Research And Design Of System For On-line Monitoring Particle In Wind Turbine Lubricants

ABSTRACT:In the working process of wind turbines,the internal components will produce wear particles because of the friction.These wear particles circulate in the tube with lubricants.By monitoring the wear particles in the lubricants,we can know the appropriate time to change lubricants and abrasion level of the wind turbines,which can increase the safety of wind turbines and extend its service life.To solve the difficulties of wear particle signal detection,this paper proposes wear particle signal detection algorithms.Moreover,this paper studies and designs a kind of system for on-line monitoring particle in wind turbine lubricants.This paper is focus on the system for on-line monitoring particle in wind turbine lubricants.This paper has studied the design requirements of the system,algorithmn for detecting wear particles,and overall design of the system.(1)The overall design framework is proposed in this paper,which consists of wear particle sensor,wear particle signal data processing,and host computer.The wear particle signal data processing part should have these functions,such as data collection,data processing,data storage and data communication.With regard to each function,this paper proposes the corresponding strategies and selects the best one.(2)This paper analyzes the difficulties of wear particle detection,mainly including high frequency interference,baseline deviation,spike noise,inertia fake crest,and chopped waveforms,which can result in wear particle detection error.To eliminate the high frequency signal interference,this paper adopts Polynomial Fitting Moving Smoothing method as the filtering algorithm.The peak-seeking algorithm can determine the crest and trough of wear particle waveforms,which solves these difficulties,including baseline deviation,spike noise,and inertia fake crest.This paper proposes two kinds of classification algorithms to solve the chopped waveforms problem,which can classify the wear particle signals based on the positions of the crest and trough.Eventually the wear particle information is extracted from the sampling data.(3)This paper designs and implements the system for on-line monitoring particle in wind turbine lubricants.In the software part,the sampling data is transfered by DMA.The data is stored in FRAM via I2C bus.Moreover,the communication with host computer is realized by Modbus protocol.In the hardware part,this paper designs the circuit based on DSP.The parameters of the wear particle detection algorithm are set by experiemnts.The data continuity testing verifies the feasibility of tansfering data by DMA.The filtering testing verifies the effectiveness of the filtering algorithm.As the wear particle detection experimental results shown,the wear particle detection algorithm achieves good results for the ferromagnetic wear particle with more than 100?m diameter and non-ferromagnetic wear particle with more than 500?m diameter.The results from the host computer verify the effectiveness of the whole system.