Model-Free Predictive Direct Control of Unified Power Quality Conditioner Based on Ultra-Local Model

Abstract

To enhance the robustness of the unified power quality conditioner (UPQC) with finite control set model predictive control (FCS-MPC), direct control in the dq coordinate system faces challenges such as complex coordinate transformation, system coupling issue and phase-locked-loop delays. In this study, a predictive direct control strategy based on an ultra-local model (ULM) is proposed. A model-free parametric predictive direct current control scheme for the UPQC within the αβ framework is crafted by amalgamating an ultra-local model with predictive direct control. Derived from the generalised instantaneous power theory and the active power equilibrium within the UPQC system, the current command generation mechanism of UPQC is established, and the parallel active power filters (PAPF) current command generation mechanism is developed by integrating dead-beat control alongside the strategy for maintaining a consistent baseline voltage magnitude for the connected load. This approach effectively navigates the complex coordinate transformation and system coupling issue, realises no phase-locked loop, no system parameters and no PI outer-loop controller control and simplifies the control system structure. Simulation results show that even with 50% parameter mismatch, the proposed strategy can still maintain the grid current THD at < 2%. Finally, we verify the feasibility of the strategy through simulations and experiments.