RESUMO
The relationship existing between operating parameters like speed of operation, applied pressure, feed rate and dependent variables like oil recovery and throughput capacity was investigated while mechanically expressing oil from kernels of Pentaclethra macrophylla. A 3-factor, 5-levels central composite design of response surface methodology was used for modeling and optimization of the process. The speed was varied over 15, 30, 45, 60 and 75 rpm using pulley arrangement, applied pressure was varied over 5, 10, 15, 20 and 25 MPa by adjusting the wormshaft distance of the oil expeller and the feed rate was varied over 100, 200, 300, 400 and 500 g/min by regulating the quantity of kernels fed into the expeller. Developed models were validated by comparing predicted values with experimental values. In optimizing the process, the oil recovery and throughput capacity were maximized while the independent variables were set at ranges. Optimum oil recovery of 73.2% and throughput capacity of 4.18 kg/h was obtained at 45 rpm speed of operation, 20 MPa applied pressure and 300 g/min feed rate. A quadratic polynomial model was developed for the oil recovery while a two-factorial-interaction model was developed for the throughput capacity. The results showed that speed of operation, applied pressure and feed rate had significant influence on the oil recovery and throughput capacity. The models developed were valid as they showed a good agreement between the variable due to low variations between calculated and predicted values.
