[ 1]。
Among all kinds of faults of steam turbine generator set, vibration fault is a kind of fault that has great influence on production and operation. On the one hand, the diagnosis of vibration fault is complicated and the processing time is long; On the other hand, once the vibration fault diverges and causes an accident, the impact and consequences are very serious [2].
1 condition monitoring and fault diagnosis of large steam turbine generator set
Due to the power demand and financial shortage in China, it is very important to reduce the failure rate of old units and extend their service life [3]. At present, only a part of the operation monitoring of various large turbine-generator sets in domestic power plants are equipped with vibration monitoring systems of Bentley Company of the United States or Philips Company of Germany, and many of them are relatively backward and imperfect. Therefore, it is necessary to study the fault diagnosis technology of large turbo-generator set.
With the increase of unit capacity, vibration faults become more and more complicated. At present, on-line monitoring equipment generally has only a few vibration expert systems, which are not perfect. Using advanced detection and diagnosis instruments and scientific and effective technical methods for on-site fault diagnosis is the main method for fault diagnosis and prediction analysis of various units in power plants at present [4].
At present, western developed countries, mainly the United States, are in a leading position in the comprehensive research of on-line monitoring and diagnosis technology for large steam turbine generator sets: on the one hand, the signal processing and data analysis technology in the United States is developing rapidly, and these processors, analyzers and data acquisition systems are the foundation and core of mechanical equipment condition monitoring and an inseparable part of the subsequent technology (fault diagnosis) development; On the other hand, several professional companies in the United States, such as Bentley, Bay, etc., have been engaged in the research on the operation and monitoring of large power plants, as well as the reliability, safety, maintainability and economic management technology of the units. They have established a huge database management system and carried out expert system research for more than 40 years, with strong data and software strength. In addition, there are many famous diagnostic instrument companies in the world, such as B&K in Denmark, Schenk in Germany and Takeda Riken in Japan, which produce various analytical instruments and software systems for equipment diagnosis. However, foreign online monitoring systems, on-site diagnostic instruments and diagnostic management software are generally very expensive, and due to the lack of localization, there are some problems such as inconvenient maintenance and use, which are difficult to popularize in grass-roots power plants in China.
The equipment diagnosis technology of industrial enterprises in China started from 1983, and was mainly used in petrochemical, metallurgical and electric power industries in the early stage. After entering the 1990s, it quickly penetrated into all major industries of the national economy. Among them, the fault diagnosis of rotating machinery is the most widely used and involves the most industries, such as turbine-generator sets in power industry, compressors in petrochemical industry, various aero-engines in aviation industry and so on. On-line monitoring and fault diagnosis technology for large turbo-generator sets, as a key scientific and technological project in the Seventh Five-Year Plan and the Eighth Five-Year Plan, continued to be supported during the Ninth Five-Year Plan, and its significance is obvious. Some universities such as Xi Jiaotong University, Harbin Institute of Technology, Tsinghua University and some research institutes such as Xi Thermal Energy Engineering Research Institute are generally at the leading level in the research of fault mechanism and diagnosis technology of large steam turbine generator sets. However, due to the increasing capacity of large turbo-generator assembly machine in recent years (for example, 700 MW turbo-generator set has been put into production in China), the research on the mechanism, fault characteristics and on-site diagnosis methods of many common faults of large turbo-generator set needs to be further deepened. In addition, in the field of signal acquisition, fault diagnosis instrument and data management software development, although some domestic universities and research institutes have introduced their own products, such as 902 and 903 portable data collectors introduced by Beijing Tong Zhen Testing Technology Research Institute, QLSA-W vibration and noise testing analyzer introduced by Chongqing University Testing Center, PDM2000 data acquisition analyzer and management software introduced by Dalian University of Technology, etc. However, with the rapid development of computer technology, especially microprocessor and software technology, the above-mentioned equipment and software systems have some limitations in performance index, reliability and adaptability of software to data acquisition equipment of different companies.
2. Main contents and general situation of fault diagnosis technology research
For more than 30 years, fault diagnosis technology has been absorbing the new achievements of various scientific and technological developments, and the theory and application of diagnosis have made great progress. It involves system theory, cybernetics, information theory, detection and estimation theory, computer science and many other aspects. It has become a new interdisciplinary subject integrating basic disciplines such as mathematics, physics, mechanics, chemistry, electronic technology, information processing, artificial intelligence and related professional disciplines. The main contents of fault diagnosis technology research include the following four aspects: fault mechanism; Fault information processing technology; Fault source separation and location technology; Research on the application of artificial intelligence technology.
2. 1 Study on fault mechanism [5 ~ 7]
The study of fault mechanism is based on reliability and fault physics, studying the physical or mathematical model of fault, and carrying out physical simulation or computer simulation. Its purpose is to understand the formation and development process of the fault, make clear the dynamic characteristics of the fault, so as to further master the typical fault signals, extract the fault symptoms and establish the fault model mode. The study of fault mechanism is the basis of fault diagnosis and an important guarantee to obtain accurate and reliable diagnosis results.