Price: 10.00
Tensile testing of polymers usually involves measurement of three quantities: tensile strength,
elongation (i.e. strain at break), and modulus. Each of these quantities is subject to measurement
error. The total error for any property is made up of component errors from several sources. A
sensible approach to error reduction is to analyze the entire measurement process, estimate the
magnitude of each separate error source, and then take steps to reduce the largest errors. A very
useful tool for this purpose is Analysis of Variance, a statistical tool whose name suggests its
function, the separation and estimation of sources of variance (which is a measure of variability).
This paper describes work which was done during a major quality improvement project for reducing the error in tensile testing. One Quality Improvement Team had previously carried out studies using Analysis of Variance, and had discovered that, in the case of polymers having a Tg well above room temperature which were UV-cured near room temperature, a major source of error arose from factors occurring before testing was done. In other words, samples taken from the same source and prepared under supposedly identical conditions did not have identical physical properties. In contrast, polymers having a Tg near or below room temperature did not have this problem. A second team was formed to continue the work of the first. It began by trying to identify all of the factors which could affect the condition of a polymer at any stage during preparation, or at any time prior to testing. The list of potential causes included not only deliberate steps in sample preparation, but also accidental or unintentional factors.
