A LSTM-based model predictive control method for unlocking the potential of building energy flexibility in Malaysian commercial buildings.

Buildings are regarded as promising energy flexibility resources due to their significant energy consumption and better integration into the electricity grid. To fully exploit the potential of building flexibility in tropical climates, optimized operational strategies need to be developed in which cost savings and thermal comfort should be considered. Model predictive control (MPC) is widely acknowledged as one of the effective methods for developing optimal strategies. However, the practical implementation of traditional MPC is often hindered by substantial computational burdens. This study proposes a Long Short-Term Memory based model predictive control (LSTM-LBMPC) method employing LSTM neural networks to learn and imitate MPC behavior from a dataset containing optimal control trajectories. This method eliminates the need for online optimization, significantly reducing dependency on computational resources. The simulation experiment was performed on a Malaysian commercial office building with a variable air volume (VAV) cooling system under tropical climate conditions. The results indicate that, compared to the baseline control strategy, traditional MPC and LSTM-LBMPC reduced energy costs by 13.89% and 12.75%, respectively, and peak electrical loads by 30.20% and 27.8%, respectively, without compromising thermal comfort. Especially, compared to traditional MPC, LSTM-LBMPC can significantly reduce the computational cost by as much as 99.8%, with only a small trade-off in performance.