Intelligent real time control strategy and power management based on MPC and LSTM-TCN model for standalone DC microgrid with energy storage

Standalone microgrids powered by renewable energy face major challenges of stability and reliability due to the intermittent nature of those energy sources and fast load shifting. To mitigate these challenges, an effective control strategy and power management are required to ensure power balancing and minimizing fluctuations. This paper presents a novel intelligent control and power management strategy for standalone DC microgrids. The primary objectives of this control strategy are real-time voltage regulation and power balancing, as well as preventing the energy storage system from overcharging and over discharging. The microgrid contains a PV system with energy storage systems, including a battery and supercapacitor. The proposed control strategy is based on a LSTM-TCN model and model predictive control (MPC). The LSTM-TCN model forecasts the microgrid disturbances including environmental conditions (irradiance and temperature) and the load demand. To effectively integrate the forecasted values in the MPC architecture, the sigmoid function is applied, enabling a smooth transition between the actual system states and predicted ones especially during high variation of the disturbances. Performance evaluation of the proposed control strategy conducted through comparisons with established control methods under the variation of environmental conditions and load demand. Results show that the proposed control approach provides excellent voltage stability, fast response time, and low overshoot, performing better than other control strategies, especially during high load variation.