Adaptive Optimal Control of Faulty Nonlinear DC Microgrids With Constant Power Loads: Dual-Extended Kalman Filter Approach

This article investigates the problem of estimating actuator fault and states and controlling the bus voltage in direct current microgrids (DC MGs) with linear and nonlinear constant power loads (CPLs). It is considered that the DC MG states are not fully measurable and the utilized sensors are not ideal and noisy. Additionally, the actuator fault occurs and it is modeled as an additive term in the power system dynamics. These issues, including nonlinearities, un-measurable states, noisy measures, and actuator fault indispensably degrade the operation of the DC MG. To solve this issue, initially, a dual-extended Kalman filter (dual-EKF) is suggested for the fault and state estimation. It decomposes the process of estimating the state and actuator fault to reduce the online computational burden. For the control purpose, a linear parameter varying (LPV) model predictive control (MPC) is suggested to regulate the current and voltage of the DC MG. It benefits the nonlinear system modeling of LPV representation and constrained-based design procedure of the MPC to result in an accurate and low online computational burden dealing with system constraints. By deploying the overall robust adaptive dual-EKF estimation-based LPV-MPC, there is no need to have any prior knowledge of all system states and actuator faults in prior. The theoretical analysis and controller design are validated by numerical simulations on a typical islanded DC MG and comparisons are done with state-of-the-art estimation and control strategies.