基于双通道多特征融合的电力负荷智能感知方法

负荷识别是分析用户用电行为的主要工具之一.提高负荷识别的精度对于开展用能监测服务、实现节能降损具有重要意义.提出了一种基于双通道多特征融合的电力负荷智能感知方法.首先,从电器设备的基本属性出发,分析了电流、谐波、功率等数值特征以及电压-电流(V-I)轨迹图像特征对负荷识别的影响;其次,考虑了特征之间的互补性,分别采用主成分分析-神经网络(principal component analysis-back propagation,PCA-BP)算法和卷积神经网络算法将数值特征和图像特征以不同通道在高维空间进行深度融合;最后,采用Softmax分类算法对融合后的高级特征和设备标签进行有监督的学习,从而实现了不同类别电器设备的有效辨识.算例测试结果显示,所提出方法的负荷识别准确率高达94.55％.该结果充分说明了将多种特征进行高维空间融合,可以更全面、立体地反映设备的本质属性,提高算法的识别精度.

Intelligent Power Load Identification Method Based on Dual-Channel And Multi-Feature Fusion

Load identification is one of the main tools to analyze the behavior of electric power consumers. It is of great significance to improve the accuracy of load identification for power monitoring and energy saving. In this paper, an intelligent load identification method based on dual channel and multi-feature fusion was presented. Starting from the basic properties of electrical equipment, the influence of current, harmonic, power and other numerical features as well as V-I image features on load recognition was analyzed; then, considering the complementarity between the features, the PCA-BP neural network and convolutional neural network algorithms were used to deeply fuse numerical features and image features in high-dimensional space with different channels; finally, the Softmax classification algorithm was applied to learn the advanced features and equipment labels in a supervised way, so as to realize the effective identification of different types of electrical equipment. The test results show that the load identification accuracy of this method is as high as 94.55%. The results fully demonstrate that the multi-feature fusion in high-dimensional space can reflect the essential attributes of the equipment in a more comprehensive and stereoscopic manner, which leads to a significant improvement in the accuracy of load identification.