Artificial neural networks (ANNs): A new paradigm for the study of drying kinetics and sorption isotherm

Abstract

Artificial neural networks (ANNs) have been used for a wide variety of application areas in science and engineering. ANN has become well recognized powerful tool for academicians
and scientists as the technique provides practical advantages such as adaptability, learning and training potential, and real time processing over conventional empirical models. Empirical models and Artificial Neural Networks (ANNs) have been employed for prediction of Equilibrium Moisture Content (EMC), study of drying kinetics, and sorption isotherm in some plant materials and other agro-industrial products. Understanding of temperature and relative humidity alongwith the moisture sorption phenomena of agro-industrial products provides vital information for the modifications
associated with thermodynamics of a system. Isotherm characteristics of natural products are used for designing, modelling and optimizing post harvest procedures such as drying, aeration and storage. Thus the moisture sorption characteristics, in terms of equilibrium moisture contents (EMC), enthalpy and entropy of heat sorption of natural products are imperative to be investigated. The optimization
of drying conditions, simulation and modeling of sorption isotherms can be studied using various non-linear regression, EMC models and the multilayer artificial neural network (ANN) approach. In this paper, we have discussed the applicability of ANN for estimation of EMC in plant materials and
other agro-industrial products and have also presented a comparative view of ANN technique and different
empirical models including GAB, Smith, Henderson, Oswin, Halsey and D’Arsy-watt models for drying kinetics and sorption isotherm studies.