考虑水声距的晶粒尺寸超声衰减评价模型  

Ultrasonic attenuation evaluation model of grain size considering water depth

作　　者：张晨昕[1] 李雄兵[1,2] 宋永锋[1] 刘锋[2] 刘希玲[1] 

机构地区：[1]中南大学交通运输工程学院,湖南长沙410075 [2]中南大学粉末冶金国家重点实验室,湖南长沙410083

出　　处：《中南大学学报（自然科学版）》2016年第4期1151-1157,共7页Journal of Central South University:Science and Technology

基　　金：国家高技术研究发展规划(863计划)项目(2012AA03A514);国家自然科学基金资助项目(61271356;51575541);中国博士后科学基金资助项目(2014M562126);中南大学中央高校基本科研业务费专项资金资助项目(2015zzts209)~~

摘　　要：超声衰减法评价金属材料晶粒尺寸时,会因忽略水声距设置造成系统误差而降低其评价精度。针对该问题,利用主成分分析法（PCA）对材料衰减系数测量中水声距的影响进行分析;根据PCA的数据空间降维投影特性,去除数据相关性并抑制数据噪声干扰,构建与水声距和晶粒尺寸均相关的综合衰减系数评价模型,并选取304不锈钢试块进行实验。在水声距13.8~156.9 mm范围内任取8个值,针对金相法测定晶粒粒度为72.35μm的304不锈钢试块,对比传统衰减评价模型与该模型的评价精度,验证其有效性。研究结果表明：传统衰减法与综合模型评价对金相法结果的相对误差分别为17.55%和6.49%;综合模型可抑制水声距调整精度对晶粒尺寸评价结果的不利影响,提高金属材料晶粒尺寸超声无损评价方法的实用性和可靠性。Ultrasonic attenuation was investigated as a nondestructive determination method of grain-size, but the neglect of water depth setting can introduce systematic error and reduce the precision of grain-size evaluation. A synthetic model dependent on both water depth and grain-size was established based on principal component analysis（PCA）, which can reduce noise as well as the dimensions of the feature vector. It can also eliminate superfluous characteristics and data correlation.304 stainless steel blocks were used to conduct the experiment. Numerical calculations were performed for the block with the grain size of 72.35 μm determined by the metallographic method when the water depth varied between 13.8 mm and 156.9 mm. The results show that the relative errors of conventional method and the proposed model are 17.55% and 6.49%, respectively, thus the presented model can inhibit the effect of water depth on grain-size evaluation and improve the applicability and reliability of ultrasonic nondestructive evaluation.

关 键 词：超声 衰减 水声距 晶粒尺寸 PCA 

分 类 号：TG115.28[金属学及工艺—物理冶金]