Reliability and Availability Analysis

Abstract

This study presents a critical review on the following topics: Some important continuous distributions. - Methods to estimate the parameters.

- Methods to estimates the indices of reliability

This study presents a proposed mathematical model which characterizes the distribution of the time to failure for the system components, whether the distribution has a monotonic hazard such as Weibull, Gamma, Normal distribution....etc., or non monotonic hazard and decreasing failure rate DFR in wearout region such as lognormal, Burr and inverse Gaussian distributions, consequently we can also estimate the mean life time of the component. In the design and manufacturing of both consumer and commercial products an important design criteria is the length of the warranty period measured in calendar time. It is very important to know the expected number of failures during the warranty period. This study also provides a statistical method that will predict the cumulative failures over the warranty period based on laboratory life data and the usage rate data for the products that are operated for only a limited fraction of the total available time. This method uses joint distribution of the usage rate distribution and the laboratory life distribution to translate usage time to calendar time and not the mean usage since the product use in field also follows abroad statistical distribution. The proposed technique permits an accurate calculation of the cumulative field failure over prospective warranty periods and thus is useful as a statistical method in design optimization as well as in providing decisions on product release.