Thermal power equipment semantic knowledge base, construction method and zero sample fault diagnosis method

Abstract

The invention discloses a thermal power equipment semantic knowledge base, a construction method and a zero sample fault diagnosis method. The method comprises the following steps: extracting fault attribute information from a fault diagnosis case text containing expert knowledge summarized in a thermal power generation process, coding the fault attribute information into an attribute vector, and combining data corresponding to a case to train an attribute discriminator, so as to establish mapping between data and fault case attributes, establish a'data-attribute-attribute discriminator 'ternary semantic knowledge base, and improve the fault diagnosis efficiency. And the problem of zero sample fault diagnosis of high-end thermal power equipment is solved. According to the method, expert knowledge and a data driving method are creatively combined, when a new fault occurs, an attribute discriminator is applied to judge the attribute of the new fault, and the attribute is coded into an attribute vector, so that the fault mode is determined based on the attribute shared between the fault modes, and migration and sharing of knowledge between the faults are realized. The method has a good diagnosis effect on faults without training data, and the zero sample fault diagnosis problem encountered in high-end thermal power equipment is well solved.

Description

technical field [0001] The invention belongs to the field of fault diagnosis of thermal power equipment, and particularly relates to a semantic knowledge base of thermal power equipment, a construction method and a zero-sample fault diagnosis method. Background technique [0002] With the rapid development of my country's economy, the society's increasing demand for electricity has become a pillar industry in my country's national economy. The next 20 to 30 years will be a critical period for the adjustment and transformation of energy production and consumption patterns and energy structure. However, due to the abundant coal resources in my country, coal energy is still the most important source of electricity in my country. According to statistics, by the end of 2016, coal still accounted for more than 60% of my country's primary energy consumption structure. And in a long period of time, my country's coal-based production status will not change. In recent years, in ord...

AI Technical Summary

Problems solved by technology

However, the above methods rely heavily on data for modeling, but in the actual industrial production process, the data of failure modes is very scarce and difficult to obtain, and has obvious long-tail distribution characteristics, that is, a small number of failures with higher frequency can accumulate some data, and many failures occur very rarely or even do not occur, and there is no corresponding training data

This limits the application of traditional data-driven methods in practical processes

[0004] In the production process of power plants, a large number of fault case texts containing expert knowledge are often accumulated, but currently these unstructured text information with "expert knowledge" has not been fully utilized

Images