Monte Carlo Simulation of Reliability Problems
NORTHWESTERN UNIV EVANSTON IL DEPT OF MECHANICAL AND NUCLEAR ENGINEERING
Pagination or Media Count:
The procedures for analog Monte Carlo simulation of Markov processes are examined. Two variance reduction techniques are then included in a nonanalog formulation to increase the sampling efficiency for highly reliable systems,, and a method for incorporating uncertainty in failure and repair rate data is outlined. Models for three classes of component dependencies appearing in reliability and availability problems are incorporated into the Markov formulation. They are 1 shared repair crews between components, 2 load sharing between components, and 3 standby mode. Results are given for a series of model problems to demonstrate the efficiency of the methods as well as the effects of the dependencies on system unreliability and unavailability.
- Operations Research