A Precise Prediction Model Tool for Distribution Network Reliability Indicators Based on Big Data and Multi method Integration

Aiming at the problem of low accuracy in predicting the efficiency indicators of intelligent distribution networks, a precise prediction model tool for the efficiency indicators of intelligent distribution networks is proposed. On the basis of determining the evaluation of the efficiency indicators of the intelligent distribution network, expert methods and interpretive structural models are first used to mine the internal and external multiple information, clarify the hierarchy of each information's impact on the indicators, and on this basis, comprehensively identify the influencing factors of each indicator as much as possible. Then, using the partial least squares method - variable projection importance analysis method, the VIP value of the influencing factors is calculated to accurately measure the importance of their impact on the target variable (studied indicator value), achieving dimensionality reduction of the influencing factors. Next, establish a benefit simulation model based on system dynamics model, set boundary conditions, conduct variable analysis, simulate and calculate the predicted values of each indicator, and verify the accuracy of the results using historical big data. After the correctness and accuracy of the model are verified, this tool is used to predict reliability indicators for future years.