Fan gearbox state recognition method based on likelihood learning machine

Abstract

The invention relates to a fan gearbox state recognition method based on a likelihood learning machine. The fan gearbox state recognition method based on the likelihood learning machine comprises thefollowing steps: 1) extracting the kurtosis from a vibration signal of the fan gearbox as a fault characteristic quantity of the fan gearbox; 2) using the likelihood learning machine to learn the kurtosis to generate a Gaussian model as a state monitoring model, that is, a Gaussian monitoring model; and 3) using the generated Gaussian monitoring model to judge the state of the gearbox. Compared with the traditional method of fault diagnosis by signal processing, the traditional method for fault recognition needs to rely on the experience of the operating personnel, but the fan gearbox state recognition method based on the likelihood learning machine does not rely on the experience of the operating personnel and has the characteristics of intelligent recognition; and compared with intelligent algorithms in the prior art, the intelligent algorithms in the prior art are prone to over-fitting problems, and the invention has good generalization and good portability.

Description

technical field [0001] The invention relates to the field of power system fault monitoring and diagnosis, in particular to a method for identifying the state of a wind turbine gearbox based on a likelihood learning machine. Background technique [0002] As the installed capacity of wind power continues to increase, the maintenance of wind farm equipment adopts the method of regular maintenance and after-event maintenance. The maintenance cost accounts for as much as 30% of the total operating cost of the wind farm. The planetary gearbox is a key component of the wind turbine transmission system. Under alternating loads, the failure rate exceeds 60% of the total failures of wind turbines. How to monitor the operation status and fault diagnosis of wind turbine gearboxes to reduce operation and maintenance costs and improve the safety and reliability of wind turbines is an important aspect of power system fault monitoring and diagnosis. field hotspots. [0003] Uncertain facto...

AI Technical Summary

Problems solved by technology

[0005] At present, signal analysis needs to combine the experience of operating personnel for fault judgment and identification, so the efficiency is low, and the intelligent algorithm has the problem of overfitting and low generalization performance

Images