基于改进卡尔曼滤波的电力计量数据深度融合  

Deep fusion of power metering data based on improved Kalman filter

作　　者：吴海杰 符艺超 谢敏 WU Haijie;FU Yichao;XIE Min(China Southern Power Grid Digital Grid Group(Hainan)Co.,Ltd.,Haikou 570010,China)

机构地区：[1]南方电网数字电网集团(海南)有限公司,海南海口570010

出　　处：《电子设计工程》2025年第5期111-114,119,共5页Electronic Design Engineering

基　　金：南方电网有限责任公司科技项目(ZN-YD-007)。

摘　　要：当前的电力计量数据融合方法并未对复杂数据进行时间配准,导致数据深度融合效果较差。为了有效解决这一问题,提出了基于改进卡尔曼滤波的电力计量数据深度融合方法。引入分布图法改进传统卡尔曼滤波算法,初始化卡尔曼滤波参数并填充时序数据。利用拉格朗日插值方法所构建插值基函数对数据进行降噪处理。利用小波变换方法预处理电力计量数据,结合时间配准与LEACH协议构建出多通道融合路径,实现电力计量数据的深度融合。分析实验结果可知,该方法的最大融合误差仅为0.2,能够使全部数据融合到一起,融合效果较好。The current power metering data fusion method does not carry out time registration for complex data,resulting in poor data deep fusion effect.In order to solve this problem effectively,a deep fusion method of power metering data based on improved Kalman filter is proposed.The distribution graph method is introduced to improve the traditional Kalman filter algorithm,initialize Kalman filter parameters and fill the time series data.The interpolation basis function constructed by Lagrange interpolation method is used to reduce the noise of the data.Wavelet transform method is used to preprocess power metering data,and multi-channel fusion path is constructed by combining time registration and LEACH protocol to realize deep fusion of power metering data.The analysis of experimental results shows that the maximum fusion error of this method is only 0.2,and all the data can be fused together,and the fusion effect is better.

关 键 词：改进卡尔曼滤波 电力计量数据 深度融合 时间配准 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]