Robust deep Koopman model predictive load frequency control of interconnected power systems

Load Frequency Control (LFC) plays an essential role in interconnected power systems for maintaining power system frequency by regulating power generation to match load demand. However, nonlinearities and external disturbances pose significant challenges for LFC of interconnected power systems. In this paper, a robust deep Koopman model predictive control (MPC) method is introduced for effective LFC of a nonlinear interconnected power system. The method uses Koopman operator to obtain a linearized system model and introduces a deep neural network (DNN) to approximate this operator. By applying the Koopman operator to acquire the linearized system model, a linear MPC controller can be used for LFC of the nonlinear interconnected power system. Furthermore, a feedback controller is proposed to incorporate with the MPC controller to mitigate approximation errors and external disturbances, thereby enhancing the robustness and control performance of the controller. The robustness and stability of the proposed controller are studied using a Lyapunov-based technique, and the advantages and effectiveness of our approach in addressing the challenges of LFC in interconnected power systems are demonstrated through simulations.