A simulation methodology for large scale reliability systems

Abstract

In recent years, the assessment of the reliability and availability of complex systems has played an increasingly important role in the analysis, design, and operation of these systems. A very important element in the design and operation of a system is the estimation of the impact of the logistic support on the system availability. Analytical methods for this complex problem are in general intractable, and simulation techniques that specifically address the problem are scarce and with limited capabilities. This research provides a powerful methodology that can model the effect of such logistic characteristics as inventory, transportation, and facilities on the reliability and availability of large scale systems. Some of the most important features of the current implementation of the simulation methodology are: (1) No assumptions about the topology of the reliability block diagram are required, except that it must have a single source and sink node. (2) In-site and off-site repair facilities and the effect of the number of repairmen on the availability and reliability of the system are treated. (3) Different failure modes are allowed for each of the lowest level components. (4) The effect of the inventory level of specified components on system availability is predicted. (5) Delays such as inspection and transportation times can be included. (6) Input and output formats are specifically designed for large scale reliability problems. (7) The methodology is straightforward and very easy to use. The methodology is implemented via a program written in Pascal (SIMULAV). It consists of three main elements: a process oriented simulation language, a minimal cut-sets algorithm, and a simulation model. The simulation model estimates the reliability and availability of a complex system considering its logistic support. It is implemented using the process oriented simulation language and it uses the minimal cut-sets to determine the status of the system (up or down). The methodology is partially verified by modeling a problem for which an analytical solution exists. Also, the current implementation of the methodology is compared against an implementation written in a general purpose simulation language (SIMAN). It is shown that general purpose simulation languages are not suited for implementing this methodology. A complex reliability system is modeled to show some of SIMULAV's capabilities.