Prediction of various engine-out parameters by use of artificial intelligence techniques

Since experimental studies on internal combustion engines are both time consuming and costly, various modelling techniques can be used to predict experimental results. Artificial intelligence techniques are prominent among other techniques. These techniques can tackle complex, non-linear problems. In this study, four different engine models were handled which intended to estimate various engine-out parameters. In Model 1, performance and emission parameters of a diesel engine operated with biodiesel-alcohol mixtures were estimated. In Model 2, performance and emission parameters of a diesel engine operated with diesel fuel with nanoparticle additives were estimated. In model 3, various operational parameters of a spark ignition engine operated with 95 RON gasoline fuel were estimated. In model 4, vibration characteristic of a diesel engine operated with different diesel-biodiesel blends with HHO gas addition into intake manifold were estimated. Regression analysis, artificial neural networks and adaptive neuro fuzzy inference system methods were used to make predictions. In was concluded that, regression analysis is not capable of predicting the parameters accurately. On the other hand, artificial neural networks and adaptive neuro fuzzy inference system made more accurate estimations.