Enhancing Stability of DC Cascaded Systems With CPLs Using MPC Combined With NI and Accounting for Parameter Uncertainties

Compared with the traditional passive- and active-damping techniques for mitigating the impedance mismatch between the input impedance of constant power loads (CPLs) and the output impedance of the dc link within dc microgrids, there is a growing interest in leveraging model predictive control (MPC). MPC is proven to have the advantage of addressing the impedance mismatch issue, which leads to system instability and oscillations, owing to its constant handling, multivariable control, predictive capabilities, and optimization-based control. Nevertheless, the generalized MPC faces some technical challenges as it involves intricate enumeration and operations in switching state selection and observer design, thereby increasing computational complexity. Moreover, issues such as variable switching frequency and a lack of protection mechanisms impede its widespread application. To overcome these challenges, this article proposes an MPC technique employing a nonlinear inductor (NI) in the dc cascaded system to enable its faster and more stable transitions. The proposed controller utilizes the dead-beat principle to simplify the predictive process and alleviate the digital implementation burden in the control stage. The incorporation of the NI extends the stability range of the control-to-output transfer characteristics of the power conversion stage with CPL. Subsequently, two Luenberger observers are introduced, capable of handling load uncertainty and adapting to changes in system parameters. The selection criteria and stability analysis of the observer parameters are discussed. The proposed MPC is simple to implement and operates at a fixed switching frequency. Validation through simulation and experimental results confirms the effectiveness of the technique in improving system stability.