ANN-BASED PREDICTION OF DIESEL ENGINE PERFORMANCE AND EMISSIONS USING TAMARIND BIODIESEL BLENDED WITH IRON OXIDE (II, III) NANOPARTICLES

This paper presents an artificial neural network model that was developed to accurately predict the performance and emission characteristics of a diesel engine fueled with tamarind biodiesel blends enhanced with iron (II, III) oxide nanoparticles. The model uses input parameters, such as the biodiesel blend ratio, nanoparticle concentration, engine load, and calorific value. Kolmogorov’s theorem is applied to vary neurons number from 5 to 9 in the hidden layer, while the Levenberg-Marquardt (Trainlm) algorithm is used for training with the backpropagation method for prediction. Optimal ANN architectures for performance parameters, including specific fuel consumption (SFC) and brake thermal efficiency (BTE) were 4-5-1 and 4-6-1, with respective RMSE/MAPE values of 0.0101/1.865 and 0.6153/1.982. For emission parameters (HC, CO, NOx, and smoke), the best models were 4-9-1, 4-5-1, 4-6-1, and 4- 6-1, with RMSE of 0.0087, 0.1850, 19.7955, and 1.0502 and MAPE values of 5.1731, 2.9295, 1.6329, and 1.3463, respectively. These results highlight the effectiveness of ANN in modeling complex engine behaviors and the potential of nanoparticle-enriched tamarind biodiesel as a sustainable, low-emission alternative fuel.