Abstract:Air conditioning load occupies a large proportion of residential electricity consumption, and monitoring its operation is of great significance, which can guide residents to use electricity reasonably, ensure the stable operation of the power grid, and at the same time provide decision-making support for demand response to alleviate the contradiction between supply and demand. At present, the air conditioning load monitoring work is relatively small, mostly carried out as part of non-intrusive load monitoring, and the selected input data are mainly considered to be generalized, with less attention paid to the operation law of air conditioning and the influence of meteorological factors, resulting in the accuracy of air conditioning monitoring is still insufficient. To address this problem, this paper proposes an air conditioning monitoring method based on expert knowledge and twin branch networks. Based on Long Short Term Memory Neural Network(LSTM), a temporal branch is constructed to deeply mine the temporal characteristics of air conditioner operation, and based on Back Propagation Neural Network(BP), an expert knowledge electrical feature branch is constructed to capture the operation pattern of air conditioning. Meanwhile, the experimental results of this paper on one year of real data from 20 users show that compared with the traditional LSTM model, the model can realize accurate identification of air conditioning loads and provide support for the assessment of demand response potential of air conditioning loads.