“…It is also difficult to use any direct measurement or observation for verifying these instantaneously [16,17,18,19] Therefore, developing physical models, which will inherently reflect the main characteristics of the injection, wall wetting, evaporation and mixing, is paramount for predicting the amount of fuel available for combustion, which is directly linked to the torque output of the engine. The concept of model based engine calibration is not a new field of study [30,69]; however, it has received more attention recently for the above stated reasons. Some of the models rely heavily on experimental data and statistical inference, some rely on a semi-empirical model, [20,21,49] and some are based on experimental data and statistical inference for training a neural network based model [22].…”