作者：游剑, 邹伟健, 晋刚, 雷玉, 朱世超, 宋卓明
作者单位：华南理工大学聚合物成型加工工程教育部重点实验室, 聚合物新型成型装备国家工程研究中心, 广东 广州 510641
关键词：超声在线测量;复合材料;填充含量;支持向量机回归
摘要：
利用高压毛细管流变仪在稳定的温度和流速条件下挤出低密度聚乙烯（LDPE）/玻璃微珠（GB）复合材料，并使用超声波测量复合体系中玻璃微珠的填充含量，定义时域上的声衰减参量S1和频域上的特征向量M0，并分别使用线性回归（LR）和支持向量机回归（SVR）的方法建立S1和M0测量填充含量的定量模型，模型预测结果与离线的灰分法测量结果进行对比。实验结果表明，基于自定义特征参数的两个模型的预测填充含量偏差最大不超过1.2%，最小仅为0.17%，都能满足在线实时测量的需求。其中频域方法相比时域方法预测准确性更高，与离线的灰分法测量精度相近。
Ultrasonic online measurement of the particle filling content in composites
YOU Jian, ZOU Weijian, JIN Gang, LEI Yu, ZHU Shichao, SONG Zhuoming
The Key Laboratory of Polymer Processing Engineering of Ministry of Education, The National Engineering Research Center of Novel Equipment for Polymer Processing, South China University of Technology, Guangzhou 510641, China
Abstract: Low density polyethylene (LDPE)/glass microbead (GB) composite was extruded by high pressure capillary rheometer at a steady temperature and flow rate, and the content of glass beads in the composite system was measured by ultrasonic wave. Two kinds of sound attenuation characteristic parameters of the sound attenuation variate S1 in the time domain and the characteristic vector M0 in the frequency domain were defined to describe the changes of the filler content. The linear regression (LR) and support vector machine regression (SVR) method were used to establish the quantitative models of the filling contents, which was compared with the offline ash comparison of measurement results. The experimental results show that the predicted filling content deviations of the two models based on the self-defined feature parameters were no more than 1.2% and the minimum values were only 0.17%, which can meet the requirements of on-line real-time measurement. The frequency domain method is more accurate than the time domain method, and it's accuracy. was similar to that of the off-line ash measurement.
Keywords: ultrasonic online measurement;composites;filler contents;SVR
2018, 44(12): 111-116 收稿日期: 2018-05-08;收到修改稿日期: 2018-06-23
基金项目: 国家重大科学仪器设备开发专项（2012YQ230043）；国家自然科学基金（11572129）
作者简介: 游剑(1994-),男,湖北武汉市人,硕士研究生,专业方向为聚合物超声在线检测
参考文献
[1] ZHAO P, PENG Y, YANG W, et al. Crystallization measurements via ultrasonic velocity:Study of poly(lactic acid) parts[J]. Journal of Polymer Science Part B Polymer Physics, 2015, 53(10):700-708
[2] KAZYS R, MAZEIKA L, SLITERIS R, et al. Measurement of viscosity of highly viscous non-Newtonian fluids by means of ultrasonic guided waves[J]. Ultrasonics, 2014, 54(4):1104-12
[3] PUTZ V, BURZIC I, MIETHLINGER J, et al. Adaptation of in-line ultrasonic velocimetry to melt flow measure-ment in polymer extrusion[J]. Measurement Science & Technology, 2013, 24(10):107002-107007(6)
[4] PRASAD V S K, BALASUBRAMANIAM K, KANNAN E, et al. Viscosity measurements of melts at high temperatures using ultrasonic guided waves[J]. Journal of Materials Processing Technology, 2008, 207(1-3):315-320
[5] STEINHOFF B, LELLINGER D, ALIG I. Monitoring of polymer melt processing[J]. Measurement Science & Technology, 2010, 21(6):062001
[6] KAZYS R, SLITERIS R, REKUVIENE R, et al. Ultrasonic technique for density measurement of liquids in extreme conditions[J]. Sensors, 2015, 15(8):19393-19395
[7] SUN Z, JEN C K, YAN J, et al. Application of ultrasound and neural networks in the determination of filler dispersion during polymer extrusion processes[J]. Polymer Engineering & Science, 2005, 45(6):764-772
[8] TATIBOUËT J, HUNEAULT M A. In-line ultrasonic monitoring of filler dispersion during extrusion[J]. International Polymer Processing Journal of the Polymer Processing Society, 2013, 17(1):49-52
[9] 林莉, 李喜孟. 超声波频谱分析技术及其应用[M]. 机械工业出版社, 2009.
[10] 王雷, 晋刚, 谢云, 等. 玻璃微珠填充PS复合材料中超声波传播特性研究[J]. 塑料科技, 2014, 42(4):43-47
[11] 谢云, 基于超声信号的无机粒子高填充聚合物分散状态表征研究[D]. 广州:华南理工大学, 2015.
[12] 苏高利, 邓芳萍. 关于支持向量回归机的模型选择[J]. 科技通报, 2006, 22(2):154-158
[13] 杜树新, 吴铁军. 用于回归估计的支持向量机方法[J]. 系统仿真学报, 2003, 15(11):1580-1585
[14] LI C H, HOU H, LIU Y L, et al. An automatic method for selecting the parameter of the normalized kernel function to support vector machines[J]. Geoscience and Remote Sensing Symposium, 2012, 45(1):836-839
[15] BRERETON R G, LLOYD G R. Support vector machines for classification and regression[J]. Analyst, 2010, 135(2):230
