以不同晶粒尺寸的304奥氏体不锈钢试样为研究对象，采用频率5 MHz的超声换能器对其进行水浸横波检测，通过快速傅里叶变换（FFT）分析裂纹缺陷和晶粒反射信号的能量特征，以便排除缺陷信号的干扰；通过超声检测得到无裂纹缺陷试样的相对二阶非线性系数，分析了其与平均晶粒尺寸的关系。结果表明：当在1.6~4.4 MHz频段的能量占比低于1%时，可认为奥氏体不锈钢中无裂纹缺陷存在；相对二阶非线性系数随着奥氏体不锈钢平均晶粒尺寸的增大而减小，因此可利用该参数来表征奥氏体不锈钢的晶粒尺寸。

Taking 304 austenitic stainless steel samples with different grain size as research object, water-immersed shear wave testing was carried out with a ultrasonic transducer with frequency of 5 MHz. The energy characteristics of crack defect and grain reflection signals were analyzed by fast Fourier transform (FFT) to eliminate the interference of defect signals. Relative second-order nonlinear coefficients of samples without crack defects were obtained by ultrasonic testing, and the relationship between the coefficient and the average grain size was analyzed. The results show that when the energy proportion in 1.6-4.4 MHz band was less than 1%, we considered that no crack defects existed in the austenitic stainless steel. The relative second-order nonlinear coefficient decreased with the increase of average grain size of the austenitic stainless steel, so the parameter could be used to charaterize the grain size of austenitic stainless steel.