间歇结晶过程粒径分布的建模与控制

发布时间：2018-02-27 23:42

本文关键词： 粒径分布数粒平衡方程状态观测器低阶矩网络模型迭代学习控制　出处：《北京化工大学》2015年硕士论文　论文类型：学位论文

【摘要】：近年来,化工行业和医学行业的飞速发展,使得结晶过程面临着巨大的挑战和机遇。为了满足工业日益发展的需要,人们对晶体的形貌、粒度和分布的要求也越来越高。因此,实现结晶过程的在线粒径分布控制是企业取胜的关键性所在,然而目前对粒径分布的直接在线测量缺乏有效的手段,如何实现粒径分布的在线测量并达到优化控制的目的成为结晶过程亟待解决的问题。本文致力于结晶过程粒径分布的在线控制研究,来提高终点产品质量。本文首先介绍了结晶过程的反应机理,包括成核速率模型、生长速率模型、溶解度模型、物料衡算模型和数粒衡算模型的关系和原理。在此基础上,本文研究了结晶模型的数值求解方法、建立结晶过程的机理模型,分析不同的操作变量对粒径分布的影响。由于结晶过程的粒径分布不可测量或需要花费昂贵的代价,通常采用粒径分布的矩值代替粒径分布。针对矩模型不能直接观测晶体分布的问题,本文在矩模型中应用状态估计的方法获得粒径分布,通过观测矩值的变化估计溶液浓度的变化,进而实现直接对粒径分布的观测,对整个动态过程的研究及控制起到了重要的作用。结晶过程的机理模型通常是在理想情况下建立的,精度有时候难以满足要求。本文采用数据建模的方法,求得结晶过程分布与矩的关系,建立操作变量与低阶矩之间的网络模型,降低建模难度。然后根据给定的优化性能指标,推导出自适应迭代学习控制率对粒径分布的低阶矩进行控制,从而实现对粒径分布的间接控制。最后将该方法应用到草酸钴结晶过程中,仿真结果证明了该方法的可行性。
[Abstract]:In recent years, with the rapid development of chemical industry and medical industry, the crystallization process is facing great challenges and opportunities. The key to the success of enterprises is to realize the on-line particle size distribution control in crystallization process. However, the direct on-line measurement of particle size distribution is lack of effective means at present. How to realize the on-line measurement of particle size distribution and achieve the purpose of optimizing control become the urgent problem of crystallization process. This paper is devoted to the research of on-line control of particle size distribution in crystallization process. Firstly, the reaction mechanism of crystallization process was introduced, including nucleation rate model, growth rate model, solubility model, material balance model and particle balance model. In this paper, the numerical solution method of crystallization model is studied, the mechanism model of crystallization process is established, and the influence of different operating variables on particle size distribution is analyzed. The particle size distribution is usually replaced by the moment value of the particle size distribution. In view of the problem that the moment model can not directly observe the crystal distribution, the particle size distribution is obtained by using the method of state estimation in the moment model. The change of solution concentration is estimated by the change of observed moment, and the observation of particle size distribution is realized directly. It plays an important role in the study and control of the whole dynamic process. The mechanism model of the crystallization process is usually established under ideal conditions, and the precision is sometimes difficult to meet the requirements. In this paper, the method of data modeling is used. The relationship between the distribution of the crystallization process and the moment is obtained, and the network model between the operation variables and the lower moments is established to reduce the difficulty of modeling. The adaptive iterative learning control rate is derived to control the lower moment of particle size distribution, and the indirect control of particle size distribution is realized. Finally, the method is applied to the crystallization process of cobalt oxalate, and the simulation results show that the method is feasible.
【学位授予单位】：北京化工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TQ026.5

【参考文献】