The results of a statistical evaluation of ten impact test methods are presented in this paper. The concept of “good impact test” has two aspects: the quality and the applicability of the test results. The applicability of various test methods for predicting in-service performance is the subject of a companion paper. This paper presents an evaluation of the quality of the test output, as well as an analysis of the correlation between various test outputs.
The ten test methods evaluated—three pendulum, one tensile, four drop dart, one flex, and one driven dart—are described briefly. Three of the test instruments were instrumented. Seven materials, ranging from unfilled thermoplastics to reinforced thermosets, were tested by each of the methods.
Evaluation of a test's output quality is developed in terms of the range of applicability and the degree of material differentiation. Specific factors that serve as criteria include the existence and distinctness of failure, and the spread and scatter of measured results. The test methods are ranked from “highly desirable” to “severely restricted.”
Evaluation of the correlation of material ranking by the various test methods is presented. In general, results from different impact tests do not correlate. However, there are instances of strong or fair correlation. The degree of correlation is explained in relation to the similarity of test stress states and measured characteristics. It is meaningful to classify impact tests according to stress states and measured characteristics. Material selection for impact performance should be based on test methods that simulate the stress states, controlling variables, and failure limits of the intended application.