Deep bed drying of malt

Abstract

During the production of malt, germinated barley is dried from 70 - 85% to 2.4 - 4% moisture content (d.b.) requiring high energy and a long drying process. This study is concerned with developing a mathematical simulation model for deep bed drying of malt to provide design data and data on temperature and moisture content histories at different depth of the bed for use by the design engineers and maltsters for minimizing energy consumption and improving quality. In order to do this certain physical and thermal properties were also determined. The properties of physical dimensions, 1000 grain weight, bulk density and dry bulk density were determined experimentally and each of these properties except bulk density was found to be linearly dependent on moisture content. Specific heat was also determined experimentally and also found to be linearly dependent on moisture content. The ratio of latent heat of malt to latent heat of free water was determined from existing experimental data on equilibrium moisture content and found to be a function of moisture content. Shrinkage of the mal t bed was de termined experimentally and found to be a non-linear function of moisture content reduction. The heat transfer coefficient of the malt bed was also determined experimentally and found to be a function of air flow rate. The thin layer drying experiments were conducted under controlled conditions of the drying air using an automatic continuous weighing system. Numerical procedures were developed to fit the experimental data to the single exponential equation, the Page equation and the double exponential equation. The drying constant was expressed as a function of the drying air temperature and the dynamic equilibrium moisture content was expressed as a function of drying air temperature and relative humidity. Two mathematical models, model 1 similar to Newcastle University model and model 2 similar to the Michigan State University model were programmed in FORTRAN to predict temperature and moisture content changes with time and position. Shrinkage and variable dry bulk density effects and a simple comprehensive condensation procedure were incorporated in both models. The set of partial differential equations for model 2 was solved by an exponential approximation and using central difference values updated by an iteration. Five deep bed experiments were conducted in the laboratory and the suspended thermocouple technique for the measurement of temperature was found quite satisfactory. Both models agreed well with the experimental data. Reasonable agreement was also obtained with commercial kiln data. Practically there is no difference between the predictions of the two models and model 1 takes less computing time. A short section 1S included to illustrate the use of the programme for simulation. Typical commercial conditions are used and the effect of gas fired, indirect fired and recirculating gas fired conditions is examined.