Application of artificial neural network to predict specific fuel consumption and exhaust temperature for a Diesel engine Adnan Parlak a, * , Yasar Islamoglu b , Halit Yasar b , Aysun Egrisogut b a Sakarya University, Technical Education Faculty, Mechanical Education Department, 54187 Esentepe, Sakarya, Turkey b Sakarya University, Mechanical Engineering Department, 54187 Esentepe, Sakarya, Turkey Received 24 March 2005; accepted 5 October 2005 Available online 18 November 2005 Abstract The ability of an artificial neural network model, using a back propagation learning algorithm, to predict specific fuel consump- tion and exhaust temperature of a Diesel engine for various injection timings is studied. The proposed new model is compared with experimental results. The comparison showed that the consistence between experimental and the network results are achieved by a mean absolute relative error less than 2%. It is considered that a well-trained neural network model provides fast and consistent results, making it an easy-to-use tool in preliminary studies for such thermal engineering problems. Ó 2005 Elsevier Ltd. All rights reserved. Keywords: Specific fuel consumption; Exhaust temperature; Injection timing; Diesel engine; Neural networks 1. Introduction The digital computer provided a rapid means of per- forming many calculations involving the artificial neural network (ANN) methods. Along with the development of high speed digital computers, the application of ANN approach could be progressed a very impressive rate. In recent years, this method has been applied var- ious disciplines including automotive engineering, in forecasting of engine thermal characteristics for different working conditions. Some researchers studied this method to predict internal combustion engine character- istics. Artificial neural network approach has been used by Xu et al. [1], in forecasting engine systems reliability, Yuanwang et al. [2], to analyze the effect of cetane num- ber on exhaust emissions from engine, Korres et al. [3], to predict diesel lubricity, Lucas et al. [4], to model Die- sel particulate emission, Hafner et al. [5], for diesel engine control design, Shayler et al. [6], in automotive engine management systems, Tan and Saif [7], to model the intake manifold and throttle body processes in an automotive engine. In the existing literatures, it was shown that the use of ANN is a powerful modeling tool that has the ability to identify complex relationships from input–output data. However, no investigation to predict specific fuel con- sumption and exhaust temperature for Diesel engine using ANN approach appears to have been published in the literature to date. Therefore, the present work investigates the applicability of ANN method for pre- dicting the aforementioned parameters. 2. Data gathering method In the present study, the Ricardo E6 type, single-cyl- inder, four stroke, water cooled, and pre-combustion chamber engine was used. A schematic diagram of the experimental setup, and the test engine picture used 1359-4311/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.applthermaleng.2005.10.006 * Corresponding author. Fax: +90 264 3460262. E-mail address: parlak@sakarya.edu.tr (A. Parlak). www.elsevier.com/locate/apthermeng Applied Thermal Engineering 26 (2006) 824–828