Research of performance on a spark ignition engine fueled by alcohol–gasoline blends using artificial neural networks. (5th December 2015)

• Record Type: Journal Article
• Authors: Kapusuz, Murat; Ozcan, Hakan; Yamin, Jehad Ahmad
• Is Part Of: Applied thermal engineering. Volume 91(2015:Dec.)
• Page Start: 525
• Page End: 534
• Publication Date: 2015-12-05
• Subjects: Artificial neural network -- Spark ignition engine -- Modeling -- Ethanol -- Methanol; Heat engineering -- Periodicals; Heating -- Equipment and supplies -- Periodicals; Periodicals; 621.40205
• DOI: 10.1016/j.applthermaleng.2015.08.058
• Languages: English
• ISSNs: 1359-4311

Abstract

Abstract: In this paper, we investigate various alcohol–unleaded gasoline mixtures that can be used with no modifications in a spark-ignition engine. The mixtures consisted of 5%, 10% and 15% ethanol, methanol together and separately. Based on the recommendations of the Jordanian Petroleum Company (JoPetrol), total alcohol content should not exceed 15–20% owing to safety and ignition hazards. Optimizations for the use of alcohol were made for the maximum torque, maximum power and minimum specific fuel consumption values. For torque 0.9906, for brake power 0.997, and for brake specific fuel consumption 0.9312 regression values for tests have been obtained from models generated by the neural network. According to the modeling and optimizations, use of fuel mixture containing 11% methanol–1% ethanol for performance, and fuel mixture containing 2% methanol for BSFC were found to have better results. Moreover, the paper demonstrates that ANN (Artificial Neural Network) can be used successfully as an alternative type of modeling technique for internal combustion engines. Highlights: ANN model was developed and verified. Effects of alcohol–gasoline blends on performance of a SI engine are fairly simulated. Effects of alcohol–gasoline blends on performance of a SI engine are optimized.