Wind power intelligent fault diagnosis method based on big data

Abstract

The invention discloses a wind power intelligent fault diagnosis method based on big data. The wind power intelligent fault diagnosis method comprises the steps of: subjecting a fault list and data files corresponding to the fault list to collection and data arrangement; establishing a main fault sorting model based on a Bayesian network; establishing a main fault positioning model based on a deeplearning algorithm; and performing real-time fault diagnosis and positioning based on the main fault sorting model and the main fault positioning model. According to the wind power intelligent faultdiagnosis method, the Bayesian network algorithm and the deep learning algorithm are combined, the problem of manual feature screening is avoided, the fault diagnosis result can be quickly obtained based on a fault sorting result and a fault positioning result, and the wind power intelligent fault diagnosis method has high anti-interference capability.

Description

technical field [0001] The invention relates to the technical field of wind power, in particular to an intelligent fault diagnosis method for wind power based on big data. Background technique [0002] Wind power plants are generally built in remote areas, and many major equipment are located in the nacelle, at a height of more than 80 meters from the ground, which increases the difficulty of wind power troubleshooting. When an actual fault occurs, the alarm of the main fault will be accompanied by a large number of other non-main fault alarms. In actual production, determining the main fault in the entire fault alarm sequence is the most urgent and basic requirement. Building a wind power fault diagnosis model can quickly and accurately locate major faults, reduce operation and maintenance costs, and improve operation and maintenance efficiency. [0003] At present, the methods of fault diagnosis mainly include: (1) Wavelet analysis method. As a brand-new time-frequency ...

AI Technical Summary

Problems solved by technology

[0004] However, the existing fault diagnosis methods all have deficiencies: the wavelet analysis method needs to use the vibration monitoring data of wind power, cannot give the diagnosis result in real time, and requires a long time for diagnosis; the neural network method needs to realize the feature expression based on manual experience, which is laborious , time-consuming, and requires high manual experience; the expert system method is too dependent on expert experience, and the anti-interference ability is weak

Images