Optimization of process parameters using a hybrid intelligent system model and evaluation of physicochemical properties of microwave roasted Chironji (Buchanania lanzan) kernels

Abstract

Microwave (MW) roasting of Chironji kernels was carried out at different combinations of microwave power and roasting time. The impact of MW roasting on bulk density, hardness, water absorption capacity (WAC), water activity (WA), browning index (BI), total phenolic content (TPC) and antioxidant activity (AA) were evaluated. Significant reduction in bulk density, hardness, and WA was noticed upon roasting, while a substantial increase in WAC, BI, TPC, and AA was found. A hybrid intelligent system model was used for achieving the optimal conditions of MW roasting parameters. The model comprised of several steps such as a general full factorial experimental design, computation of Taguchi’s quality loss function, principal component analysis, grey relational analysis, and process optimization using combined artificial neural network—genetic algorithm technique. Under optimal condition of 700 W power and 5.5 min of roasting time, decrease in bulk density of 8.02 ± 0.15%, hardness of 16.63 ± 0.17 N, WAC of 1.02 ± 0.06 g/g, WA of 0.0541 ± 0.0009, BI of 76.79 ± 1.23, TPC of 26.74 ± 0.32 mg GAE/g, and AA of 73.23 ± 0.86%, were recorded for roasted Chironji kernels. Further, the SEM analysis confirmed microstructural changes in the kernels roasted at optimal conditions.