Predictive modeling of performance of a helium charged Stirling engine using an artificial neural network


ÖZGÖREN Y. Ö., Cetinkaya S., SARIDEMİR S., ÇİÇEK A., KARA F.

ENERGY CONVERSION AND MANAGEMENT, cilt.67, ss.357-368, 2013 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 67
  • Basım Tarihi: 2013
  • Doi Numarası: 10.1016/j.enconman.2012.12.007
  • Dergi Adı: ENERGY CONVERSION AND MANAGEMENT
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Sayfa Sayıları: ss.357-368
  • Anahtar Kelimeler: Beta type Stirling engine, Helium, ANN, Engine performance, RHOMBIC-DRIVE MECHANISM, SPARK-IGNITION ENGINE, SURFACE-ROUGHNESS, EXHAUST EMISSIONS, SYSTEM, CYCLE, TEMPERATURE, EFFICIENCY, DISPLACER, PISTON
  • Gazi Üniversitesi Adresli: Evet

Özet

In this study, an artificial neural network (ANN) model was developed to predict the torque and power of a beta-type Stirling engine using helium as the working fluid. The best results were obtained by 5-11-7-1 and 5-13-7-1 network architectures, with double hidden layers for the torque and power respectively. For these network architectures, the Levenberg-Marquardt (LM) learning algorithm was used. Engine performance values predicted with the developed ANN model were compared with the actual performance values measured experimentally, and substantially coinciding results were observed. After ANN training, correlation coefficients (R-2) of both engine performance values for testing and training data were very close to 1. Similarly, root-mean-square error (RMSE) and mean error percentage (MEP) values for the testing and training data were less than 0.02% and 3.5% respectively. These results showed that the ANN is an acceptable model for prediction of the torque and power of the beta-type Stirling engine. (C) 2012 Elsevier Ltd. All rights reserved.