Influence of Statistical Distributions on Availability and Inspection Interval of Protective Devices

Protective devices are designed to protect people, the environment and material assets under emergency situations. If protective devices do not work well, serious consequences may be resulted. It is critical to pay special attention to their maintenance. For this reason, many availability models have been developed to obtain an optimal inspection interval and to maximize their availability. However, few attention have been paid to the relationship between the statistical distributions used to describe the lifetime of protective devices and their optimal inspection interval and maximum availability. Furthermore, the problem that might occur when the normal distribution takes negative values has not been considered yet in protective device maintenance. This thesis aims to calculate the optimal inspection interval and maximum availability for the Weibull, normal, truncated normal and exponential distributions. Also, the relationship between these statistical distributions, and the availability and the inspection interval is studied. Finally, this thesis intends to study the problem that arises when the normal distribution might take negative values. To meet these objectives, an existing availability model, which considers constant time between inspections, is adapted to the Weibull, normal, truncated and exponential distributions. After adapting the model to each distribution, the effects of each distribution’s parameters on the optimal inspection interval and maximum availability are analyzed. It is not recommended to use the normal distribution if it has a large number of negative values while the truncated normal distribution is suggested as a possible approach to replace the normal distribution. This analysis help us to have a understanding on what is the performance and limitations of each of the four distributions.Outgoin