Continuous Control Set Model Predictive Control of a Bridgeless-Boost Three-Level Active Rectifier

A problem of the switched reluctance drive is its natural torque pulsations, which are partially solved with finite control set model predictive control strategies. However, the continuous control set model predictive control, required for precise torque stabilization and predictable power converter behavior, needs sufficient computation resources, thus limiting its practical implementation. The proposed model predictive control strategy utilizes offline processing of the magnetization surface of the switched reluctance motor. This helps to obtain precalculated current references for each torque command and rotor angular position in the offline mode. In online mode, the model predictive control strategy implements the current commands using the magnetization surface for fast evaluation of the required voltage command for the power converter. The proposed strategy needs only two lookup table operations requiring very small computation time, making instant execution of the whole control system possible and thereby minimizing the control delay. The proposed solution was examined using a simulation model, which showed precise and rapid torque stabilization below rated speed.

Download Full-text

A Comparison of Finite Control Set and Continuous Control Set Model Predictive Control Schemes for Model Parameter Mismatch in Three-Phase APF

Frontiers in Energy Research ◽

10.3389/fenrg.2021.727364 ◽

2021 ◽

Vol 9 ◽

Author(s):

Jianfeng Yang ◽

Yang Liu ◽

Rui Yan

Keyword(s):

Model Predictive Control ◽

Predictive Control ◽

Parameter Uncertainty ◽

Prediction Error ◽

Continuous Control ◽

Power Electronic ◽

Parameter Mismatch ◽

Three Phase ◽

Finite Control ◽

Control Set

Model predictive control (MPC) methods are widely used in the power electronic control field, including finite control set model predictive control (FCS-MPC) and continuous control set model predictive control (CCS-MPC). The degree of parameter uncertainty influence on the two methods is the key to evaluate the feasibility of the two methods in power electronic application. This paper proposes a research method to analyze FCS-MPC and CCS-MPC’s influence on the current prediction error of three-phase active power filter (APF) under parameter uncertainty. It compares the performance of the two model predictive control methods under parameters uncertainty. In each sampling period of the prediction algorithm, different prediction error conditions will be produced when FCS-MPC cycles the candidate vectors. Different pulse width modulation (PWM) results will be produced when CCS-MPC solves the quadratic programming (QP) problem. This paper presents the simulation results and discusses the influence of inaccurate modeling of load resistance and inductance parameters on the control performance of the two MPC algorithms, the influence of reference value and state value on prediction error is also compared. The prediction error caused by resistance mismatch is lower than that caused by inductance mismatch, more errors are caused by underestimating inductance values than by overestimating inductance values. The CCS-MPC has a better control effect and dynamic performance in parameter mismatch, and the influence of parameter mismatch is relatively tiny.

Download Full-text