Offshore wind turbine generator gear system fault diagnosis method based on Park transformation

Abstract

The invention discloses an offshore wind turbine generator gear system fault diagnosis method based on Park transformation. The method comprises the following steps of: collecting three-phase current parameters of an offshore wind turbine generator stator; preprocessing the three-phase current parameters through park conversion to determine a fault characteristic quantity; supplying the fault characteristic quantity to an auto-encoder in a sample learning mode for sample training; obtaining a hidden layer representation method by the auto-encoder, enabling an encoder to finely adjust the parameters of the whole system according to a training result after training of a plurality of hidden layers to obtain a neural network model with feature extraction and mode recognition functions, and calculating a threshold; and inputting test online data into the trained neural network model to obtain an abnormal score score of a test set, and drawing an exponentially weighted moving average value control chart to judge whether a fan gearbox has a fault or not. The method is better in noise reduction capability, is higher in diagnosis precision, and can meet the precision and real-time requirements of offshore wind power.

Description

technical field [0001] The invention belongs to the field of fault diagnosis of wind power generation, and in particular relates to a fault diagnosis method for a gear system of an offshore wind turbine based on Park transformation. Background technique [0002] At present, there have been many studies on how to diagnose the faults of the wind turbine gear system, such as an online fault diagnosis system for wind turbine gearboxes recorded in Chinese patent document CN201310301699.0, and a vibration-based A wind turbine gearbox fault diagnosis method based on signal blind source separation and sparse component analysis, a genetic algorithm combined with stacked noise reduction sparse autoencoder described in Chinese patent document CN201710599192.6, a wind turbine described in Chinese patent document CN202010134735.9 Gearbox fault diagnosis method and system, Chinese patent document CN201610916137.0 records a wind power gearbox fault diagnosis method based on adaptive resona...

AI Technical Summary

Problems solved by technology

[0003] The disadvantages of the above existing technologies are: the above technical solutions are mainly aimed at onshore wind turbines, while for offshore wind turbines, the wind speed at sea is unpredictable, and these working conditions will cause the double-fed generator gearbox to be subjected to a large load The failure rate is high and online monitoring will be greatly affected by noise. Many traditional methods are difficult to meet the requirements of accuracy and real-time

Images