| The dry grind process is used to produce about two thirds of total U.S. fuel ethanol. The simultaneous saccharification and fermentation (SSF) is the most important unit operation in the dry grind corn process. At present, static controllers that do not respond to changing conditions during the SSF process are used to control SSF in most dry grind plants. Use of optimal controllers based on continuous process measurements can provide performance improvements by dynamically optimizing the fermentation conditions. A controller for the SSF process was developed based on models for starch hydrolysis, yeast metabolism and metabolic flux analysis for yeast. To optimize the SSF process, the controller utilized temperature, pH, glucose, maltose, maltotriose, glycerol, ethanol and organic acid measurements to regulate the temperature, pH and glucoamylase dose.; Starch structure (amylose and amylopectin) was modeled using a cluster model. Modeled molecules were used to simulate liquefaction and saccharification processes. A cybernetic model for yeast metabolism was developed. The yeast metabolism model simulations over a range of temperature, pH, organic acids, initial inoculum levels and initial glucose concentrations conformed to hypothesized trends and to observations from other researchers. Simulations converged to expected results and exhibited continuity in predictions for all ranges of variables simulated. The control problem was formulated as a scalar performance criteria minimization and was solved using an iterative algorithm based on steepest descent technique. A fermentation system was built and calibrated.; The controller maintained glucose concentration of <2.0% w/v in the fermenter throughout the SSF process. Reduced glucose concentrations minimize the osmotic shock to yeast and inhibit the growth of other competing microorganisms. Use of the optimal controller resulted in 50% reduction in glucoamylase amount required for the SSF process under varying operating conditions as compared to the standard SSF process. Optimal controller significantly improved final ethanol concentrations as compared to the conventional process without optimal controller under conditions of temperature and pH disturbances. Use of the optimal controller in conventional dry grind ethanol processes can result in estimated cost savings up to {dollar}1 million for a 151 million L/yr (40 million gal/yr) dry grind plant. |
