In this paper, dynamic models have been developed to predict the air temperature, specific humidity and drying rate in an industrial seed grain dryer. An industrial dryer was utilised for experimental measurements. The dyer was modelled as a system of serial cells characterised by heat and mass transfer with air back-mixing. Model equations were solved numerically based on the Levenberg–Marquardt algorithm in MATLAB. The model response was optimised by signal matching technique. The sensitivity of the model parameters on the prediction accuracy was assessed using Monte Carlo simulation tests. The accuracy of the models was statistically checked using the coefficient of determination (R2), the root mean square error (RMSE), and the probability value (p value) approach. The results revealed a good agreement between the measured data and model predictions with the highest mean relative deviation (MRD) of 2.37% and p value of 0.0298 at 5% significance level. The model accuracy was highly sensitive to the parameter that defines the air residence time inside the drying bin. The dryer exhibited air back-mixing level of 45.8% and moisture decay rate constant (k) of 0.028 per hour. The convective heat transfer between the air and seed grain was determined as 0.96 kJ/h m2 °C. The dynamic models developed here can adequately predict the outlet air temperature, specific humidity and solids temperature and subsequently the drying curve. The models could help to provide real-time insights of drying characteristics, which is necessary for the achievement of a stable and effective control of the drying process.
Read full abstract