An ANN based modelling, forecasting, and experimental study of emissions and performance parameters running on microalgae biodiesel-nanoparticles blended fuel

In this current analysis, biodiesel derived from third-generation microalgae spirulina is assessed as a replacement for diesel. Also, the influence of nano additives Al₂O₃ and MgO in microalgae spirulina amalgams on compression ignition engine attributes is studied. Additionally, using an Artificial Neural Network (ANN), an optimization model was designed to characterize the test variables. Test investigation revealed that, at peak load, the overall drop in BTE was observed to be 3.55, 3.18, and 4.16 respectively, for Al₂O₃, MgO doped blends, and biodiesel mixture (without additive), than SB0. On the other hand, the BSFC of Al₂O₃, MgO doped blends, and biodiesel mixture enhanced by 5.77, 5.31, and 7.49%, respectively. At peak load, CO₂ emittants of Al₂O₃, MgO doped blends and spirulina biodiesel mixture were 9.35%, 10.11%, and 7.584% higher than SB0. NO_X of Al₂O3, MgO-doped blends, and spirulina biodiesel amalgams decreased by 7.79%, 8.3%, and 10.31%, respectively, at full load. On the other hand, a reduction of HC of about 13.37%, 13.66%, and 12.91% for amalgams with Al₂O₃, MgO additives, and without additives is observed. Three sets of transfer functions namely soft plus, softmax, and sigmoid along with optimizers adagrad, nadam, and adam were tried to develop the model. The model trained with Adam optimizer with 16 neurons in the hidden layer stands out as the best among other algorithms with the least MSE and strong r. The r and MSE of engine characteristics NO_X, HC, CO₂, BTE, and BSFC are 0.99992, 0.99888, 0.99959, 0.99989, 0.99969 and 0.00002, 0.00015, 0.00007, 0.00003, 0.00004 respectively. This research reveals that the ANN strategy can be employed effectively to predict engine characteristics.