<![CDATA[This study aims to optimize the performance of fuel-injected motorcycles through the application of a Deep Neural Network (DNN) in the Electronic Control Unit (ECU) and the use of ethanol-pertalite fuel blends. The ethanol blends used in the study were 0%, 5%, 10%, 15%, and 20%. Fuel consumption tests were conducted using the standard ECE/324 driving cycle, and emission tests were performed according to Euro 4 standards. Tests were conducted on a real track to evaluate fuel consumption performance and exhaust gas emissions. The results indicate that the 15% ethanol blend (E15) provided optimal engine efficiency, while the 20% ethanol blend (E20) resulted in the largest reduction in carbon monoxide (CO) and hydrocarbon (HC) emissions. Furthermore, the DNN model with 50 neurons and a Sigmoid activation function demonstrated the best balance between accuracy (R=0.9868) and generalization (MSE=0.3843) in optimizing ignition timing and injection timing. In conclusion, the ethanol blends and the application of DNN in the ECU have proven effective in enhancing fuel efficiency and reducing exhaust emissions, supporting the development of more sustainable transportation technologies.]]>
Copyrights © 2025
