Sample Size and Error Level for Testing Single Fiber Strength  
Author Ningxiang Wu


Co-Author(s) Liyang Xie


Abstract The accuracy of single fiber strength prediction is associated with sample size. This paper discussed the effects of sample size to the variation of Weibull parameter estimation in detail by Monte Carlo simulation. By simulation results, the assumptions that σ-bar-hat is subject to normal distribution and m-hat is subject to lognormal distribution were presented and the models of confidence bounds calculation for σ-bar and m were derived based on the relationship between sample size and maximum permissible error level. Both the models were verified by Likehood Ratio method using simulation data and experiment data.


Keywords single fiber strength, Weibull distribution, sample size, error level, interval estimation
    Article #:  22338
Proceedings of the 22nd ISSAT International Conference on Reliability and Quality in Design
August 4-6, 2016 - Los Angeles, California, U.S.A.