Evaporative Cooler Climate Prediction using Artificial Neural Network

  • Ibrahima Sow Pan African University Institute of Basic Sciences, Technology and Innovation (PAUSTI), Nairobi, Kenya
  • Evan Murimi Jomo Kenyatta University of Agriculture and Technology (JKUAT), Nairobi, Kenya
  • Urbanus Mutwiwa Jomo Kenyatta University of Agriculture and Technology (JKUAT), Nairobi, Kenya

Abstract

Fruits and vegetables are highly perishable and need to be stored in a controlled environment in order to elongate their shelf-life. In the tropics, some areas experience high ambient air temperature with corresponding low relative humidity thus making it possible to apply the principle of evaporative cooling in managing the temperatures in agricultural structures. Predicting the internal climate in an evaporative cooler is important so as to enhance the environmental conditions thus increase the shelf life of fruits and vegetables. However, this prediction is a complex operation. In this paper a neural network approach based on non-linear auto regressive model with exogenous input (NARX) has been developed for predicting internal temperature and relative humidity of an evaporative cooler. The resulting models driven by the Levenberg-Marquardt back propagation algorithm showed high performance for the prediction of the internal variables of the cooler, which makes it possible to avoid the loss of the products a few days after the harvest. The prediction models will be used for control purposes.
Published
Mar 11, 2020
How to Cite
SOW, Ibrahima; MURIMI, Evan; MUTWIWA, Urbanus. Evaporative Cooler Climate Prediction using Artificial Neural Network. JOURNAL OF SUSTAINABLE RESEARCH IN ENGINEERING, [S.l.], v. 5, n. 3, p. 113-127, mar. 2020. ISSN 2409-1243. Available at: <http://sri.jkuat.ac.ke/ojs/index.php/sri/article/view/808>. Date accessed: 05 july 2020.

Most read articles by the same author(s)

Obs.: This plugin requires at least one statistics/report plug-in to be enabled. If your statistics plugins provide more than one metric then please also select a main metric on the admin's site settings page and/or on the journal manager's settings pages.