A Mathematical Modelling Approach Based on Artificial Neural Network for Estimation of Key Parameters in Internal Combustion Engine

In a conventional powertrain driven by Internal combustion (IC) engines, there are many key sensors such as intake manifold temperature sensor, exhaust temperature sensor, Nitrogen oxide (NOx) sensors etc. to monitor engine performance and emissions. For enabling various diagnostics strategies and engine monitoring, virtual sensors or modelled values of such sensors play an important role. Conventional strategies incorporate the use of regression models, map-based models and physics-based models. There are a few drawbacks with conventional models in terms of accuracy and model calibrations efforts. Data driven models or neural networks have fairly better accuracy and reliability for estimating complex parameters. Representing the neural network with a mathematics-based model would help to eliminate drawbacks associated with conventional modelling approach. The proposed methodology uses artificial intelligence technique called artificial neural network (ANN) for estimation, for example, intake manifold temperature estimation. The data driven model is built in python. Modelling process of ANN comprises of feature selection, data scaling, and training/testing with predefined set of neurons in each layer. Once the ANN is trained, weights and biases of each neuron and intermediate connections are obtained. The relationship between each neuron in the input layer, hidden layers and output layers is established using the weights and biases. Subsequently a mathematical model is built using the above information to replicate the results obtained by ANN. For initial validation, the ANN was tested with real world vehicle data. Statistical analysis and time series analysis between actual intake manifold temperature and estimated intake manifold temperature is done for different engine operating conditions. Based on the analysis, it was concluded that the results obtained from ANN demonstrate high accuracy. Furthermore, the mathematical model was validated against the results obtained by ANN. A fairly accurate match was observed between the ANN output and the results obtained by mathematical model.