MODELLING OF ADSORPTION REFRIGERATION MACHINE BY ANN

摘要

This paper proposes a new approach for modeling the solar-driven adsorption refrigerator system withrnActivated carbon/methanol working couple. The use of artificial neural networks (ANNs) has beenrnproposed to predict the coefficient of performance (COP) using many inputs related to the geometry,rncouple adsorbent/adsorbate, meteorological parameters, etc.rnThe statistical coefficient was near to r ≈1 for the COP with minimal error, these accuracies are acceptablernwhich means, the ANN is capable to predict the COP of any prototypes before design. The present methodrngreatly reduces the time required by design engineers to find the optimum solution and in many casesrnreaches a solution that could not be easily obtained from simple modeling programs. The importance ofrnthe ANN approach, that it is possible to find solutions that make solar-energy applications more viable andrnthus more attractive to potential users.