A study on the hydrogen consumption calculation of proton exchange membrane fuel cells for linearly increasing loads: Artificial Neural Networks vs Multiple Linear Regression


ÖZÇELEP Y., SEVGEN S., ŞAMLI R.

RENEWABLE ENERGY, vol.156, pp.570-578, 2020 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 156
  • Publication Date: 2020
  • Doi Number: 10.1016/j.renene.2020.04.085
  • Journal Name: RENEWABLE ENERGY
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, PASCAL, Aerospace Database, Aquatic Science & Fisheries Abstracts (ASFA), CAB Abstracts, Communication Abstracts, Compendex, Environment Index, Geobase, Greenfile, Index Islamicus, INSPEC, Pollution Abstracts, Public Affairs Index, Veterinary Science Database, DIALNET, Civil Engineering Abstracts
  • Page Numbers: pp.570-578
  • Istanbul University Affiliated: No

Abstract

This paper presents an experimental study about the proton exchange membrane fuel cell (PEMFC) behavior on linearly increasing loads. The study mainly based on the effect of the linear load slope on hydrogen consumption for 0-600 W range and 0-100 Watt/s slope. Experimental results are processed by Artificial Neural Networks (ANN) and Multiple Linear Regression (MLR). The relationship between total consumed energy, peak power, slope and hydrogen consumption are discussed and novel equations are presented. The average error rates of ANN and MLR are 0.3189%, and 0.1124% while the average R-2 values are 0.9965 for ANN simulation and 0.9545 MLR simulation. We presented that the energy and exergy efficiency are decreased 6%, cost of the energy is increased 13% with the increasing slope of the power. We also performed the sensitivity and uncertainty analysis. The results give information to hydrogen system designers about an effective way to reach hydrogen consumption by performing both of the modelling processes successfully. (C) 2020 Elsevier Ltd. All rights reserved.