基于动态仿真的复杂过程系统的定量多元HAZOP分析

发布时间：2019-01-26 11:48

【摘要】：危险与可操作性分析(HAZOP)是一种被广泛应用于过程危害分析的工具。以往的HAZOP分析利用不同专业的专家经验进行分析,近来,基于知识或模型的专家系统也被应用到此分析中。HAZOP需要占用大量的时间和资源,为了节约时间,节省经费,获得更全面的结果,计算机辅助HAZOP分析技术应运而生。基于知识的专家系统就是这样一种尝试,但是由于专家经验的有限性,结果展示的不全面。若仿真技术准确而广泛的应用,则基于模型的专家系统能够提供全面的分析结果。本论文将渣油催化裂化作为研究对象。渣油催化裂化在石油化工过程中是一个非常重要的化学过程,它是将高沸点长链烃类液体蒸发并断裂成小的、短的分子的过程。同时,它也是一个非常复杂的化工过程,许多变量之间相互影响,误差传递难以预测。本论文的目的是使HAZOP更少的依赖于专家系统,从而克服传统的HAZOP的缺点。本论文的方法是使用动态仿真为复杂的化学过程提供高保真的数据,从而对生产线的多级效应进行研究。定性分析通过观测数据变化的趋势和重构有向图和连接矩阵来监测过程的发展。它通过使用模拟获得的数据,给出在正常工作条件下和引入偏差效果下的设计意图数据。定量多级分析能够被进一步分成两部分,单变量引起的偏差和多变量引起的偏差。在第一部分中介绍两个单变量偏差,即由蒸汽增长引起的偏差和原油流率增长引起的偏差。这种偏差引起的产品质量变化的效果不足以采取补救措施,且这种效果随着生产线而消失。在第二部分,三个变量引起的偏差被结合起来考虑。例3和例4考虑了不同的引导参数与同一参数结合。第一种结合使偏差效果达到最小化,第二种结合使偏差增大,甚至引起了无计划的停工,而不是预期的暂时的产品质量的损失。常规HAZOP由于传递的本质,由本论文提供的方案结果很容易被忽略错过。这在工厂检修时候,是极其不利的,对设备和人员可能会造成经济损失和伤害。本论文精确地阐释了单变量与多变量偏差的多级效应,而且为HAZOP的定量分析提供了帮助,使它的分析结果更加全面精确,提高了由偏差引起的可能性危害的可预测性,以及时对偏差引起的后果进行补救。
[Abstract]:Hazard and operability analysis (HAZOP) is a widely used tool for process hazard analysis. In the past, HAZOP analysis made use of expert experience of different specialties. Recently, expert system based on knowledge or model has also been applied to this analysis. HAZOP needs to occupy a lot of time and resources, in order to save time, save money, In order to obtain more comprehensive results, computer-aided HAZOP analysis technology emerged as the times require. The expert system based on knowledge is such an attempt, but because of the limitation of expert experience, the result is not comprehensive. If the simulation technology is applied accurately and widely, the expert system based on model can provide comprehensive analysis results. In this paper, residuum catalytic cracking is studied. Residuum catalytic cracking is a very important chemical process in petrochemical process. It is the process of vaporizing high boiling point long chain hydrocarbon liquid and breaking into small and short molecule. At the same time, it is also a very complex chemical process, many variables affect each other, error transfer is difficult to predict. The purpose of this thesis is to make HAZOP rely less on expert system and overcome the shortcoming of traditional HAZOP. The method of this thesis is to use dynamic simulation to provide high fidelity data for complex chemical processes, so as to study the multilevel effect of production line. Qualitative analysis monitors the development of the process by observing data trends and reconstructing digraphs and connection matrices. By using the simulated data, the design intention data under normal working conditions and under the effect of introducing deviations are given. Quantitative multilevel analysis can be further divided into two parts: the deviation caused by single variable and the deviation caused by multivariate. In the first part, two single-variable deviations are introduced, that is, the deviation caused by steam growth and the deviation caused by the increase of crude oil flow rate. The effect of the product quality change caused by this deviation is not enough to take remedial measures, and the effect disappears with the production line. In the second part, the bias caused by the three variables is considered in combination. Examples 3 and 4 consider different boot parameters combined with the same parameter. The first combination minimizes the deviation effect and the second causes the deviation to increase and even causes the unplanned shutdown rather than the expected temporary loss of product quality. Due to the nature of transmission, the results provided by this paper can easily be ignored by conventional HAZOP. This is extremely unfavorable when the plant is being overhauled and may cause economic losses and injuries to equipment and personnel. In this paper, the multilevel effects of univariate and multivariate deviations are explained accurately, and the results of quantitative analysis of HAZOP are helpful to make its analytical results more comprehensive and accurate, and the predictability of possible hazards caused by deviations is improved. To remedy the consequences of the deviation in a timely manner.
【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】：TE624.41

【相似文献】