Using a combined neural network ─ genetic algorithm approach for predicting the complex rheological characteristics of microfluidized sugarcane juice

Abstract

The effect of multi-cycle multi-pressure microfluidization on the rheological properties of sugarcane juice has been reported for the first time in this work. Microfluidization pressure was varied in defined steps from 50 to 200 MPa with 1–7 cycles of processing. Rheological measurements were carried out at 30, 40 and 50 °C. Increase in microfluidization pressure up to 150 MPa showed an increase in consistency index. Beyond 150 MPa, an observed reduction in consistency index was attributed to the reduction in carbohydrate polymer chains, agglomeration, and molecular degradation. At low processing pressures, microfluidized sugarcane juice exhibited pseudoplastic behavior but showed a dilatant tendency at higher pressures (>150 MPa). Microfluidization increased the flow behavior index by 24–38% as compared to unprocessed sugarcane juice. No prominent trend in rheological data was observed with change in microfluidization cycles. Consistency index was reduced with increase in temperature and was found in the range of 1.5 × 10−3 ─ 5.3 × 10−3 Pa.sn whereas the flow behavior index was found in the range of 0.71 ─ 0.99. A GA mediated ANN model with 12 hidden layer neurons was found to predict the rheological properties of microfluidized sugarcane juice with a reasonable level of accuracy (R2 = 0.86–0.93).