Effective control of spark ignition engines (SIE) under all operating conditions is essential for achieving high fuel economy, low emissions and high vehicle performance. Design and development of high performance control system is a challenging problem due to the variety of engine operating regimes, the complexity of nonlinear physical and chemical engine processes, a number of unmeasurable variables which directly affect important engine variables, multiplicity of control inputs and outputs, process/measurement noise and load disturbances. In this paper, the most important problems of torque tracking and air-to-fuel ratio (AFR) stabilization at the stoichiometric level are addressed. To provide a suitable solution for this problem, a data driven approach based on the design of direct and inverse models is proposed. The inverse model is represented by a grey box with a selected fixed structure, outputs which are the control variables and a set of input variables as nonlinear functions of the engine state and regulated variables. The direct model is also represented as a grey box, but the regulated variables are the model outputs and the control variables are the model inputs. The parameters of the grey box models are estimated through an offline identification procedure using vehicle data and a special representation of the models in the form of linear regressions. The controller is designed to maintain the combined gain of tandem "inverse model direct model" close to unity at all engine operating regimes. Two approaches for parameter estimation are proposed and justified. One approach is based on the substitution of the regulated desired value in the inverse model for its current value, and the other is based on the pseudo inverse of the direct model. Both approaches result in the design of a feedforward controller. In practice, the feedforward controller is augmented by a PID controller to provide improved performance in the presence of modeling errors and external disturbances. The final controller is robust to uncontrollable disturbances. Test results demonstrating the performance of the algorithms are presented and discussed.
[1]
Simon Haykin,et al.
Neural Networks: A Comprehensive Foundation
,
1998
.
[2]
Kumpati S. Narendra,et al.
Identification and control of dynamical systems using neural networks
,
1990,
IEEE Trans. Neural Networks.
[3]
B. Widrow,et al.
Adaptive inverse control
,
1987,
Proceedings of 8th IEEE International Symposium on Intelligent Control.
[4]
Richard D. Braatz,et al.
On the "Identification and control of dynamical systems using neural networks"
,
1997,
IEEE Trans. Neural Networks.
[5]
John B. Heywood,et al.
Internal combustion engine fundamentals
,
1988
.
[6]
Derong Liu,et al.
A biologically inspired adaptive nonlinear control strategy for applications to powertrain control
,
2009,
2009 IEEE International Conference on Systems, Man and Cybernetics.
[7]
Myoungho Sunwoo,et al.
Feedback Error Learning Neural Networks for Air-to-Fuel Ratio Control in SI Engines
,
2003
.
[8]
Lennart Ljung,et al.
System Identification: Theory for the User
,
1987
.
[9]
Vadim I. Utkin,et al.
Automotive engine diagnosis and control via nonlinear estimation
,
1998
.