Using artificial neural networks for representing the air flow rate through a 2.4 liter VVT engine

The emerging Variable Valve Timing (VVT) technology complicates the estimation of air flow rate because both intake and exhaust valve timings significantly affect engine's gas exchange and air flow rate. In this paper, we propose to use Artificial Neural Networks (ANN) to model the air flow rate through a 2.4 liter VVT engine with independent intake and exhaust camshaft phasers. The procedure for selecting the network architecture and size is combined with the appropriate training methodology to maximize accuracy and prevent overfitting. After completing the ANN training based on a large set of dynamometer test data, the multi-layer feedforward network demonstrates the ability to represent air flow rate accurately over a wide range of operating conditions. The ANN model is implemented in a vehicle with the same 2.4 L engine using a Rapid Prototype Controller. Comparison between a mass air flow (MAF) sensor and the ANN model during a typical dynamic maneuver shows a very good agreement and superior behavior of the network during the transient. Practical recommendations regarding the production implementation of the ANN are provided as well.

[1]  M. S. Goodman,et al.  Transient Air/Fuel Ratio Control of an S.I. Engine Using Neural Networks , 1996 .

[2]  Zoran Filipi,et al.  Quasi-Dimensional Computer Simulation of the Turbocharged Spark-Ignition Engine and its Use for 2- and 4-Valve Engine Matching Studies , 1991 .

[3]  Zoran Filipi,et al.  Variable geometry turbine (VGT) strategies for improving diesel engine in-vehicle response: a simulation study , 2004 .

[4]  C. F. Aquino.  Transient A/F Control Characteristics of the 5 Liter Central Fuel Injection Engine , 1981 .

[5]  Ulf Holmberg,et al.  Different Strategies for Transient Control of the Air-Fuel Ratio in a SI Engine , 2000 .

[6]  Joseph R. Asik,et al.  Transient A/F Estimation and Control Using a Neural Network , 1997 .

[7]  C. R. Ferguson,et al.  A Turbulent Entrainment Model for Spark-Ignition Engine Combustion , 1977 .

[8]  D.R. Hush,et al.  Progress in supervised neural networks , 1993, IEEE Signal Processing Magazine.

[9]  Ulrich Lenz,et al.  Artificial Intelligence for Combustion Engine Control , 1996 .

[10]  Ben Shannon,et al.  Further refinement and validation of a turbulent flame propagation model for spark-ignition engines , 1980 .

[11]  Elbert Hendricks,et al.  Mean Value Modelling of Spark Ignition Engines , 1990 .

[12]  Mrdjan J. Jankovic,et al.  Methods to Reduce Air-charge Characterization Data for High Degree of Freedom Engines , 2004 .

[13]  David J. C. MacKay,et al.  A Practical Bayesian Framework for Backpropagation Networks , 1992, Neural Computation.

[14]  Jacek Czarnigowski,et al.  Hybrid Air/Fuel Ratio Control Using the Adaptive Estimation and Neural Network , 2000 .

[15]  Elbert Hendricks,et al.  Modelling of the Intake Manifold Filling Dynamics , 1996 .

[16]  Kurt Hornik,et al.  Approximation capabilities of multilayer feedforward networks , 1991, Neural Networks.

[17]  Kishan G. Mehrotra,et al.  Elements of artificial neural networks , 1996 .

[18]  M Jankovic,et al.  Cylinder air-charge estimation for advanced intake valve operation in variable cam timing engines , 2001 .

[19]  Enrico Mattarelli,et al.  Simulation of a High Performance Turbocharged S.I. Engine , 1998 .

[20]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.