Active power distribution network user power consumption behavior variable weight combination prediction model and prediction method

The invention discloses an active power distribution network user power consumption behavior variable weight combination prediction model and prediction method, and relates to the field of active power distribution network load prediction. The system comprises an Elman neural network module used for solving a weight sequence of a combined prediction model, and a variable weight combined predictionmodule based on an ordered induction weighted average operator. The variable weight combination prediction module obtains a weight sequence of each single user power consumption behavior prediction model output by a network by taking an induced ordered weighted average operator of each single model as input of an Elman neural network, and synthesizes a prediction result of each single user powerconsumption behavior prediction model through the weight sequence. And a combined prediction result with higher precision is obtained. According to the method, the strong mapping capability of the Elman neural network is fully