CNN-BiLSTM-AE-based short-term power load prediction method

• Main Authors: XU QIHANG, CAI HANFEI, CHEN XI
• Format: Patent
• Language: Chinese; English
• Published: 24.10.2023
• Subjects: CALCULATING; CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTINGELECTRIC POWER; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; ELECTRICITY; GENERATION; PHYSICS; SYSTEMS FOR STORING ELECTRIC ENERGY; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

The invention relates to a short-term load prediction method and system based on a combined deep learning model. In the invention, a short-term power load prediction model is developed, and a real load data set in a certain region of Australian is utilized to evaluate the short-term power load prediction model. The method comprises the following steps: firstly, preprocessing input data to delete missing, redundancy and outliers; next, different normalization technologies are applied to better represent the input data, so that an effective model is generated; furthermore, a new model CNN-BiLSTM-AE is developed by the invention. The model compares three prediction modules: CNN, BiLSTM-AE and FC. Firstly, two CNN layers are used for extracting information from several variables in a data set, then the information is input into BiLSTM-AE, BiLSTM converts a sequence into a coded feature vector, and then decoding is carried out through another BiLSTM. The coded feature vector layer copies the coded sequences, and f