Artificial neural networks-based performance and emission characteristics prediction of compression ignition engines powered by blends of biodiesel derived from waste cooking oil

Compression ignition engines are essential for power generation, yet the use of fossil fuels like petrol and diesel results in the emission of harmful toxins into the atmosphere, leading to ecological disruption and exacerbating environmental issues. To address these challenges, there is a growing interest in alternative fuels, with vegetable oil-based biodiesels emerging as a promising option due to their renewable nature and comparable performance characteristics to that of diesel. The main aim of this study is to develop artificial neural network (ANN) models for predicting biodiesel performance and emission characteristics, aiming to reduce the reliance on resource-intensive physical testing through accurate property-based predictions. The optimum artificial neural network model topologies for the prediction of the performance parameters, i.e., brake specific fuel consumption, brake specific energy consumption, brake thermal efficiency, and exhaust gas temperature, were 6–5-1, 6–3-1, 6–3-1, and 6–5-1, respectively. The mean square error, root mean square error, mean absolute deviation and mean absolute percentage error for the brake thermal efficiency prediction were 0.0397, 0.1993, 0.1234, and 0.5599, respectively. The sensitivity analysis results showed that the artificial neural network model developed for the prediction of brake specific fuel consumption is highly sensitive to torque; similarly, the model for brake specific energy consumption is sensitive to brake power, and models for brake thermal efficiency and exhaust gas temperature are sensitive to fuel consumption. The results of this study aid in the broader adoption of biodiesel (i.e. derived from waste cooking oil) as a viable alternative fuel, mitigating greenhouse gas emissions and promoting sustainable energy practices for a greener future.