With the intelligent development of Electric Power System (EPS), the intelligent demand for the cabinet management of electric tools and appliances is more and more urgent. This article focuses on the processing and application of multi-source data of power instrument cabinets under the Internet of things (IoT) environment, aiming at realizing its intelligent identification and accurate prediction of operation state. By analyzing the composition, operation principle and characteristics of multi-source data of power equipment cabinet in IoT environment, an intelligent identification algorithm model of multi-source data based on deep belief network is constructed, and different types of data are input into the model for learning and training after preprocessing, feature extraction and fusion. Operation state prediction combines time series decomposition with long-term and short-term memory network (LSTM) algorithm, and inputs historical data into LSTM network for training. The results show that the intelligent identification of multi-source data is accurate and effective. The prediction accuracy of the running state prediction model is 85%, the recall rate is 80%, the F1 value is 82%, the false alarm rate is 12%, and the missing report rate is 20%. Compared with traditional linear regression and simple neural network model, this model has obvious advantages. It can be seen that the method proposed in this article has effectively improved the intelligent management level of power instrument cabinets.
Chen et al. (Sun,) studied this question.