啤酒发酵过程的建模仿真与控制

针对啤酒发酵过程的特点，建立了数学模型，设计了多变量温度控制系统，并提出一种面向工业过程的多变量时滞系统仿真方法。该控制系统已在自行研制的AUTO－2000集散控制系统中实现，并成功用于安徽圣泉啤酒厂的啤酒发酵温度控制，现场运行情况表明，系统可靠性高，适应性强，有效提高了啤酒的质量和产量，取得了每年1千多万元的经济效益。

Modeling Simulation and Control for Beer Fermentation Process

Beer fermentation is the key process of beer produ ct ion. The output and quality of beer depends directly on whether or not the varia tion of temperature and pressure in fermentation process satisfies technological requirements. Thus, it is very important to improve production automation effic iently. Mechanism analyses and tests show that the fermentation pot has great ti me delays, strong interaction and time variant. The author s gained its mathematical model by square wave modeling and least square identif ication, and established the fermentation temperature multivariable control syst em according to the features of beer fermentation process. After compensating fo r the system with time delays by multivariable Smith Predictors, a precompensato r was designed using auto weighted pseudo dominance algorithm of Nyquist Array method so that the system was diagonal dominance. A main controller matrix by si ngle variable control theory was then designed in order for the system to have satisfactory properties. Lastly, the simulation method for a multivariable system with time delays oriented to industrial process is presented, and the res ulting controller parameters are optimized by simulation. This system is implemented in the AUTO 2000 Distributed Control Systems develop ed and successfully applied to beer fermentation temperature control in Sheng Q uan Brewery, Anhui. It decreases the effect of time delays, but makes the upper and middle temperatures of the pot satisfy the technological requirements simult aneously. It improves output and quality of beer efficiently, thus gaining 10 mi llion yuan and more in profits every year. The practical operation shows that th e control system has high reliability and good robustness. It also pro ves that the modeling simulation and control method presented in this paper is e ffective,thus it can be used to control a variety of multivariable industrial processes with great time delays and strong interaction.