硅橡胶表面硬度的激光诱导击穿光谱分析

硅橡胶材料的老化状态是输电线路运维的重点关注内容,现有研究认为硬度是表征其老化程度的重要指标之一。目前复合绝缘子的伞裙硬度主要是在停电间隙用邵氏硬度计测量,现场应用限制较大。激光诱导击穿光谱(laser-induced breakdown spectroscopy,LIBS)是一种元素测量技术,利用高能脉冲激光烧蚀样品表面产生等离子体,分析等离子体光谱得到样品成分元素信息。已有LIBS研究中某些材料的表面硬度与材料特征元素的离子原子谱线强度比值以及等离子体激发温度之间存在较强的相关性,但在复合材料中这种相关性较弱,通常需要对大量谱线进行筛选以得到满足条件的谱线,且对全谱信息的利用程度不高。研究了不同氢氧化铝、白炭黑填料含量的高温硫化硅橡胶表面硬度与LIBS光谱的关系,发现表面硬度与填料含量之间具有较强的相关性,并通过主成分分析对不同硬度的硅橡胶光谱数据进行降维,结合神经网络可对不同硬度的样品进行区分。结果表明,LIBS技术与机器学习算法相结合可以准确测量硅橡胶硬度,这种方法利用了更多光谱信息,且不需要在众多谱线中寻找特征分析线,因此分析精度与速度得以提升。

Measurement of Surface Hardness of Silicone Rubber via Laser-induced Breakdown Spectroscopy

Ageing condition of silicone rubber material is an important content raising serious concern in operation and maintenance of power transmission lines. Some researches show that hardness is one of the most important properties related to aging condition of silicone rubber composite material. Currently the most common way to measure the surface hardness of silicone rubber material is to use Shore Tester, while it is not practical for field measurement. Laserinduced breakdown spectroscopy(LIBS) is an elemental measuring method. During the test a sample is ablated with a pulsed laser, then plasma is induced. After a few microseconds,the plasma emits line spectra to be used for calculating the ingredient information of the sample. Some studies show that there are strong linear correlations among surface hardness for some kinds of material, the ionic-to-atomic line intensity ratio of the characteristic element in the material and excitation temperature of the laser-induced plasma. However, it is found out that the above relationship is weak for composite material,and it is time-consuming to search some specific lines from the spectral data set for this method, and little use is made for the whole information. This paper studies the relationship between hardness of silicone rubber for different filler content and corresponding LIBS spectra, and a strong dependency between hardness and filler content is discovered. The spectral data dimension of silicone rubber for different hardness is reduced with principal component analysis(PCA) method, then theprincipal components are used to distinguish samples with different hardness combined with artificial neural network(ANN) algorithm. Results indicate that combination of LIBS technique and machine learning algorithms works well on hardness measurement of silicone rubber material. This method can make full use of the data without spectral line selection process so that the measurement accuracy and testing speed are improved.