Research on Real-time State Identification Model of Electricity-heat System Considering Unbalanced Data

Abstract

It is meaningful to study the real-time state monitoring and recognition of integrated energy system and grasp its state in time for stable operation of it. A state identification method based on multi-class data equalization and Light Gradient Boosting Machine (LightGBM) is proposed for integrated energy systems. First, Latin hypercube sampling is used to simulate the load of the system in a day, and different system states are set up. Then, the energy flow model is used to calculate the system power flow under different states, and the feature indexes are obtained. Aiming at the unbalanced data, the synthetic minority oversampling technology is used to preprocess data to achieve the balance of data sets in each state. The pre-processed data is utilized to train the LightGBM classifier, and the optimal hyperparameters of the model are obtained based on the K-fold cross-validation and grid search method. Finally, the preprocessed data set is used to verify the proposed method. The calculation results show that the accuracy of the identification model reaches 99.2%. Compared with other traditional methods, the model can accurately identify the operating state of the electricity-heat energy system at any time section.