Power load probability prediction method based on constrained parallel LSTM quantile regression

Abstract

The invention discloses a power load probability prediction method based on constrained parallel LSTM quantile regression, and the method comprises the steps: collecting the load power and impact factor data of a plurality of sample days, and forming a data set; setting model hyper-parameters; establishing a constrained parallel LSTM model, and pre-training each quantile LSTM in the constrained parallel LSTM model to obtain a weight and offset parameter set; performing overall training on the constrained parallel LSTM model, performing fine adjustment on the weight and offset parameters in thetraining process, and determining the optimal weight and offset parameters of the constrained parallel LSTM model; inputting the verification set into the trained constraint parallel LSTM model, andselecting an optimal hyper-parameter of the model according to the verification error; and inputting the test sample into the constrained parallel LSTM model with the optimal hyper-parameter, and carrying out inverse normalization on a prediction result output by the constrained parallel LSTM model. According to the method, quantile regression prediction of the power load is carried out by adopting the constrained parallel LSTM model, so that the predicted load probability distribution is more reasonable, and intersection between quantile prediction values is avoided.

Description

technical field [0001] The invention belongs to the field of power load forecasting, and in particular relates to a power load probability forecasting method based on constrained parallel LSTM quantile regression. Background technique [0002] Short-term power load forecasting is the basis for safe and economical operation of power systems, and provides important information for power system planning and operation, energy trading, unit start-up and shutdown, and economic dispatch. Improving the accuracy of load forecasting can help improve the utilization of electrical equipment and minimize energy waste. [0003] At present, load probability forecasting methods mainly include interval estimation, kernel density estimation and quantile regression. The first two methods are mainly based on the parameter statistics estimation probability distribution of the point forecast error, while the quantile regression can directly explain the relationship between the response variable ...

AI Technical Summary

Problems solved by technology

[0006] The technical problem of the present invention is that the quantile prediction value of the existing quantile regression method of electric load has a crossover phenomenon, which leads to unreasonable

Images