Joint Control of Storage Voltage Prediction and Grid-forming Converter using LSTM

Conventional grid-forming converter (GFC) control design requires comprehensive consideration of source-side capacity configuration. However, the constraints of source-side energy storage devices make the parameters of GFC a technical challenge of misfit. Therefore, in this paper, we propose short-term prediction of GFC capacity and converter control based on long short-term memory network (LSTM). First, we need to construct the state models of the energy storage device and the storage converter (SC). Then, the operation data of SC is collected and trend prediction is realized based on LSTM. Finally, the joint control strategy of GFC and SC is proposed and the prediction data is incorporated. Experiments in the hardware-in-the-loop platform show that the joint control method can stably control the power flow of SC and GFC.