Impact prediction model of acetone at various ignition advance by artificial neural network and response surface methodology techniques for spark ignition engine

In this study, it was aimed to predict and optimize the effects of acetone/gasoline mixtures on spark ignition engine responses at different engine speeds and ignition advance values with artificial neural network and response surface methodology. The regression results obtained from response surface methodology show that absolute variance ratio values for all answers are greater than 0.96. Correlation coefficient values obtained from artificial neural network were obtained higher than 0.91. Mean absolute percentage error values were between 0.8859% and 9.01427% for artificial neural network, while it was between 1.146% and 8.957% for response surface methodology. Optimization study with response surface methodology revealed that the optimum results are 1700 rpm engine speed, 2% acetone ratio and 11 degrees before top dead center ignition advance with a combined desirability factor of 0.76523%. Additionally, in accordance with the confirmation analysis among the optimal outcomes and the estimation outcomes, it was stated that there is a great harmony with a maximum error percentage of 7.662%. As a result, it is concluded that the applied response surface methodology and artificial neural network models can perfectly provide the impact of acetone percentage on spark ignition engine responses at different engine speeds and ignition advance values.

Automated summary

This study aims to predict and optimize the effects of acetone/gasoline mixtures on spark ignition engine responses using artificial neural network and response surface methodology. The results show that the applied models can accurately assess the impact of acetone percentage on engine performance.