面向需求有界变化的石化生产过程优化与控制方法研究

Due to the variation of the requirement of petrochemical production market, the more targeted optimization and control method has been an urgent need to adapt to the petrochemical producing processes. This proposal aims to design a novel framework for dealing with the bounded requirement fluctuation from the points of modeling, interval optimization and real-time control method. First of all, based on the technology of interval analysis, with the known bounded information, the original process model is reorganized to establish a new dynamic model. Furthermore, the interval optimization of product process will be proceeded. The quick solving method for the process with complex constrain conditions is proposed. And the feasible optimal zone is achieved. Finally, considering the real constrain of equipment, the product market bounded variation is integrated into the cost functions. The real-time optimization and control of producing process is implemented to deal with the real-time market requirement fluctuation, and to find the new trade-off point between the efficiency of producing process and the maximum profit of production. This study will be efficiently devoted to solve the process problem due to the petrochemical market requirement fluctuations, which provide the strong support to the theory and engineering application. The achievements of this proposal will also be more important for the optimization and energy-saving technology of petrochemical product processes.

面对石油化工类产品市场需求的波动现状，石化生产过程迫切需要开发出更加有针对性的优化和控制解决方案。本课题将从系统建模、区间优化和实时控制的角度设计一种能够有效处理需求有界变化的过程优化与控制方法。首先，利用需求变化的上下界信息，结合原有过程模型，建立以区间分析技术为核心的有效反映需求变化的新模型。其次，对生产过程进行区间优化，提出处理带有复杂约束的优化问题快速求解方法，得到最优的可行区间。最后，结合实际的生产装置约束条件，把石化产品市场需求的变化信息融入到控制性能指标中，设计并实现对生产过程的实时优化和控制，以期达到实时反映产品需求的有界波动情况，并能够在生产效率和产品利润之间找到更好的平衡。本课题的研究将为有效解决当前石油化工类产品需求波动带来的生产过程运行问题提供有力的理论支撑和技术应用基础，对于石化生产过程的优化与节能技术研究也具有重要的学术价值。

面对工业产品市场需求的波动现状，工业生产过程需要更加有针对性的优化和控制方法。本项目面向需求有界变化，开发了复杂工业生产过程的建模、优化和控制方法。主要研究内容有：1）针对复杂工业生产过程的不确定性和信息不完整情况，研究了非对称信息的自组织区间二型模糊神经网络建模方法和基于神经网络和自适应加权融合的建模方法等。2）针对工业生产产品市场需求经常发生变化的问题，面临的优化问题是多目标的，结合建立的模型研究了非策略博弈Q-学习优化方法和多目标群体仿生智能优化算法等。3）针对工业生产过程经常遇到不确定性、干扰、时滞、故障等问题，严重地危害了系统的控制效果，导致产品的质量下降，采用模型预测控制、迭代学习控制等研究了可靠的工业生产过程先进控制方法。4）将理论研究成果用于生产实际，在乙烯裂解炉装置上进行应用，取得了很好的控制效果。本项目发表学术论文30篇，其中SCI论文22篇，获授权发明专利3项，软件著作权3项，省级科技成果奖励2项，培养博士研究生4名，硕士研究生3名，青年教师4名。