Abstract
The purpose of this study is to analyze the performance, the pollutant emissions, and the neural network modeling of a compression ignition (CI) engine operating on nanoparticles‐diesel fuel. Alumina nanoparticles that ranged in dosing level from 20 to 80 ppm were used as additives in diesel fuel along with the simultaneous usage of a 2% by volume surfactant. To achieve the uniform dispersion of AL2O3 nanoparticles, an ultrasonic vibrator was used. The results indicate a perceptible effect on engine performance and a decrease in the specific fuel consumption as compared to similar procedures that used base fuels. In addition, it was found that the use of this fuel resulted in lower amounts of NOx, HC, and CO emissions than that of diesel fuel. A generalized regression artificial neural network (GRNN) model was developed to predict a correlation between brake power, fuel consumption, HC, CO, NOx using different amounts of nanoparticles and speeds as input data. Predictive ability of this neural network is investigated considering mean square error (MSE) and correlation coefficient (R) values. The predicted results of the model led to the MSE values of 9.6346 × 10−5, 8.6470 × 10−4, 0.0213, 0.0088, and 2.5836 × 10−4 for power, fuel consumption, HC, CO, and NOx, respectively. Also, the R values that were obtained for these outputs are: 0.99999, 0.99912, 0.98506, 0.99977, and 0.9998, respectively. © 2015 American Institute of Chemical Engineers Environ Prog, 35: 540–546, 2016
