With the escalating requirement for global sustainable energy solutions and the complexities linked with the complete transition to new technologies, internal combustion engines (ICEs) powered with biofuels like ethanol are gaining significance over time. However, problems linked to the performance and emissions of such ICEs necessitate accurate prediction and optimization. The study employed the integration of artificial neural networks (ANN) and multi-level historical design of response surface methodology (RSM) to address these challenges in alignment with the Sustainable Development Goals (SDGs). A single-cylinder spark ignition (SI) engine powered with ethanol-gasoline blends at different loads and speeds was used to gather data. Among six initially trained ANN models, the most efficient model with a regression coefficient (R2) of 0.9952 (training), 0.98579 (validation), 0.98847 (testing), and 0.99307 (overall) was employed to predict outputs such as brake power, brake specific fuel consumption (BSFC), brake thermal energy (BTE), concentration of carbon dioxide (CO2), carbon monoxide (CO), hydrocarbons (HC), and oxides of nitrogen NOx. Predicted outputs were optimized by incorporating RSM. On implementing optimized conditions, it was observed that BP and BTE increased by 19.9%, and 29.8%, respectively. Additionally, CO, and HC emissions experienced substantial reductions of 28.1%, and 40.6%, respectively. This research can help engine producers and researchers make refined decisions and achieve improved performance and emissions. The study directly supports SDG 7, SDG 9, SDG 12, SDG 13, and SGD 17, which call for achieving affordable, clean energy, sustainable industrialization, responsible consumption, and production, taking action on climate change, and partnership to advance the SDGs as a whole respectively.
Read full abstract