Research On Fault Diagnosis Technology Of Cloud Platform Wind Turbine

With the increasing pollution of non renewable energy and other non renewable energy sources on the environment,wind power generation has been developed rapidly in China.The increase in the single capacity of the fan and the complexity of the structure have greatly increased the probability of the failure of the unit.At the same time,along with the expansion of information acquisition system based on the wind turbine condition monitoring of the breadth and depth of strengthening,generated data showed mass characteristics,mass condition monitoring data storage and how to deal with growing,rapid and effective fault diagnosis and early warning has become an important research topic.In this paper,distributed storage and cloud computing are introduced to study the above topics.Using cloud computing distributed and highly fault-tolerant computing characteristics to improve data storage speed,distributed data storage mode increases the capacity and security of storage data.This paper describes the implementation of several key steps.Firstly,This paper expounds the structure and common faults of the wind turbine,introduces the commonly used methods and principles of the fan fault diagnosis,and briefly introduces the cloud computing platform.Secondly,On the basis of the above contents,the method of using Hadoop platform to realize the fault diagnosis of the wind turbine is put forward,and the process of the whole method is briefly described.Taking the storage of the data of the mechanical vibration signal of the wind turbine as an example,the concrete realization process of the distributed storage in the diagnosis process is introduced.It includes signal de-noising,eigenvalue extraction,and the writing and writing process of distributed storage.Data storage and speedup experiments are carried out to prove the efficiency of distributed storage for large amounts of data storage.Thirdly,Taking the prediction of brake shoe clearance as an important parameter of the mechanical braking system of wind turbine as an example,the realization process of distributed computing based on wavelet neural network algorithm is introduced in detail,and the simulation experiment is carried out to verify the good prediction effect.Finally,In order to improve the convenience and security of the fault management of the wind turbine,a logon system for the fault management of the wind turbine is designed based on the knowledge of Web.