Determination of bio-diesel engine combustion pressure using neural network based model

Combustion pressure analysis is an important aspect to be studied in the research and development of internal combustion engines. However, measurements of incylinder combustion pressure for a complete range of testing are time-consuming and costly, as it required high accuracy pressure sensor systems. Alternatively, a simulation model based on the computer program can be used to retrieve those parameters. This study focused on developing the prediction model to determine the combustion pressure of diesel engines by employing artificial neural network methods. Input data for training, testing, and validation of the model were obtained from laboratory engine testing. The biodiesel blends percentage, engine loads, engine speeds and crank angle position were selected as the input parameters. The performance of the ANN model was validated against the experimental data. The results show that the developed model successfully predicted the engine combustion pressure with a higher correlation coefficient (R-value) between 0.99968-0.99973, means that the model produces 99% of prediction accuracy. In addition, the prediction errors occurred within a small range of values. This study revealed that the neural network approach is able to predict the combustion pressure of the diesel engine with high accuracy