme,and 20 data of the fourth window were obtained,and the last data were taken as the initial value of prediction.The prediction deviation was calculated by the mathematical expectation of the deviation between the first 19 transformed data of the fourth window and the 19 measured data corresponding to the third window.The final predicted value was the sum of the initial predicted value plus the value,which was the predicted deviation multiplied by the correction coefficient.The result of simulation,which was compared to other well-known algorithms,indicates that the proposed algorithm is better than several other algorithms,especially in improving the accuracy of extreme point prediction.The data analysis also shows that the relative error of the predicted data of electric field to ground of 10 KV overhead line meets the requirements of field target,and the predicted data can be used as technical index and accident analysis basis.

关 键 词：对地电场预测 滑动窗口 db2小波提升 预测偏差 数学期望 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]