1. Science
  2. Publications
  3. Information Processing Systems
  4. 1(152)'2018
  5. Estimating the system lower confidence limit of physically realizable system using normally distributed and statistically independent load-strength models

Estimating the system lower confidence limit of physically realizable system using normally distributed and statistically independent load-strength models

V. Dubnytskyi, A. Kobylin, O. Kobylin
Annotations languages:


Description: The method of lower confidence limit estimation of physically realizable system using normally distributed and statistically independent load-strength models is offered. To calculate the intervals of the statistically independent load and strength distribution histograms, empirical kurtosis distribution is used. It is suggested to consider each interval of the obtained histograms as a classically calculated interval number. In order to estimate system confidence interval, the ratio level of each interval of generalized load histogram to each interval of generalized strength is calculated. The system confidence interval is estimated by the lower tail of the frequency estimate for those cases where the left boundary of the interval ratio is more than one. For "normally distributed load and strength" version the analytical reliability calculations are performed. As a result of its comparison with the statistical analysis, it is shown that for an operated system the suggested method gives more cautious results. The suggested method can be used to determine the confidence level of a system operating under the conditions of severe environment.


Keywords: confidence, histogram, Gamma Distribution, normal distribution, kurtosis, interval measurements

Reference:
Dubnitskii, V.Iu., Kobilіn, A.M. and Kobylin, O.A. (2018), “Otsenka nizhnei granitsy nadezhnosti fizicheski realizuemoi sistemy v protsesse ee ekspluatatsii pri proizvolnykh zakonakh raspredeleniia obobshchennoi nagruzki i prochnosti” [Estimating the system lower confidence limit of physically realizable system using normally distributed and statistically independent load-strength models], Information Processing Systems, Vol. 1(152), pp. 53-60. https://doi.org/10.30748/soi.2018.152.08.