Robust Adaptive Control of DC Microgrid Converters via DE‐RAUKF‐MPC Integration

The stability of DC microgrid converters is challenged by the negative incremental impedance of constant power loads (CPLs) and by performance degradation caused by parametric uncertainties and non‐Gaussian mixed noise. Existing control strategies face a fundamental trade‐off: Nonlinear disturbance observers (NDOs) are prone to noise amplification, whereas integral‐augmented model predictive control (MPC) suffers from dynamic conflicts between model prediction and integral correction. To address these dual challenges, this paper proposes an integrated control framework combining a disturbance‐estimating RAUKF (DE‐RAUKF) with a physics‐informed MPC. The proposed method has three main features. First, it combines a standard augmented‐state disturbance‐estimating observer structure with a dual adaptive covariance mechanism, in which Huber/Mahalanobis‐driven measurement adaptation and physical‐subspace process adaptation jointly address impulsive noise and structural model mismatch. This observer achieves accurate estimation under parameter perturbations and impulsive noise interference. Second, an active disturbance compensation mechanism injects the real‐time disturbance estimate directly into the MPC prediction model. This physics‐informed strategy helps linearize the nonlinear CPL dynamics from a closed‐loop perspective, thereby avoiding the convergence latency associated with parameter‐identification‐based compensation. Third, an adaptive penalty alternating direction method of multipliers (ADMM) solver balances solution accuracy and computational efficiency during aggressive transients. Simulation results demonstrate that the proposed strategy outperforms the NDO‐based sliding mode control and integral‐augmented MPC baselines. Under severe CPL conditions, it achieves steady‐state precision with RMSE together with strong transient robustness, validating its effectiveness for high‐reliability DC microgrid voltage regulation.