油气井后处理生产应急智能监控系统研究

发布时间：2018-07-01 09:48

本文选题：安全生产 + 预测预警　；参考：《北京邮电大学》2013年硕士论文

【摘要】：本文首先阐述了净化厂的基本情况,包括生产设备、工艺流程、DCS系统,然后根据生产流程中各个采集数据点的关系,将数据点分为几个不同的模块,每个模块可以单独表示生产工艺中的一个流程。之后,针对项目建设的需求提出了对预警系统的设计思路,同时介绍了作为系统构架平台的ECM,提出了模型构建和数据处理的基本流程和方法。本文的主要研究对象是对预测预警平台系统的构建,因此本文随后研究数据分析的各种方法的理论知识,确定了PCA为系统中需要应用的主要的分析方法,并重点介绍了PCA算法的计算流程,另外,对于PLS、神经网络、小波、EWMA等分析方法也做了总体的概述。之后,提出了针对脱硫、脱水、硫磺回收、蒸气四个生产单元的模型设计方法。最后,在ECM中以主成分分析法PCA为主要模块建模,建立了脱硫、脱水、硫磺回收、蒸汽等生产单元的故障预测预警的模型,并且对模型中的每个元件进行了详细的介绍；在压缩机的模型中,由于开关机状态的分辨比较复杂,所以提出了以时间平移为核心的算法来确定机器的开关机状态。通过建模后组成的系统研究长寿分厂设备运行的状态,保证了监测对象的安全生产。本项目实现了对净化厂生产过程的实时监控,使得在生产中的可能发生危险之前及时发现潜在的问题并排除成为了可能,为净化厂安全生产提供了保障。
[Abstract]:This paper first describes the basic situation of the purification plant, including the production equipment, process flow and DCS system, then according to the relationship between the data points collected in the production process, the data points are divided into several different modules. Each module can represent a single process in the production process. After that, the design idea of early warning system is put forward according to the requirement of project construction. At the same time, the ECM, as the platform of system architecture, is introduced, and the basic flow and method of model construction and data processing are put forward. The main research object of this paper is the construction of the prediction and early warning platform system, so this paper then studies the theoretical knowledge of various methods of data analysis, and determines PCA as the main analysis method that needs to be applied in the system. The calculation flow of PCA algorithm is introduced in detail. In addition, the analysis methods such as PLS, neural network and wavelet EWMA are also summarized. After that, the model design method for desulfurization, dehydration, sulfur recovery and steam production unit is put forward. Finally, the principal component analysis (PCA) is used as the main module in ECM to model the failure prediction and warning model of desulfurization, dehydration, sulfur recovery, steam and other production units, and each component of the model is introduced in detail. In the compressor model, due to the complexity of the state resolution of the switch machine, an algorithm based on time translation is proposed to determine the state of the machine switch machine. After modeling, the system is used to study the operation state of the plant equipment, which ensures the safety of the monitoring object. The project realizes the real-time monitoring of the production process of the purification plant, which makes it possible to find the potential problems and eliminate the potential problems in time before the possible danger in the production, thus providing a guarantee for the safe production of the purification plant.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP277;TE687

【参考文献】