Simplified Decoupler-Based Multivariable Controller with a Gain Scheduling Strategy for the Exhaust Gas Recirculation and Variable Geometry Turbocharger Systems in Diesel Engines

Abstract

This paper presents a simplified decoupler-based multivariable controller with a gain scheduling strategy in order to deal with strong nonlinearities and cross-coupled characteristics for exhaust gas recirculation (EGR) and variable geometry turbocharger (VGT) systems in diesel engines. A feedback controller is designed with the gain scheduling strategy, which updates control gains according to engine operating conditions. The gain scheduling strategy is implemented by using a proposed scheduling variable derived from indirect measurements of the EGR mass flow, such as the pressure ratio of the intake, exhaust manifolds, and the exhaust air-to-fuel ratio. The scheduling variable is utilized to estimate static gains of the EGR and VGT systems; it has a large dispersion in various engine operating conditions. Based on the estimated static gains of the plant, the Skogestad internal model control (SIMC) method determines appropriate control gains. The dynamic decoupler is designed to deal with the cross-coupled effects of the EGR and VGT systems by applying a simplified decoupler design method. The simplified decoupler is beneficial for compensating for the dynamics difference between two control loops of the EGR and VGT systems, for example, slow VGT dynamics and fast EGR dynamics. The proposed control algorithm is evaluated through engine experiments. Step test results of set points reveal that root-mean-square (RMS) error of the gain-scheduled feedback controller is reduced by 47% as compared to those of the fixed gain controller. Furthermore, the designed simplified decoupler decreased the tracking error under transients by 14-66% in various engine operating conditions.