Development of super‐atmospheric oxidation chamber for orthodox tea processing and its validation through neural network approach.

BACKGROUND Orthodox tea is known for its distinct aroma and superior quality. However, the oxidation step that is most crucial for developing these attributes needs precise control and is also time consuming. In the present study, a super‐atmospheric oxidation chamber (SOC) was designed to address this issue. Oxidation of hand‐rolled tea leaves was performed at an elevated oxygen concentration inside the SOC. Both pressure and processing time were varied over the ranges 101–303 kPa and 0–150 min, respectively. Theaflavin (TF), thearubigin (TR) and total color (TC) were determined experimentally and then predicted through artificial neuron network (ANN) modeling. RESULTS: An increasing oxygen concentration and attainment of 100% relative humidity in the compressed air was recognized at elevated air pressure. The observed rate of formation or depletion in TF, TR and TC values was faster with compression of the ambient air. The experimental data points were predicted based on an ANN network containing 18, 10 and 16 neurons in the hidden layer since lowest mean square error and mean absolute error values were observed for the given experimental TF, TR and TC values respectively. CONCLUSION: Considering the significant decrease in oxidation duration and minimal mechanized complexity of the design, the set‐up developed in the present study can be upscaled for industrial application in orthodox tea processing. © 2019 Society of Chemical Industry [ABSTRACT FROM AUTHOR]