摘要
为提升智能电表量测数据挖掘效果,该文研究一种基于小波去噪和时频分析的智能电表量测数据挖掘方法。应用小波变换阈值去噪方法对数据集进行去噪,通过惩罚策略选择阈值后,对存在噪声的数据集进行运算,获取去噪后的数据集,根据去噪后的小波系数与存在噪声的小波系数获取最优阈值函数;利用自适应最优径向高斯核时频分析方法,有效将最优阈值函数的数据集分离为自分量信号与互分量信号,精准挖掘智能电表仿真模型数据库内数据,完成智能电表量测数据信息的输出。实验结果表明,所研究方法去噪效果较好,相对误差保持在2%以内,挖掘精度维持在96%以上,应用性能较好。
In order to improve the efficiency of intelligent meter measurement data mining,a method of intelligent meter measurement data mining based on wavelet denoising and time-frequency analysis is studied.The wavelet transform threshold denoising method is applied to denoise the dataset.After selecting the threshold through a penalty strategy,operations are performed on the noisy dataset to obtain the denoised dataset,and the optimal threshold function is obtained based on the denoised wavelet coefficients and the noisy wavelet coefficients;Using the adaptive optimal radial Gaussian kernel time-frequency analysis method,effectively separating the dataset of the optimal threshold function into self component signals and cross component signals,accurately mining data in the smart meter simulation model database,and completing the output of smart meter measurement data information.The experimental results show that the noise removal effect of the proposed method is good,the relative error is maintained within 2%,and the mining accuracy is maintained above 96%,with better application performance.
作者
杨元
郭庆
YANG Yuan;GUO Qing(Ordos Institute of Technology,Ordos 017000,China;Ordos Substation of Inner Mongolia Autonomous Region Environmental Monitoring Station,Ordos 017000,China)
出处
《电子设计工程》
2024年第7期78-81,86,共5页
Electronic Design Engineering
关键词
智能电表
量测数据
小波去噪
AORGK时频分析方法
小波变换阈值
intelligent meter
measurement data
wavelet denoising
AORGK time-frequency analysis method
wavelet transform threshold
