Adaptive and predictive control of fuel injection with time delay consideration for air–fuel ratio regulation of gasoline direct injection engines. (July 2017)

• Record Type: Journal Article
• Authors: Feng, Yu; Jiao, Xiaohong; Wang, Zhijing
• Is Part Of: Proceedings of the Institution of Mechanical Engineers. Volume 231:Number 8(2017:Aug.)
• Page Start: 1022
• Page End: 1033
• Publication Date: 2017-07
• Subjects: Adaptive control -- air charge prediction -- air–fuel ratio -- gasoline direct injection engine -- time delay -- unscented Kalman filter; Mechanical engineering -- Congresses; Transportation engineering -- Congresses; 629.2
• DOI: 10.1177/0954407016670527
• Languages: English
• ISSNs: 0954-4070

Abstract

Accurate air–fuel ratio control is a key affecting factor for improving fuel economy and reducing exhaust emissions for internal combustion engines. Challenging issues in air–fuel control are the accurate estimation of cylinder air charge for achieving the stoichiometric in-cylinder air–fuel ratio and the disposition of measurement time delay from the oxygen sensor for removing its limits on the achievable feedback performance. In this article, based on hybrid discrete–continuous-time descriptions for the cylinder air charge dynamics and air–fuel feedback regulation controlled plant, a novel fuel injection controller with adaptive feedback and predictive feedforward is designed to ensure accurate air–fuel control of a gasoline direct injection engine. The feedforward fuel injection is determined based on the cylinder air charge prediction using unscented Kalman filter for the compensation of the injection delay and modelling error and the attenuation of the measurement noise. The feedback fuel compensation is designed as a proportional-integral structure with adaptive gains by means of an adaptive stabilization method of uncertain input delayed systems for the management of the transport delay and parameter uncertainty. The effectiveness of the proposed fuelling control against time delay, modelling error, measurement noise and parameter uncertainty is demonstrated by the simulation utilizing experimental data from a real V6 GDI engine.