Currently, fuel economy and thermal efficiency are more important to all engines. Efficiency is increased with cooled air by intercooler. Most of researches regarding engineering problems generally deal with experimental studies. But, the experimental researches are quite expensive and time consuming. In the last decades, Neural Networks (NN) had been used increasingly in a variety of engineering applications. The objective of the study is to investigate the adequacy of neural networks (NN) as a quicker, more secure and more robust method to determine the effects of intercooling on performance of a turbocharged diesel engine’s specific fuel consumption. The data are obtained from experimental research that is performed by the author. NN based model is developed, trained and tested through a based MATLAB program by using of these data. In the study, break specific fuel consumption (BSFC, g/kWh) was analysed with intercooling and without intercooling. The statistical analysis is performed to explain the performance of the NN based model. NN based model outputs are also compared with the experimental results. The statistical results and the comparison demonstrated that the NN based model is highly successful to determine the effects of intercooling on performance of a turbocharged diesel engine’s specific fuel consumption.

Key words: Neural networks, intercooling, specific fuel consumption, scaled conjugate gradient algorithm, diesel engine.