Research Of Ultrasonic Nondestructive Testing Signal Processing Method

Equipment metal plate is the key part of mechanical equipment,its quality directly affects the reliability of mechanical equipment and possible white dots on their plate inside,for detecting defects such as cracks and inclusions.Ultrasonic nondestructive testing has the advantages of high detection accuracy,low cost of detection device,harmless to human and environment,and suitable for defect detection of mechanical equipment board.Ultrasonic testing method is the most widely used nondestructive testing method,ultrasonic signal processing technology is the main research topic in the field of ultrasonic nondestructive testing.Ultrasonic flaw echo signals is a kind of time-varying signal,the non-linear and non-stationary characteristics of traditional Fourier transform is mainly used for linear,smoothness of signal processing,and based on the Fourier transform of related transformation such as fast Fourier transform,wavelet transform has the limitations of Fourier analysis signal,so this paper mainly studied the Hilbert Huang transform analysis method,Hilbert-Huang transform was used to deal with the flaw signal of mechanical equipment plate in the experiment.A series of related researches are carried out on the existing problems in the process:Firstly,the basic principle of ultrasonic nondestructive testing and the propagation characteristics of ultrasonic wave in medium are discussed,and the corresponding detection method is used to detect the mechanical equipment.The defects,positions and causes of mechanical equipment are analyzed.The qualitative,quantitative and positional analysis of the defects of typical ultrasonic signals were carried out.Secondly,the system construction of the simulated ultrasonic testing machine is realized.In the experiment,the defect signals of four different crack depths were collected,the time domain waveform is analyzed.Due to the limited information capability of time domain waveform,it is extended to time-frequency domain for analysis.In time frequency domain,the traditional signal processing algorithm and its inapplicability are introduced.Aiming at the characteristics of ultrasonic echo signal,this paper mainly studies the time-frequency analysis method for dealing with non-stationary signals,namely the Hilbert-Huang transform method,Using the Hilbert-Huang transform method to deal with the problem of ultrasonic defect echo in the experiment,The problem is endpoint effect and modal confusion.Thirdly,aiming at the problem of endpoint effect,the two methods of improvement are analyzed,which is the image extension method and the boundary local feature scale extension method.Through the simulation experiment the comparative analysis,Finally,the ultrasonic defect echo in the experiment was extended by the boundary local characteristic scale extension method;For modal confusion.The EEMD decomposition method is introduced,and compared with EMD method about decompose the synthetic signal.The results show that the EEMD method is effective for the problem of modal aliasing caused by EMD decomposition.The EEMD method is applied to the echo signal of ultrasonic defect,and the decomposition results show that the mode confusion is eliminated.Based on the EEMD method,the optimal noise reduction model of the original ultrasonic defect signal was established.Finally,the principle of Hilbert spectrum and marginal spectrum is introduced.The spectrum analysis and comparison of the Hilbert marginal spectrum and the fast Fourier transform of a known frequency amplitude signal are presented.Showing the advantages of Hilbert spectrum analysis.Furthermore,the Hilbert spectrum and the marginal spectrum of the ultrasonic defect echo signals of four different crack depths in the experiment were presented.And analyze the various defect information represented by the spectrogram.Finally,the purpose of detection and identification of internal defects of mechanical equipment is achieved.