A cost function for continuous-flow grain drying was developed , which relates grain moisture content , grain temperature and thermal damage , fuel used for drying and elapsed time to the cost of drying per tonne of output grain . A simplified model of a grain dryer was developed for rapid calculation of these parameters during simulation of unsteady-state drying .
The cost function and dryer model were used in a long-range predictive control algorithm which , for each control action of an existing feedforward-plusfeedback system , determined and applied nearoptimal values of target moisture content and drying air temperature . The complete system was tested using computer simulation of a mixed-flow dryer to represent a real dryer , and the overall cost of drying was compared with that given by a system with the lowest cost combination of a fixed target moisture content and drying air temperature . The system demonstrated a well-behaved response over a range of input grain moisture content from 0 ΠΈ 20 to 0 ΠΈ 30 dry basis and showed an 18% saving in time and a cost saving of 0 ΠΈ 30 Β£ t Οͺ 1 . The cost incurred by grain recirculation and the problem of control instability caused by recirculation were also investigated .