ANN prediction approach analysis for performance and emission of antioxidant-treated waste cooking oil biodiesel

For the purpose of lowering hazardous emissions and enhancing performance of diesel engine, waste cooking oil biodiesel has emerged as a feasible and promising biofuel. In this research paper, 300 and 400 ppm doses of tert-butylhydroquinone (TBHQ) and diphenylamine (DPA) antioxidants were added to waste cooking oil biodiesel of 20% volume to evaluate performance and emission parameters in unmodified diesel engine. An artificial neural network model was developed to predict brake thermal efficiency (BTE), brake specific energy consumption (BSEC), nitrogen oxide emission (NOx), carbon monoxide emission (CO), hydrocarbon emission (HC), and smoke opacity by considering load, blends, and type of antioxidant in different doses as input. Prediction and validation were carried out using the findings of the experiments. The quasi-Newton method algorithm was used to predict data that best fits with linear regression analysis. The result showed at full load, BTE and BSEC have R 2 values of 0.985 and 0.995, respectively. The recommended ANN model's accuracy and performance were acceptable. At full load, the brake thermal efficiency increased, and brake specific energy consumption was reduced for fuel blend with antioxidant in respect of without antioxidant blend. NOx emission was reduced by 2.32, 5.24, 7.35, and 12.44% for 300-doses DPA blend, 300-doses TBHQ blend, 400 doses TBHQ blend, and 400 doses DPA antioxidant blend, respectively, compared to without antioxidant blend. The adoption of ANN to predict performance and emission can speed up and lower the running cost of understanding output behavior. Graphical Abstract