基于HOG和SVM的非侵入式负荷识别

Non-Intrusive Load Identification Based on HOG and SVM

智能电网时代,准确高效的居民用电负荷评估对改善和调节电力网络的传输结构至关重要。对用户用电的电流、有功功率和谐波电流等数据进行数据预处理,基于对偶树复小波变换对数据降噪,建立基于HOG和SVM分类识别模型提取数据特征,并进行负荷识别。非侵入式负荷识别极大地降低数据收集和分析成本,对居民使用电器类型和数量的实时监测,为准确估算居民用电载荷提供可靠依据。

In the era of smart grid,accurate and efficient residential power load assessment is essential to improve and adjust the transmission structure of the power network.Data preprocessing is performed on user electricity current,active power and harmonic current data.Data is denoised based on dual-tree complex wavelet transform.To extract data features and identify load,the classification and recognition models based on HOG and SVM are established.Non-intrusive load identification greatly reduces the cost of data collection and analysis and monitors the type and quantity of electrical appliances which is used by residents.It can provide a reliable basis for accurately estimating the electrical load of residents.