基于复杂网络的化工过程SDG模型研究

发布时间：2018-07-10 15:12

本文选题：故障诊断 + 化工过程　；参考：《青岛科技大学》2015年硕士论文

【摘要】：随着化工流程规模日益增大和化工生产装置间关系更趋复杂,化工生产操作弹性逐渐减弱,高温高压、低温真空、有毒腐蚀等苛刻条件相对突出。为了确保化工生产过程安全顺利的进行,化工过程故障的识别与诊断必不可少。基于符号有向图(SDG)的化工故障诊断理论及应用研究方兴未艾,同时,此理论方法在建模复杂、整体性不高、易于漏掉重要变量等方面体现出一定程度的欠缺。目前,复杂网络理论已经在电网、交通等领域的故障诊断方面开展了相关研究,是处理过程复杂性、多变性的有效手段之一。本文试图在SDG化工网络诊断理论研究的基础上,利用复杂网络理论对故障网络进行关键节点的识别和社团结构的定量划分,以期解决SDG化工网络诊断中面临的不足,研究工作主要体现在以下几方面：首先,针对故障诊断中符号有向图的建模方法,对传统符号有向图中的流程图建模方法进行了改进,并提出了构建基于SDG模型的网络邻接矩阵建模的新方法。其次,提出了基于复杂性网络与SDG结合的化工过程诊断中关键节点的识别模型和方法,并在对化工SDG网络进行复杂性网络特征的计算和判断、中心性计算和社团划分理论的基础上,对其进行关键节点的识别以及重点保护区的划分,并根据关键节点和社团划分的结果,绘制出其关键节点及路径保护图,以此对网络的安全运行进行相应的理论指导。最后,选取了丙烯聚合工艺、合成氨工艺以及TE过程为例,依据以上建立的基于复杂网络理论与SDG所建立的模型,对其SDG网络从网络复杂性特征、关键节点的识别和重点保护区确定三个方面进行了案例分析和验证研究。结果均表明,所建立的模型合理可行且与前人相比在简便性和准确性等方面具有一定优势,为化工生产故障诊断提供了新的研究思路和成果。
[Abstract]:With the increasing scale of chemical process and the more complex relationship between chemical production units, the operational elasticity of chemical production gradually weakened, high temperature and high pressure, low temperature vacuum, toxic corrosion and other harsh conditions are relatively prominent. In order to ensure the safety and smooth progress of chemical production process, it is necessary to identify and diagnose the chemical process faults. The theory of chemical fault diagnosis based on symbolic directed graph (SDG) and its application are in the ascendant. At the same time, this theory and method shows a certain degree of deficiency in modeling complexity, low integrity, easy to miss important variables and so on. At present, the complex network theory has carried out the related research in the power network, transportation and other fields of fault diagnosis, which is one of the effective means to deal with the complexity and variability of the process. Based on the research of SDG chemical network diagnosis theory, this paper attempts to use the complex network theory to identify the key nodes of the fault network and to quantitatively divide the community structure, in order to solve the shortage in the SDG chemical network diagnosis. The research work is mainly reflected in the following aspects: firstly, aiming at the modeling method of symbolic directed graph in fault diagnosis, the flow chart modeling method of traditional symbolic directed graph is improved. A new modeling method of network adjacency matrix based on SDG model is proposed. Secondly, the identification model and method of key nodes in chemical process diagnosis based on the combination of complex network and SDG are proposed, and the characteristics of complex network are calculated and judged. On the basis of the theory of centrality calculation and community division, the key nodes are identified and the key protected areas are divided. According to the results of the division of key nodes and communities, the key nodes and path protection maps are drawn. In this way, the safe operation of the network is guided by the corresponding theory. Finally, the propylene polymerization process, ammonia synthesis process and te process are selected as examples. According to the model based on complex network theory and SDG, the complexity of SDG network is analyzed. The identification of key nodes and the determination of key protected areas are studied in three aspects: case analysis and verification. The results show that the proposed model is reasonable and feasible, and has some advantages in simplicity and accuracy, which provides a new idea and achievement for fault diagnosis in chemical production.
【学位授予单位】：青岛科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TQ086

【参考文献】