Forecast of charging processes of a local charging infrastructure for potential flexibility provisioning based on methods of artificial intelligence

New solutions are required in order to meet the upcoming challenges faced by the electrical distribution grid due to the increasing share of renewable energies and power-intensive loads such as charging stations (CSs) for electric vehicles (EVs) or heat pumps. In this contribution, methods of artificial intelligence (AI) are used to forecast the standing times of electric vehicles at charging stations. By forecasting these standing times, a more effective charging is possible. Smart charging management systems (CMS) are capable to reduce electrical distribution grid's bottlenecks, by regulating the operational behaviour of the charging stations. In order to train the AI, data of a local charging infrastructure are used. Through a structured dataset, a cross validation is done to find the most accurate AI model to predict the standing time of the electric vehicles at the charging stations. The forecast of the trained AI model is analysed with a confusion chart. At last, the AI forecast is validated through a new dataset and the suitability of AIs in smart CMS is discussed.