主元分析法在火电厂故障检测中的应用及效果研究

发布时间：2018-01-14 01:00

本文关键词：主元分析法在火电厂故障检测中的应用及效果研究　出处：《华北电力大学》2016年硕士论文　论文类型：学位论文

【摘要】：火电企业技术改造是当今技术创新及应用研究的主流方向之一,其中,技术改造主要是通过改进物理结构或加装新设备来实现,如加装脱硫装置、对磨煤机改进等。然而,技术改造并非只有物理改造,还应包括电子技术的改进,如对控制算法、检测方法的改进。在这些方面,国内相关研究并不多。本文通过研究火电企业运行过程,发现在故障检测方面,国内绝大部分火电厂均使用固定的控制限来报错,而火电厂实际运行过程却并非持续稳态运行,这就导致了在火电厂增减负荷时系统出现误报或数据的丢失,严重影响监测的稳定性,从而影响火电企业的经济性。主元分析方法(PCA)作为多元统计方法的一种,其对过程控制的故障检测与诊断方法不依赖于系统的数学模型,因此是广泛应用在工业领域的统计检测方法。基于主元分析方法的过程检测方法,由于充分的利用了主元分析算法在处理线性相关数据时降维的功能,使得对多变量生产过程的检测可在低维变量控件实现。因此,考虑在火电企业故障检测中使用主元分析方法。故障检测中的PCA方法原理是在稳态过程中收集数据,用固定控制限的SPE、T_2和T_H~2来监测。然而,对于那些过程暂态值必须考虑的系统(如火电厂运行系统),用固定控制限会引起误报和丢失数据,这会严重影响监测系统的稳定性,并带来不必要的经济损失。本文立足于这一点,对主元分析方法进行改进,使得其能应用在动态稳态经常切换的火电厂运行系统。本文总结了在火电厂故障检测领域在的国内外的研究成果,指出了该领域需要深入研究的问题;针对目前存在的问题,提出了基于方差的自适应控制限的改进方法,从而来克服暂态过程中因条件改变而产生的误报问题;选取某火电厂的相关数据,利用matlab软件对该方法进行了仿真,验证了该技术改造的可行性;通过建立评价指标,运用层次分析法确定指标权重,运用模糊评价法来对方案进行综合评价。结果表明,该技术改造可以为火电企业获得经济效益与技术上的改进。
[Abstract]:Technical transformation of thermal power enterprises is one of the main direction of the technology innovation and application of the technology is mainly realized through the improvement of the physical structure or the installation of new equipment, such as the installation of desulfurization equipment for mill improvement. However, technological transformation is not only a physical transformation, but also should include the improvement of electronic technology, such as the control algorithm, the improvement of the detection method. In these areas, not many domestic research. Through the research of thermal power enterprise operation process, found in fault detection, most of the domestic thermal power plants are used to control limit fixed error, while the actual operation process of power plant is not sustained steady operation. This resulted in the the system of thermal power plant load changes when the loss of false positives or data, seriously affect the stability of monitoring, thus affecting the economy of power enterprises. Principal component analysis (PCA) as a multivariate system A kind of design method, the mathematical model of the method of fault detection and diagnosis process control is not dependent on the system, so the statistical detection methods are widely used in industrial field. The process of PCA detection method based on the full use of the principal component analysis algorithm for dimensionality reduction in data processing of linear correlation the function, which makes the detection of the multi variable production process can be realized in low dimension control. Therefore, consider using principal component analysis method in fault detection in thermal power enterprises. The principle of PCA method in fault detection is to collect data in the steady state in the process of using the fixed control limits of SPE, T_2 and T_H~2 to monitor. However, for the transient value must be considered in the system (such as power plant operation system), with the loss of data will cause false positives and fixed control, which will seriously affect the stability of the monitoring system, and bring economic loss is not necessary Lost. Based on this, the principal component analysis method was improved, so it can be used frequently in the dynamic steady state switching power plant operation system. This paper summarizes the fault detection in the field of thermal power plant in the domestic and foreign research results, pointed out that the field needs to be further studied aiming at the existing problems; and put forward the improved method based on variance adaptive control limit, the problem of false positives due to changing conditions to overcome in transient process; selecting the relevant data of a thermal power plant, the method was simulated by using MATLAB software, to verify the feasibility of the technical transformation; through the establishment of evaluation index, using the analytic hierarchy process the weights of the indexes, using fuzzy evaluation method to comprehensive evaluation of the program. The results show that the technology can obtain improved economy benefit and technology of thermal power enterprises.

【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TM621

【引证文献】