The effect of low-boiling carbon-neutral fuels mixed with conventional fuels on the performance of diesel engines was investigated for the decarbonization of heavy-duty vehicles. Three types of fuel were used as test fuels: diesel oil and a 6:4 mixture of diesel oil and n-paraffin fuels, n-hexane and n-heptane (DP2β and DP3β, respectively). These fuels were fed into a 1.5-liter single-cylinder, direct-injection heavy-duty diesel engine with a piston compression ratio of 23. The results showed that DP2β and DP3β reduced soot, a type of emission gas, by 45% (DP2β) and 30% (DP3β), respectively, compared to diesel oil under medium-load (IMEP: 1.4) conditions and at an EGR ratio of 30%. In addition, a simple AI model was created to predict engine power and exhaust performance from various engine control parameters, considering the current situation where engine control and calibration are becoming more complex with the introduction of new technologies. The AI model consists of an input layer, a hidden layer, and an output layer. The prediction accuracy of the AI model was improved and optimal operating conditions were analyzed. As a result, the secondary accuracy of net fuel consumption rate, Soot emissions, and NOx emissions were improved to R: 0.99, respectively. Then, analysis of optimal operating conditions was conducted using PSO, and by changing the weights of net fuel consumption rate, Soot emissions, and NOx emissions in the evaluation function of PSO, operating conditions were found that preferentially reduced the parameters with the highest weights.
Matsuda et al. (Wed,) studied this question.