Open AccessВИЗНАЧЕННЯ ФУНКЦІЇ НАДІЙНОСТІ НЕВІДНОВЛЮВАНИХ ТЕХНІЧНИХ СИСТЕМ ПРИ НЕПОВНИХ ДАНИХ
Author(s) -
Олександр Володимирович Томашевський,
Геннадій Валентинович Сніжной
Publication year - 2019
Publication title -
avacìjno-kosmìčna tehnìka ì tehnologìâ/avìacìjno-kosmìčna tehnìka ta tehnologìâ
Language(s) - English
Resource type - Journals
eISSN - 2663-2217
pISSN - 1727-7337
DOI - 10.32620/aktt.2019.8.19
Subject(s) - weibull distribution , reliability engineering , reliability (semiconductor) , failure rate , exponential distribution , consistency (knowledge bases) , computer science , survival function , probability density function , function (biology) , mean time between failures , statistics , survival analysis , mathematics , engineering , power (physics) , physics , quantum mechanics , artificial intelligence , evolutionary biology , biology
The study of the reliability of technical systems is especially important for systems whose failure to operate is associated with a danger to people's lives (for example, aircraft). The question of determining the reliability of technical systems based on complete information about their work is well studied, but in practice, we often deal with incomplete data. It is related to the fact that the time from the beginning of operation (testing) to the occurrence of failure is not always known and the data on the refusal may be incomplete (censored). Application of censored data determining reliability indicators will allow for more accurate estimates. Technical systems, depending on the characteristics of failure, can be restored or non-recoverable depending on the features of the failure. The main indicators of the reliability of non-restorable systems are the probability of failure-free operation (reliability function), the probability density of failure, the average run-up to failure, the failure rate. You can determine the reliability function for incomplete data applying STATISTICA Survival Analysis module. This module has been widely used in medical research to assess the survival function and can be applied to assess the reliability of technical systems. The Survival Analysis module contains procedures to determine the analytic form of the reliability function (survival) based on the assessment of the correspondence of the empirical distribution obtained from the experimental data with the given theoretical distribution. In conducting the study, consistency checks the following distributions: exponential, Weibull, Gompertz. The distribution parameters most closely related to the empirical one are determined using the least-squares method. The conformity assessment is carried out using Pearson's criterion. As an example, the definition of a reliability function for incomplete data examines the results of tests of the main element of the avionics units - the integral microcircuit. It is found that the distribution of Gompertz is the closest to the empirical distribution of the working time until the failure.