Title: Using some performance parameters to predict exhaust gas emissions of a turboprop engine: adaptive neuro-fuzzy method

Authors: Yasin Şöhret; Işıl Yazar; T. Hikmet Karakoç

Addresses: Keciborlu Vocational School, Suleyman Demirel University, TR-32700 Isparta, Turkey ' Department of Mechatronics, Eskisehir Vocational School, Eskisehir Osmangazi University, TR-26250 Eskisehir, Turkey ' Faculty of Aeronautics and Astronautics, Anadolu University, TR-26470 Eskisehir, Turkey

Abstract: This paper presents an exhaust gas emissions prediction model for a turboprop engine depending on some performance parameters. Within this context, experimentally collected emissions data is used to develop a model in the adaptive network-based fuzzy inference system. For system identification in the adaptive network-based fuzzy inference system, grid partitioning is preferred as the clustering method, and the accuracy of the prediction model is acceptable to the best of the authors' knowledge. The root mean square error is found to be 0.12375, 4.7332, 0.081264 and 0.033515 for the emissions index prediction of carbon monoxide, carbon dioxide, nitrogen oxides and unburned hydrocarbons, respectively.

Keywords: aircraft emissions; adaptive neuro-fuzzy inference system; ANFIS; combustion; turboprop engines; performance parameters; emissions prediction; exhaust gas emissions; neural networks; fuzzy logic; system identification; grid partitioning; clustering; carbon monoxide; carbon dioxide; CO2; carbon emissions; nitrogen oxides; NOx; unburned hydrocarbons.

DOI: 10.1504/IJSA.2016.076051

International Journal of Sustainable Aviation, 2016 Vol.2 No.1, pp.1 - 14

Received: 28 Nov 2015
Accepted: 08 Dec 2015
Published online: 22 Apr 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article