Premium
Improving black tea quality through optimization of withering conditions using artificial neural network and genetic algorithm
Author(s) -
Das Shrilekha,
Samanta Tanmoy,
Datta Ashis K.
Publication year - 2021
Publication title -
journal of food processing and preservation
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.511
H-Index - 48
eISSN - 1745-4549
pISSN - 0145-8892
DOI - 10.1111/jfpp.15273
Subject(s) - polyphenol oxidase , artificial neural network , moisture , water content , peroxidase , relative humidity , biological system , chemistry , computer science , materials science , artificial intelligence , engineering , meteorology , enzyme , biology , biochemistry , composite material , physics , geotechnical engineering
Polyphenol oxidase and peroxidase activity and moisture loss have been used for optimization of withering parameters in black tea production. Leaves of TV25 clone were withered at different air temperature (25, 30, 35°C) and relative humidity (RH) (60, 75, 85%), with constant air flow rate. With higher air temperature, Polyphenol oxidase specific activity decreased, whereas peroxidase specific activity was enhanced. A multilayered feed forward artificial neural network, with single hidden layer, was trained and validated using experimental data based on mean square error and coefficient of determination ( R 2 ). Optimized withering parameters for both crush‐tear‐curl (CTC) and orthodox tea production were obtained using Genetic Algorithms to maximize enzyme activity and bring down moisture content within desired range. A penalty function based on distance from feasible region was incorporated in Genetic Algorithm. Optimized parameters are: CTC – temperature: 29.13°C, RH: 79.58%, duration: 11.41 hr; Orthodox – temperature: 33.04°C, RH: 77.27%, duration: 10.60 hr. Novelty impact statement Polyphenol oxidase and peroxidase enzyme activity has been used as optimization criteria, along with moisture loss, for selection of withering parameters. Soft computing methods like Artificial Neural Network and Genetic Algorithms have been employed for modeling and optimization of the nonlinear relation between withering parameters and enzyme activity and moisture content with penalty function for constrained optimization of moisture content in leaves. In this study, optimum withering conditions in terms of air temperature, relative humidity, and duration were found out to obtain maximum enzyme activity and desired moisture content in leaf for producing better quality tea.