Model-data hybrid driven power system scheduling method and system

• Main Authors: ZHANG HUANMING, ZHOU RUIYE, ZHAO BIMEI, ZHAO XIANGYU, CAO SHANG, LIANG LINGYU
• Format: Patent
• Language: Chinese; English
• Published: 09.07.2024
• 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 model-data hybrid driven power system scheduling method and device, computer equipment, a storage medium and a computer program product. The method comprises the steps of obtaining an optimal solution of a preset demand based on a preset physical model and a mathematical optimization equation, and setting the optimal solution as an expert strategy; constructing a scheduling model for simulating the expert strategy based on the expert strategy, and training the scheduling model; a Gaussian mixture model is constructed, a scheduling environment is judged, and if the scheduling environment is similar to an existing sample in a preset sample pool, scheduling operation is carried out based on the trained scheduling model; otherwise, directly calling an expert strategy to carry out scheduling operation. By adopting the method, the generative adversarial imitation learning agent can be constructed by utilizing the adversarial thought of the generative adversarial network in an agent traini