嵌入式NILM电力负荷识别及特征库构建系统设计

Design of Embedded NILM Power Load Acquisition and Feature Database Construction System

针对非侵入式电力负荷识别系统负荷特征库构建困难,算法复杂,硬件成本高的问题,构建了基于STM32嵌入式处理的电力负荷采集识别系统。介绍了系统硬件结构。分析了基于嵌入式处理器的电力数据通信方式。设计了低复杂度的快速滤波算法和基于最小二乘的负荷状态监测、识别算法。基于本系统,对常见家用电力负荷实际工作状态进行了实测。测试数据分析表明本系统的综合识别率符合要求,且算法更加适用于嵌入式硬件结构。系统硬件成本低,有利于推广应用。

Aiming at the problems of difficult construction of load characteristic database,complex algorithm and high hardware cost of non-invasive power load identification system,a power load acquisition and identification system based on STM32 embedded processing is constructed.The hardware structure of the system is introduced.The power data communication mode based on embedded processor is analyzed.A low complexity fast filtering algorithm and a load condition identification algorithm based on least square are designed.Based on this system,the actual working state of common household power load is measured.The analysis of test data shows that the comprehensive recognition rate of the system meets the requirements,and the algorithm is more suitable for embedded hardware structure.The hardware cost of the system is low,which is conducive to popularization and application.