Abstract
•The effect of different alcohol-gasoline fuel blends on engine performance and emissions simultaneously was investigated.•ANNs were developed to examine the performance of fuel blends in a gasoline engine.•The GRNN was selected to predict values based on low MSE and AARD% value.
This research investigated artificial neural network (ANN) modeling to predict the exhaust emissions and engine performance. Different percentages of alcohol at various engine speeds and comparison ratios were used to obtain the required data for testing and training the proposed ANN. Six experimental datasets were used for the training process to develop an ANN model based on the standard program. Furthermore, the accuracy of the proposed model was evaluated by calculating the mean square error (MSE), regression coefficient (R2), and average absolute relative deviation (AARD%). The total values of AARD% for the proposed model were 10.50 and 15.45% for carbon monoxide and hydrocarbon emissions and 10.50 and 3.13% for torque and fuel consumption, respectively, which were acceptable errors compared with the experimental uncertainty.