THE 19 TH INTERNATIONAL CONFERENCE ON COMPOSITE MATERIALS 1 General Introduction Single fiber tensile tests are often conducted to measure fiber properties of high strength polymer fibers, such as PPTA (poly(p-phenylene terephthalamide), used in soft body armor (SBA). These tests are frequently conducted at slow deformation rates (quasi-static) [2] relative to rates that occur in fibers during ballistic impact. To overcome this measurement challenge, the Kolsky- bar apparatus has been utilized and now allows measurement of the single fiber tensile behaviors at high strain rate (HSR) [1] deformations comparable to ballistic impact. This study examined single PPTA fibers conducted at quasi-static conditions using two different gripping techniques with several different gauge lengths to investigate the gripping effects on the tensile tests. The test results were analysed in both parametric and nonparametric methods to investigate the distributions of the test data. Statistical analyses were performed to 1) compare the performance of the direct and glue-tab gripping under quasi-static and HSR conditions and 2) assess several models for the data. Distributional models are useful for characterizing material properties based on distributional parameters. In addition, distributional models are often incorporated into strength models. Based on the analyses obtained by the quasi-static tests, a fiber gripping method will be developed for high strain rate test. 2 Results and Data Analysis 2.1 Strength Data Single fiber tests have been typically conducted by using glue-tab grips with fiber lengths having a higher aspect ratio than 2000. However, HSR tests should be performed at much shorter gauge lengths in order to achieve force equilibrium during the dynamic loading. For the test with the shorter gauge lengths using the glue-tab grip, not only contributions of end effects but also the wicking of adhesives is a concern. So an alternative gripping method, the direct grip, was investigated. This method is clamping a single fiber directly and thus has no adhesive wicking. From the quasi-static tests, we measured tensile strengths for two gripping methods (glue-tab and direct grip) at four gauge lengths (2 mm, 5 mm, 10 mm, and 60 mm). Since the strength distributions for both tests were significant, the statistical analyses were carried out and analyzed in next sections. 2.2 Non-Parametric Analysis The gripping methods were compared graphically using kernel density [4] and quantile-quantile (Q-Q) [3] plots. More formally, two-sample Kolmogorov- Smirnov (KS) tests were conducted to see if the data can be described by a common distribution. Kernel density plots provide a quick summary look at a univariate set of data and can show features such as: 1) the center (location) of the data; 2) the spread (scale) of the data; 3) the skewness of the data; 4) the presence of outliers; and 5) the presence of multiple modes in the data. The kernel density estimate is defined as fy  ∑ {     }    (1) STATISTICAL ANALYSIS OF SINGLE PPTA FIBERS N. Heckert 2* , J.H. Kim 1 , W. McDonough 1 , K. Rice 3 and G. Holmes 1 1 Materials Science and Engineering Division, 2 Statistical Engineering Division, 3 Law Enforcement Standard Office (OLES), National Institute of Standard and Technology, Gaithersburg, USA * Corresponding author (alan.heckert@nist.gov) Keywords Single fiber, soft body armor, high strain rate, statistical analysis