Finite-Set Ultralocal Model Predictive Current Control with Integral Sliding Mode Observer for Quadratic Buck Converters of Data Centers

Data centers demand efficient and reliable power converters to handle their increasing energy requirements while maintaining operational stability. This paper introduces a novel control strategy for quadratic buck converters (QBCs), to improve performance in data center power systems. The proposed method applies an integral sliding mode observer (ISMO) to the finite-set ultralocal model predictive current control (FSUMPCC). The ISMO estimates unmeasured states and external disturbances, ensuring robust performance against parameter variations and external perturbations. Since neither the ultralocal model nor the ISMO, do not utilizes the system’s parameters, the proposed FS-UMPCC achieves high control precision in the presence of severe parameter mismatch. The proposed method also enhances system reliability by reducing sensitivity to model inaccuracies and external disturbances. The simulation validation demonstrates the effectiveness of the proposed approach. Results show significant control accuracy compared to the conventional finite control set model predictive control (FCS-MPC) strategy.