Electric vehicle autonomous charging control method based on deep reinforcement learning under weak link communication

• Main Authors: ZHANG QUAN, XIANG XINYU, CHEN YI, SUN YUXI, HUO YINGNING, WANG WEIFENG, XU CHUANZI, YU HEYANG, GENG GUANGCHAO, JIANG QUANYUAN
• Format: Patent
• Language: Chinese; English
• Published: 05.09.2023
• 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; ELECTRIC EQUIPMENT OR PROPULSION OF ELECTRICALLY-PROPELLEDVEHICLES; ELECTRICITY; ELECTRODYNAMIC BRAKE SYSTEMS FOR VEHICLES, IN GENERAL; GENERATION; MAGNETIC SUSPENSION OR LEVITATION FOR VEHICLES; PERFORMING OPERATIONS; PHYSICS; SYSTEMS FOR STORING ELECTRIC ENERGY; TRANSPORTING; VEHICLES IN GENERAL

The invention discloses an electric vehicle autonomous charging control method based on deep reinforcement learning under weak link communication. The method comprises the following steps: acquiring total load prediction data of household electric appliances connected with a power distribution area, acquiring historical charging information of an electric vehicle, uploading the historical charging information to a communication master station, constructing a super agent, training to obtain a super agent with an optimal decision, packaging and issuing the super agent for a charging pile of the power distribution area, and sending the packaged and issued charging pile to the power distribution area. And under the condition of weak link communication, the power distribution area and the charging pile perform autonomous control according to the real-time total load prediction data and the charging demand information by using the electric vehicle intelligent agent, and the optimal charging action of the electric v