Development of an in-cylinder processes model of a CVVT gasoline engine using artificial neural network

Today, employing model based design approach in powertrain development is being paid more attention. Precise, meanwhile fast to run models are required for applying model based techniques in powertrain control design and engine calibration. In this paper, an in-cylinder process model of a CVVT gasoline engine is developed to be employed in extended mean valve control oriented model and also model based calibration procedure. In-cylinder models are static thermos-fluid models, which predict the performance and emission index of engines based on boundary conditions of cylinder. Due to computations burden of thermos-fluid models, they are not fast enough to be used in control models. In this paper a validated thermodynamic model of engine is developed using a commercial engine analyzing software. The developed model is employed for generation input-output data sets which are used for training an artificial multi-layer neural network. In order to increase the richness of data, the Sobol method is employed to generate input data to thermodynamic model. Based on output trend, the output data are divided to two clusters and two corresponding distinct neural networks are employed. In order to validate the modeling performance the neural network results are compared to experimental results in both full and part load conditions. Comparison of neural network results with experimental results shows that the developed model is able to predict the engine emission and performance indices with required accuracy and fast enough in both full-load and part-load conditions and might be employed in extended mean value models as well as model based engine calibration with required performance.