节点文献

石化炼油生产原油库存的动态管理

【作者】 刘志军

【导师】 伍乃骐;

【作者基本信息】 广东工业大学 , 机械电子工程, 2004, 硕士

【摘要】 综合自动化是流程工业的重要发展趋势,目前在底层控制层和上层企业资源管理层取得了突破性的进展,但对中间层生产过程管理与生产过程优化,国内外都没有取得重大突破。石化炼油是一个流程工业,原油的流动是连续的。为支持企业营运和生产管理的需要,对原油库存信息进行动态管理是非常必要的。这就要求原油库存管理能提供任何时刻的库存状态,这是一个极大的挑战。 文中以典型的设备布局的炼化企业为研究对象,首先阐述了综合自动化领域的发展状况,存在的问题,及其研究的意义。在此基础上,明确了原油库存动态管理对炼化企业的经营决策和生产管理的重要意义。 为实现原油库存的动态管理,本文解决了两个关键性的问题。其一是解决了连续系统中数据处理问题,为此本文提出了连续问题离散化的方法,通过这个方法,只需要记录本文中定义的时间与事件点的库存信息,保证了数据录入方法的可行性和数据的完整性,从而解决了连续系统中海量数据的采集、存储与处理等问题;其二是在获取数据的基础上如何实现原油库存的动态管理,由此本文又提供了用于实现炼化企业原油库存动态管理的仿真算法,通过每日结束时刻库存仿真算法可获得每日结束时刻的库存状态,通过任意时刻库存仿真算法可以获得任意时刻的库存状态,这两个算法相互补充,高效地地解决了炼化企业原油库存的动态管理问题。文中还提供了一个库存动态管理的开发实例,连续问题离散化的方法和库存动态管理仿真算法在该系统中得到应用,并很好的实现了原油库存的动态管理,其中动态库存管理在计划与调度过程中也是必要的。本文最后对炼化企业的生产计划与调度问题的发展作了一个展望。

【Abstract】 Nowadays, integrated automation is an important issue for competitiveness in process industry. Effective techniques for management at the high level of corporation have been developed and distributed control system for the low level has been implemented in most process plants. However, it lacks techniques for the middle level. Thus, it is very important to develop techniques for that level. Refining industry is a typical process industry and is characterized by hybrid properties with both discrete event process and continuous material flow. In order to support dynamic production mamagement for such a system, dynamic information of crude oil inventory is necessary and inventory status at any time point is required. This is a great challenge.In this paper, we address the dynamic management of crude oil inventory. We first review the research and development status in this field and analyze the problems to be solved. Then, the goal of this paper is presented and the work is done based on a case problem, a typical refinery.To make dynamic management of crude oil inventory possible, two questions must be answered: 1) what dada should be stored, and 2) how to recover the status at any time points. To answer the first question, a method is developed to store the inventory data at only some discrete time ponts. These time points are determined by time interval and events defined in this paper. For the second question, we use soft computing to recovery the state at any time point. The method proposed has been realized for the case problem, it works well.

  • 【分类号】F407.22
  • 【被引频次】3
  • 【下载频次】297
节点文献中: 

本文链接的文献网络图示:

本文的引文网络