Non-stationary Analysis and Causal Diagnosis Method for One Million KW Ultra-Supercritical Units

Abstract

The invention discloses a multi-layer Bayesian network fault diagnosis method based on causal analysis for the non-stationary characteristics of a mega-kilowatt ultra-supercritical unit. The present invention first uses sparse co-integration analysis and Granger causality analysis to respectively extract local refined causal relationship subnetworks among non-stationary variables, and simultaneously extract global causal relationships for all variables. Finally, the Bayesian network is used to trace the fault path and locate the root fault variable. This method overcomes the influence of non-stationarity in the causal analysis of large coal-fired units, fully excavates the potential causal relationship contained in the non-stationary variables, establishes a multi-layer diagnosis model, and effectively solves the problem of locating the root cause of the complex non-stationary fault process The problem of difficult variables greatly improves the diagnostic performance of fault propagation in non-stationary processes, and helps thermal power plants to carry out effective and timely inspection and maintenance of factories, thus ensuring the safe and reliable operation of million-kilowatt ultra-supercritical generating units.

Description

technical field [0001] The invention belongs to the field of large-scale non-stationary process fault diagnosis, in particular to a Bayesian network fault diagnosis method for the non-stationary multi-layer causal structure of a mega-kilowatt ultra-supercritical unit. Background technique [0002] In the rapidly developing 21st century, the direction of large-scale and complex thermal power generation has become the mainstream. With the continuous development of science and technology and the rapid consumption of global energy, the thermal power industry has higher and higher requirements for energy efficiency. The previous million-kilowatt supercritical unit was not enough to meet the requirements, so the million-kilowatt ultra-supercritical unit began to dominate. This unit is the most advanced high-efficiency and large-capacity coal-fired power generation equipment in the world, with obvious energy efficiency advantages. It is the representative unit and the mainstream di...

AI Technical Summary

Problems solved by technology

However, in the data-driven method, the traditional method of using difference to eliminate non-stationary characteristics may lose effective causal information, resulting in incomplete or even wrong causal relationships in the established fault tracing network, which eventually leads to misjudgment of fault root variables

Therefore, ultra-supercritical units of one million kilowatts have obvious large-scale, complex and non-stationary characteristics, which bring great challenges to fault diagnosis

Images