Source load typical scene selection method and device based on scene analysis and medium

• Main Authors: YANG LULU, SHI FANGDI, LUO JINSHAN, WAN ZHIWEI, JIN QIANG, SU WEIHUA, HAN JINGLIN, WANG CONG, WANG XIAOXUE, ZHANG MINGZE, ZHANG YI, CUI KAI, ZHANG JIALEI
• Format: Patent
• Language: Chinese; English
• Published: 08.03.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

The invention relates to a source load typical scene selection method based on scene analysis, and the method comprises the steps: carrying out the processing of historical real data based on a generative adversarial network to generate original scene data according with the distribution rule of the historical real data, and the historical real data comprise historical photovoltaic output data and historical load data; the original scene data comprises a photovoltaic output sample and a load demand sample; performing scene clustering on the original scene data based on a mean shift algorithm to obtain multiple pieces of clustered scene data; and based on a preset deviation index of a typical day, calculating a difference between each clustered scene data and historical real data, and selecting the scene data with a difference load preset consistent requirement as a target source load typical scene. According to the method, high-precision scene sample data conforming to historical photovoltaic output and load