A method of accounting for censored items as part of fatigue testing of composite materials
The Aim of the paper is to create a method for including information on censorship for the purpose of adjusting the estimates of the S-N diagram of composite materials. Censored items are such items that did not fail by the end of testing, for which a certain operation time has been registered. It must be noted that, currently, researchers often ignore the operation time data of the items that did not fail by the end of testing, which does not appear to be justified in terms of cost saving and
... of cost saving and reliability of statistical conclusions. Censorship information is very important in terms of assessing durability. It only needs the right tool to use it. The proposed Method consists in bootstrapping-based simulation, a method from the group of computer-intensive methods. In the process of the method's development, the previously used approaches (for instance, as regards metals) were considered. In the examined example of the method's application, the data was taken from literary sources. Results. The paper shows an example of fatigue testing conducted by the authors that produced a large number of censored items. The results obtained using the method were compared with real data. It is shown that the quality of statistical estimation improves with the use of the method. The paper sets forth certain observations regarding the mechanical testing tool for quality control. The source of data dispersion associated with fatigue testing is discussed. Conclusions. The application of the method will help include the information on censored items in the estimation of the S-N diagram. For scientists who are involved in the experimental research of the fatigue resistance of composite materials, the suggested method might prove to be quite useful. It takes into consideration the characteristic features of the strength analysis of composite materials (large properties dispersion and absence of unlimited fatigue range). The method will allow taking onto account an important, but not always so far used information on the items that operated a certain number of cycles, but did not fail.