Model development and optimization for predicting whole kernel recovery of a palm nut cracking machine using mixed variety of palm nuts

Abstract

This work aimed at developing a mathematical model, validate and optimized for predicting whole kernel recovery as a performance parameter for a palm nut cracking machine, using mixed variety of palm nuts. The drying temperature of 105°C; five levels of cracking speed (1200, 1400, 1600, 1800 and 2000rpm); six levels of moisture content (12.4, 14.0, 16.2, 18.1, 20.0 and 27.8% w.b) and five levels of feed rate (360, 400, 450, 514.29, 600kg/h) were selected for the study. Full factorial design was used in the design of the experiment with three independent variables; cracking speed (rpm), moisture content (%) and feed rate (kg/hr), set at five, six, five level factorial respectively. Four hundred and fifty (450) samples of 1000g each of fresh palm nuts were used for the experiment. 150 samples were selected for each experimental run. Model to predict whole kernel recovery was developed using the concept of Buckingham Pi theorem. The developed models were verified and validated by fitting them into experimental data. Method of regression analysis as computed using Microsoft Excel programme of Microsoft package was used to describe the relationships, plot the graphs and compute the coefficients of determination (R² ) between the predicted and experimental values. The best combination for optimum whole kernel recovery for mixed varieties were obtained at feed rate of 580.41kg/hr, throughput capacity of 151.43kg/h, nut speed of 12.51m/s, peripheral velocity of 33.78m/s, cracking speed of 1935rpm, and moisture content of 16.69%w.b with optimum whole kernel recovery of 80.45%.