Abstract
The prediction of friction mean effective pressure (FMEP) is important when engine performance is estimated with model-based development (MBD) process. In this study, prediction of the FMEP of a single cylinder heavy duty diesel engine was attempted by utilizing a neural network technique. The experimental results, in which various operating and combustion parameters are included, taken by the engine under various operating conditions (5227 points) were evenly divided into training data (3684 points) and validation data (1543 points), and machine learning was performed by using training data. FMEP of the evaluation data were predicted by the pre-trained neural network model. As the result, it was confirmed that the predicted results of FMEP show good agreement with the actual results of FMEP of the evaluation data (experimental data).
