Research on Electric Load Identification Method of Integrated Energy System Based on Image Analysis

Aiming at the problem that the identification accuracy of traditional non-invasive load identification algorithm in low-frequency sampled data under the background of integrated energy system(IES) is low, a comprehensive energy load identification algorithm based on Markov transition field (MTF) and densely connected network (DenseNet) was proposed. The power information is converted into color image features by MTF coding to improve the identification. The image data is input into the DenseNet121 deep convolutional neural network to extract the feature information and finally realize the accurate identification of load types. The identification method is verified in many aspects on the public data set and compared with other classification algorithms. The results show that MTF images have clear and easily distinguishable feature information, which strengthens the feature expression of input network. DenseNet network's powerful deep feature capture ability can effectively solve the problem of information loss in the process of image coding, thus improving the identification accuracy of power load type under the background of IES.