Research On Vibration Detection And Noise Evaluation Of Wind Turbines

Statistical material shows that more than 60 percent of mechanical faults are caused by vibration. Vibration and noise detection is one of the important means for equipment state monitoring and fault diagnosis. The paper investigates the relevant theoretical problems of acoustic emission characteristic evaluation and vibration detection and analysis for wind turbines and develops a virtual instrument system for vibration detection and noise evaluation of wind turbines.The vibration mechanism of wind turbines and common analysis method employed for vibration signal analysis are studied. The basic parameters of vibration detection of wind turbines, the basic analysis method for vibration testing including time-domain analysis, amplitude-domain analysis, the calculation of vibration signals, spectrum analysis, time-frequency analysis, zero-phase filtering, etc. are analyzed in the paper. Why the vibration measurement and analysis is important for the fault diagnosis of wind turbines is discussed. The IEC 61400 international standard is deeply apprehended. The basic concepts in acoustic measurement are introduced. The IEC 61400-11 wind turbine noise evaluation international standard are analyzed, in which the calculations of measurement port, standard wind speed, sound power and tone are emphasized.The denoising method based on PDE for vibration signal is proposed. Firstly, the basic theories of PDE are introduced, the filtering principle of heat exchange equation is analyzed, and the relation between PDE filtering and traditional filtering is discussed theoretically. Then from the aspect of traditional filter, the design problems such as the evaluation time period, iteration time, and regulation parameter of PDE filtering are analyzed. A preliminary analysis of the extension in dimension of PDE filtering is given. The factors that affect the PDE filtering are discussed. The iteration form of PDE filtering is formed in several differential forms. The discrete PDE filtering model is build for the convenience of computer processing and the method is validated by experiments. And the comparative research with several other filtering methods is carried out. The wavelet ridge extraction is researched,The order analysis and frequency estimation based on wavelet ridge extraction in tone analysis is proposed. The elementary theories of wavelet ridge are introduced. The instantaneous amplitude and instantaneous frequency analysis based on wavelet ridge are proposed. The wavelet ridge extraction method based on reassigning algorithm and SVD is proposed. The detection and analysis method which can ensure the acoustic emission testing process of wind turbines are accurate and repeatable is discussed. Practical testing implies that the wind speed in the cabin is more relevant to the output power than the wind speed in the anemometer tower. The theoretical basis of sound power estimation formula based on sound pressure which is suitable for the mentioned process is deeply studied. The tone evaluation process based on narrow band spectrum and psychoacoustic model is analyzed in detail and the problems it exposures are discussed. After analysis above, we conclude that the order analysis and frequency estimation algorithm in tone analysis based on wavelet ridge extraction is more coincide with the concept in the IEC standard.The differences and similarities between error and uncertainty are discussed. The Evaluation and Expression of Measurement Uncertainty standards are introduced which mainly covers the standard uncertainty in the JJF1059-1999, the uncertainty in the A level assessing standard, the uncertainty in the B level assessing standard, the assessment of synthesis standard uncertainty and extended uncertainty. The uncertainty caused by the vibration and noise in the wind turbines and that caused by the virtual instrument are analyzed. According to the IEC 61400-11 standard, the analysis for the assessment of uncertainty in vibration detection, sound power uncertainty, 1/3 octave uncertainty and tone uncertainty is given.The basic concept and the features of virtual instruments are introduced. The commonly used virtual instrument development systems at home and aboard are introduced. So are the control-oriented virtual instrument virtual instruments. The vibration measurement module and noise assessment module for wind turbines are developed. The functional analysis and the constitution of the modules are also provided. The paper develops an industrial level measurement and assessment system for the IEC evaluation of acoustic emission characteristics of wind turbines whose frame is based on that of the virtual instrument and successfully put it into commercial use in the wind power industry. It fulfils the pre-evaluation for the IEC authentication in acoustic emission of a lot of new pattern wind turbines, gaining good reputation for its good performance to cost ratio.