Pareto-based modulated model predictive control strategy for power converter applications

Abstract

Modulated model predictive control (M 2 PC) allows fixed switching frequency operation of power converters, producing lower ripple and lower total harmonic distortion (THD) in the output signals than those obtained using regular model predictive control (MPC) algorithms. However, the design of the M 2 PC algorithm and its performance depend on, among other factors, the tuning of weighting factors, which are required by the cost functions typically used in predictive control algorithms. In this paper, a novel Pareto-based multi-objective M 2 PC (MO-M 2 PC) strategy is proposed. In this case, the use of weighting factors is not required, and the optimization problem is solved using a multi-objective approach. The aim of the proposed MOM 2 PC strategy is to consider practical rules, such as satisfaction of soft constraints, to calculate the control actions. The proposed Pareto-based MOM 2 PC strategy can be applied to any power converter topology. In this paper, experimental validation of the proposed methodology is pursued using a 3-phase 3 kW power converter operating as a shunt active power filter (SAPF). The proposed control is implemented in a control platform based on the dSPACE ds1103 system. The simulation and experimental results demonstrate the advantages and flexibility achieved by the proposed Pareto-based MOM 2 PC.

Pareto Optimal Weighting Factor Design of Predictive Current Controller of a Six-Phase Induction Machine Based on Particle Swarm Optimization Algorithm

“…The latter gave us the motivation to address an exhaustive comparison of the proposed MOPSO method against other Pareto-based technique (e.g. [33]) and the ANN method and/or NSGA-II as a near-future research topic.…”

Constrained Modulated Model Predictive Control for a Three-Phase Three-Level Voltage Source Inverter

“…In addition, each control strategy must also obtain the optimal region by finding the minimal cost function value, which is evaluated at each of the 24 regions. The unconstrained M 2 PC only requires nineteen evaluations of the cost function, expressed in (15), to obtain the cost of the 24 regions (17). In contrast, the proposed constrained M 2 PC requires evaluating its cost function (29) once per region.…”