Modelling the convective drying process of pumpkin (Cucurbita moschata) using an artificial neural network

This study investigated the drying kinetic of pumpkin under different drying temperatures (50, 60, 70 and 80°C), samples thickness (3, 4, 5 and 7mm), air velocity (1.2m/s) and relative humidity (40 - 50%). Kinetic models were developed using semi-theoretical thin layer models and multi-layer feed-forward artificial neural network (ANN) method. The Hii et al. (2009) semi-theoretical model was found to be the most suitable thin layer model while two hidden layers with 20 neurons was the best for the ANN method. The selections were based on the statistical indicators of coefficient of determination (R2), root mean square error (RMSE) and sum of squares error (SSE). Results indicated that the ANN demonstrated better prediction than those of the theoretical models with R2, RMSE and SSE values of 0.992, 0.036 and 0.207 as compared to the Hii et al. (2009) model values of 0.902, 0.088 and 1.734 respectively. The validation result also showed good agreement between the predicted values obtained from the ANN model and the experimental moisture ratio data. This indicates that an ANN can effectively describe the drying process of pumpkin.

Preview PDF (34) IFRJ-16272 Hashim.pdf Download (506kB) | Preview Official URL or Download Paper: http://www.ifrj.upm.edu.my/23%20(06)%202016%20supp...

Actions (login required)

View Item