African Journal of
Agricultural Research

  • Abbreviation: Afr. J. Agric. Res.
  • Language: English
  • ISSN: 1991-637X
  • DOI: 10.5897/AJAR
  • Start Year: 2006
  • Published Articles: 6574

Full Length Research Paper

Modelling of a locally fabricated flat-plate solar milk pasteuriser using artificial neural network

  Francis O. Wayua1*, Michael W. Okoth2 and John Wangoh2    
  1National Arid Lands Research Centre, Kenya Agricultural Research Institute, P.O. Box 147 (60500), Marsabit, Kenya. 2Department of Food Science, Nutrition and Technology, University of Nairobi, P. O. Box 29053 (00625), Nairobi, Kenya.
Email: [email protected]

  •  Accepted: 06 September 2012
  •  Published: 31 March 2013

Abstract

 

The objective of this work was to develop an artificial neural network model to predict milk temperature of a locally fabricated solar milk pasteuriser, based on measures of error deviation from experimental data. A three-layer feed-forward neural network model based on back propagation algorithm was developed using the Neural Network Toolbox for MATLAB®. The inputs of the model were ambient air temperature, solar radiation, wind speed, temperature of hot water, and water flow rate through the collector, whereas the output was temperature of milk being pasteurised. The optimal neural network model had a 4-4-1 structure with sigmoid transfer function. The neural network predictions agreed well with experimental values with mean squared error, mean relative error and correlation coefficient of determination (R2) of 5.22°C, 3.71% and 0.89, respectively. These results indicate that artificial neural network can successfully be used for the prediction of the performance of a locally fabricated solar milk pasteuriser.

 

Key words: Artificial neural network, modelling, solar milk pasteuriser.