Stochastic modeling and automatic control of grain dryers: Optimizing grain quality

Quality damage to grain during high-temperature drying has become a severe problem for end users in many grain-processing industries. The problem can not be solved simply by decreasing the drying-air temperature because this will decrease the dryer capacity and the energy efficiency. The major objective of this research is to minimize the quality damage through improving the dryer design and control.; The objective was pursued in two directions, i.e. modeling of the drying process and developing of the control system. Stochastic maize drying and maize-quality models were developed, providing the necessary tools for optimizing dryer design; and, automatic moisture and quality controllers were developed to minimize grain-quality deterioration. Three aspects of maize quality were investigated, i.e. moisture distribution among grain kernels, protein denaturation and viability.; The stochastic nature of the moisture content of maize before and after drying was analyzed. The moisture content distribution of maize kernels at harvest and after thin-layer drying, appeared to be normally distributed, and after crossflow-drying to be skewed. Stochastic models correctly predicted the moisture content distribution in maize kernels dried in different high-temperature dryer types, i.e. the average moisture content and its standard deviation of the dried grain.; The stochastic crossflow drying model, and the denatured protein and germination models, were combined to analyze the quality of individual kernels. The results served as a guide for operating crossflow dryers.; For the purpose of automatic moisture control, a distributed-parameter process model of crossflow drying was utilized to develop a model-predictive control system. The controller has a feedforward loop (i.e. the predictive model and the optimizer) and an feedback loop (i.e. the parameter estimator and the modifier). It was implemented and tested on a commercial crossflow dryer and controlled the moisture content of the maize at the outlet of the dryer to within ±1.3% of the set point, at inlet moisture contents ranging from 21 to 32% w.b. and drying-air temperatures from 85 to 120°C. The controller reacted properly to changes in the drying-air temperature.; The strategy of grain-quality control was analyzed. A neural network model was developed relating the quality of dried grain to various temperature/residence-time conditions. A control algorithm was selected which to varies the drying-air temperature with the objective of maintaining the grain quality close to the set point. Simulation tests showed that a more uniform grain quality is achieved by implementing the quality controller on a crossflow dryer.