Modeling of heating and cooling performance of counter flow type vortex tube by using artificial neural network


KOCABAŞ F., KORKMAZ M., Sorgucu U., Donmez S.

INTERNATIONAL JOURNAL OF REFRIGERATION-REVUE INTERNATIONALE DU FROID, vol.33, no.5, pp.963-972, 2010 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 33 Issue: 5
  • Publication Date: 2010
  • Doi Number: 10.1016/j.ijrefrig.2010.02.006
  • Journal Name: INTERNATIONAL JOURNAL OF REFRIGERATION-REVUE INTERNATIONALE DU FROID
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.963-972
  • Keywords: Refrigeration system, Vortex tube, Modelling, Neural network, Performance, Heating, Cooling, NUMERICAL INVESTIGATIONS, FEEDFORWARD NETWORKS, CRITICAL SUBMERGENCE, COMPRESSIBLE FLOW, ENERGY SEPARATION, INLET PRESSURES, NOZZLE NUMBERS, PREDICTION
  • Hacettepe University Affiliated: Yes

Abstract

In this study, the effect of the nozzle number and the inlet pressures, which vary from 150 to 700 kPa with 50 kPa increments, on the heating and cooling performance of the counter flow type vortex tube has been modeled with an artificial neural network (ANN) and multi-linear regression (MLR) models by using the experimentally obtained data. In the developed system output parameter temperature gradiant between the cold and hot outlets (Delta T) has been determined using inlet parameters such as the inlet pressure (P-inlet), nozzle number (N), cold mass fraction (mu c) and inlet mass flow rate ((m) over dot(inlet)). The back-propagation learning algorithm with variant which is Levenberg-Marquardt (LM) and Sigmoid transfer function have been used in the network. In addition, the statistical validity of the developed model has been determined by using the coefficient of determination (R-2), the root means square error (RMSE), and the relative absolute errors (RAE). R-2, RMSE and RAE have been determined for Delta T as 0.9989, 0.5016, 0.0540 respectively. Crown Copyright (C) 2010 Published by Elsevier Ltd and IIR. All rights reserved.