High-Frequency Immersion Ultrasonic Testing And Evaluation Method For Inclusions In Steel

Steel materials are important basic materials for aerospace,defense,military,transportation and shipbuilding,especially special steels such as gear steel,bearing steel and spring steel,which have higher strength,toughness,physical properties and chemistry than ordinary steel and play an important role in the construction of national major projects and high-end equipment manufacturing.With the development of national economy and technology,higher requirements have been imposed on the strength and life of steel materials.Inclusions destroy the continuity of steel matrix and affect the steel processing properties seriously,which limits the mechanical properties such as load bearing capacity and fatigue life of steel products.Rapid and reliable detection of inclusions in steel is of great significance for analyzing the source of inclusions,improving the smelting process of steel,achieving effective strategy of inclusion control,and evaluating the final quality of steel products.Metallographic statistics,electrolytic extraction and other traditional sampling-based methods not only have a long detection period,but also have limited detection area or volume,which lead to low efficiency and reliability of inclusion content evaluation and indirectly affect the steel smelting process and final product quality.In this paper,the inclusion detection method based on high-frequency immersion ultrasonic testing is applied to detect inclusions in steel rapidly and reliably.The ultrasonic parameters from echo signals of inclusions are analyzed,and the methods of identifying defect type,accurately measuring inclusion size and detecting defect in blind-zone are separately proposed.The experiment of inclusions detection is performed using Scanning Acoustic Microscope.The main contents of the paper include:(1)For the current numerical calculation model,the transmission/reception characteristics of the transducer,the distribution of incident sound field and the refraction/attenuation of acoustic wave in liquid-solid medium are not considered,which lead to the mapping relationships of characteristic parameters of echo signal and defect attribute cannot be accurately obtained.To solve this problem,the acoustic scattering model and the complete finite element model of transducer-water-steel-defect are established.The far-field scattering sound pressure and echo signals of different types of defects are calculated,and the differences of acoustic scattering between inclusions and holes are analyzed.The relationships between characteristic parameters of echo signal and sound field distribution,inclusion size,type,shape and orientation are established,which provide the basis for the identification of defect type and the accurate measurement of inclusion size.(2)Training samples for inclusions and holes are difficult to obtain from experiment accurately,resulting in machine learning methods cannot effectively distinguish inclusions and holes.To solve this problem,the ultrasonic measurement model capable of accurately predicting the echo signals of inclusions and holes is established.The sound field distribution of broadband transducer is calculated by modifying the Multi-Gaussian Beam model,and the far-field scattering amplitude of holes with any size are calculated by using the combination of Born approximation and Kirchhoff approximation.The relationships between the amplitude and peak frequency of echo signals and the size and type of defects are theoretically analyzed using the established ultrasonic measurement model.Both theoretical and experimental results show that the peak frequency decreases linearly with the increase of defect size.Combined with the amplitude and peak frequency of echo signal,the inclusions and holes can be distinguished effectively.(3)To reduce the error of measuring inclusion size by the-6dB-drop method,an accurate measurement method of inclusion size based on C-scan image restoration is proposed.The Richardson-Lucy algorithm is improved by using the total variation regularization constraint,which effectively suppresses the generation of ringing noise during image restoration.The modified algorithm is used to restore the C-scan image,which eliminates the blurring effect of point spread function on the C-scan imaging.Experiments show that the total variation regularization Richardson-Lucy algorithm can improve the accuracy of inclusion size measurement,and the measurement error of inclusion size is reduced from more than 50%to less than 10%.(4)The traditional pulse-echo method has detection blind-zone,which makes the defects in surface and subsurface of material cannot be detect effectively.To solve this problem,a blind-zone defect detection method based on the interference of leaky Rayleigh wave and reflected longitudinal wave is proposed.The interfering echo signal of leaky Rayleigh wave and reflected longitudinal wave is received and the average delay time difference of the interference ring is obtained by using the delay time of leaky Rayleigh wave in C-scan image.Combined with the quantitative relationship between the delay time difference and the defect size established by finite element method,the defect size in blind-zone is measured.Experiments show that the method of interference between leaky Rayleigh wave and reflected longitudinal wave can realize the imaging detection and quantitative characterization of defects in blind-zone.The relative error of defect size measured by the average delay time difference of interference ring is about 10%.