OPTIMAL SETTINGS OF A MELON SEED SHELLER USING STATISTICAL DESIGN OF EXPERIMENTS

Abstract

A community melon seed shelling machine, used by farmers and households to reduce the constraints associated with the manual processing, was found to produce larger quantities of unshelled seeds. In particular, the optimum seed moisture content and motor speed were unknown. The objective of this study was to determine the appropriate motor speed and moisture content of melon seeds to improve the effectiveness of seed dehulling. A full factorial design approach was used to develop a regression model and investigate the effect of moisture content and machine speed on dehulling efficiency. The results of the experiments showed that both motor speed and seed moisture content had a significant effect on seed shelling efficiency. The motor speed had the greatest effect (0.21567), followed by the moisture content (0.19732). The modelled optimization conditions were as follows: seed moisture content at 26%, motor speed at 2100 rpm. The regression model showed a coefficient of determination value of 0.999, indicating that 99.90% of the variation in response could be explained by the model; only 0.1% of the variation in response could not be explained by the model and was random. Three experimental validation runs were carried out under optimal conditions, and the highest average yield was 86.2%, corresponding to a significant improvement of 26% on the initial yield of 60 %. Other experimental design methods should be investigated to improve the results for appropriate parameter settings of the production process.