面向晶粒尺寸的超声多尺度衰减评价方法

发布时间：2018-07-10 06:14

本文选题：晶粒尺寸 + 超声无损评价　；参考：《金属学报》2015年01期

【摘要】：用小波变换获取超声波能量的时间-尺度分布,研究衰减系数随尺度的分布规律,定义加权的超声多尺度衰减系数,结合粒子群算法设计的最优尺度组合及其归一化权重分配策略,建立晶粒尺寸的超声多尺度衰减评价模型.选用304不锈钢进行实验,其衰减系数-尺度分布图表明超声波在小尺度下衰减迅速,体现了高散射材料中衰减的频率特征;而随着试样晶粒尺寸增大,整个尺度范围内的衰减都明显加剧.实验结果显示,声速法、传统衰减法与本方法的最大系统误差分别是+12.57%,+5.85%和-1.33%.对金相法测得平均晶粒尺寸为103.5 mm的验证试样用3种方法进行评价,结果分别为(110.4±7.8),(98.2±6.6)和(101.7±3.9)mm.本方法不仅可降低系统误差,且随机误差也被小波变换的恒Q滤波特性有效抑制.
[Abstract]:The time-scale distribution of ultrasonic energy is obtained by wavelet transform. The distribution law of attenuation coefficient with scale is studied, and the weighted multi-scale attenuation coefficient is defined. An ultrasonic multi-scale attenuation evaluation model for grain size is established by combining the optimal scale combination and the normalized weight allocation strategy designed by particle swarm optimization (PSO). Using 304 stainless steel as experimental data, the attenuation coefficent-scale distribution diagram shows that ultrasonic wave attenuates rapidly at small scale, which reflects the frequency characteristic of attenuation in high scattering material, but with the increase of grain size of the sample, The attenuation in the whole scale is obviously aggravated. The experimental results show that the maximum systematic errors of sound velocity method, traditional attenuation method and this method are 12.57, 5.85% and -1.33, respectively. Three methods were used to evaluate the average grain size (103.5 mm) obtained by metallographic method. The results were (110.4 卤7.8), (98.2 卤6.6) and (101.7 卤3.9) mm, respectively. This method not only reduces the system error, but also reduces the random error effectively by the constant Q filter of wavelet transform.
【作者单位】：中南大学CAD/CAM研究所;中南大学粉末冶金国家重点实验室;中国兵器科学研究院宁波分院;
【基金】：国家自然科学基金项目61271356,51205031和51105045 国家高技术研究发展计划项目2012AA03A514 湖南省自然科学基金项目14JJ2002 中国博士后科学基金项目2014M562126资助~~
【分类号】：TG115.285

【相似文献】