MODELING OF EXTRUSION PROCESS USING RESPONSE SURFACE METHODOLOGY AND ARTIFICIAL NEURAL NETWORKS

Artificial neural networks are a powerful tool for modeling of extrusion processing of food materials. Wheat flour and wheat-black soybean blend (95:5) were extruded in a single screw Brabender extruder with varying temperature (120 and 140 degrees C), dry basis moisture content (18 and 20%) and screw speed (156, 168, 180, 192 and 204 rpm). The specific mechanical energy, water absorption index, water solubility index, expansion ratio and sensory characteristics (crispness, hardness, appearance and overall acceptability) were measured. Well expanded products could be obtained from wheat flour as well as the blend of wheat-black soybean. The results showed that artificial neural network (ANN) models performed better than the response surface methodology (RSM) models in describing the extrusion process and characteristics of the extruded product in terms of specific mechanical energy requirement, expansion ratio, water absorption index, water solubility index as well the sensory characteristics. The ANN models were better than RSM models both in case of the individual as well as the pooled data of wheat flour and wheat-black soybean extrusion.