Modelling the Drying Properties of Tomato in a Hot-Air Dryer Using Hybrid ANN-GA Technique

J.B. Hussein, M.O. Oke, R.I. Ajetunmobi and F.F. Agboola

Background and Objective: Drying, a simultaneous heat and mass transfer method, is an essential process to reduce post-harvest losses of tomatoes. Thus, knowledge about the drying properties of tomatoes is necessary for designing and optimizing the drying systems. The study, therefore, investigated the modelling of the drying properties of tomatoes in a hot-air oven using the hybrid ANN-GA technique. Materials and Methods: The tomatoes were pretreated in water blanching (WAB), ascorbic acid (ASA) and sodium metabisulphite (SMB). After that, sliced into 4, 6 and 8 mm and dried at 40, 50 and 60°C air temperatures following the Taguchi experimental design. The drying properties (effective moisture diffusivity (D_eff), activation energy (E_a) and specific energy consumption (SEC)) of the dried tomatoes were determined and modelled each by hybrid ANN-GA. The highest and lowest values of correlation coefficient (R) and mean square error (MSE), respectively were used as the stopping criteria for the developed model, while R², RMSE and MAE were used to validate the reliability of the ANN-GA hybrid network. Results: The result shows a variation of 0.98×10^–10 to 6.36×10^–10 m² sec^–1 for D_eff, 12.23 to 25.76 kJ mol^–1 for E_a and 0.6247 to1.9514 kWh kg^–1 for SEC. The results of hybrid ANN-GA (R² = 0.9934, RMSE = 1.83×10^–11, MAE = 1.54×10^–11 for D_eff and R² = 0.9587, RMSE = 0.0656, MAE = 0.0501 for SEC) proved it that is capable of better prediction accuracies and generalization capability. Conclusion: The results found in this study can serve as an operational guide for drying tomato fruit on both pilot and industrial scales.

View Fulltext Back