Title: A data mining study on combustion dynamics and NOx emission of a swirl stabilised combustor with secondary fuel injection

Authors: Rouzbeh Riazi; Mohamad Asrardel; Maziar Shafaee; Shidvash Vakilipour; Hadi Zare; Hadi Veisi

Addresses: Faculty of New Sciences and Technologies, University of Tehran, Tehran, 14395-1561, Iran ' Faculty of New Sciences and Technologies, University of Tehran, Tehran, 14395-1561, Iran ' Faculty of New Sciences and Technologies, University of Tehran, Tehran, 14395-1561, Iran ' Faculty of New Sciences and Technologies, University of Tehran, Tehran, 14395-1561, Iran ' Faculty of New Sciences and Technologies, University of Tehran, Tehran, 14395-1561, Iran ' Faculty of New Sciences and Technologies, University of Tehran, Tehran, 14395-1561, Iran

Abstract: To study the relations between the amounts of NOx emission, noise level, and level of pressure fluctuations as the output quantities of an experimental swirl-stabilised combustor and two variables of overall equivalence ratio (ϕ) and secondary fuel injection rate (Qsec), as its input quantities, two different data mining approaches were employed in the present work (i.e., artificial neural network (ANN) and multiple polynomial regression (MPR) techniques). The related experiments were already carried out using four different types of secondary fuel injectors with an overall equivalence ratio (ϕ) in the range of 0.7~0.9. The results indicate that both the ANN and MPR methods have lower predicting capability for estimation of noise level and the level of pressure fluctuations compared with that of the emission index. Also the results show that the ANN has better predicting capability, for estimation of various combustor parameters, than the MPR method.

Keywords: swirl-stabilised combustor; secondary fuel injection; pressure fluctuation; emission index; MPR; multiple polynomial regression; neural network.

DOI: 10.1504/IJHVS.2017.084865

International Journal of Heavy Vehicle Systems, 2017 Vol.24 No.3, pp.215 - 238

Available online: 29 May 2017 *

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