Generator bearing fault prediction method

Abstract

The invention discloses a generator bearing fault prediction method. According to the method, first, SCADA data of a healthy wind turbine generator set is selected based on a power curve method to serve as training samples and test samples; second, a generator bearing fault prediction model of target parameters is established; third, the prediction model is utilized to predict target parameters ofthe test samples, and the target parameters are compared with actual values to obtain residual errors; fourth, a process control technology is utilized to calculate an upper limit and a lower limit of an alarm threshold and an upper limit and a lower limit of a warning; and last, the prediction model is utilized to predict an actual operation target parameter of the wind turbine generator set, the actual operation target parameter is compared with an actual value to obtain a first residual error, and the health state of a generator bearing is judged. The method has the advantages that the SCADA data of the wind turbine generator set is utilized to perform modeling prediction, the statistical process control ideology is utilized to divide the state of the generator bearing into a health state, a sub-health state and a fault state, an early warning to generator bearing failure is given in advance, and the use management capability of the bearing is improved.

Description

technical field [0001] The invention relates to the technical field of fault prediction methods, in particular to a fault prediction method for generator bearings. Background technique [0002] Wind turbine generator bearings are indispensable components of rotating machinery. Long-term work in harsh environments can easily lead to failure or failure of the inner and outer rings, rolling elements and cages of the bearings, affecting the normal operation of the entire unit and causing economic losses. At present, the fault diagnosis and early warning of bearings are mainly based on feature extraction and analysis of CMS vibration signals, but CMS monitoring equipment is not installed in every wind turbine, so it is difficult to detect and deal with bearing faults in time. Contents of the invention [0003] In view of the above-mentioned technical problems in the related art, the present invention proposes a generator bearing fault prediction method, which can overcome the a...

AI Technical Summary

Problems solved by technology

[0002] Wind turbine generator bearings are indispensable components of rotating machinery. Long-term work in harsh environments can easily lead to failure or failure of the inner and outer rings, rolling elements and cages of the bearings, affecting the normal operation of the entire unit and causing economic losses.

At present, the fault diagnosis and early warning of bearings are mainly based on feature extraction and analysis of CMS vibration signals, but CMS monitoring equipment is not installed in every wind turbine, so it is difficult to detect and deal with bearing faults in time

Images