Reliability of products or systems can be estimated by event based simulations. In these simulations the components of the product and repair resources are modelled as objects which can be in one of a number of states. For the components these states are 'waiting for next failure' event, 'waiting for repair resource' and 'waiting for repair completion'. The time of the next failure and repair completion are sampled from the failure and repair distributions and waiting for repair resources is based on queuing for those objects. Each change of state of an object is an event. The repair resources have their own states and events. At each component event the simulation checks all the RBD networks and accumulates the total down time and number of failures for each RBD. At the end of the simulation the availability, MTBF, and MTTRS can be estimated from the total down times, number of failures and total simulation times.
It is not difficult to program an event based simulation, but it is not usually necessary since there are suitable commercially available reliability simulation tools. Optimal Solutions 4U uses AvSim+.
The advantages of reliability simulations are:
A particular strength is that blocks do not have to be statistically independentTwo blocks or elements are statistically independent if knowing that one of them has failed makes it neither more probable nor less probable that the other block has failed. If blocks are not statistically independent then conditional probabilities have to be used to calculate the overall probability of failure of service. which is of great assistance in developing complex models of complex services. In addition AvSim+ provides interactive graphical methods for defining RBDA diagram that links the elements of a product that need to be working for the product to work. So long as there is a path from the start of the diagram to the end of the diagram the equipment can provide the service. The diagrams can be hierarchical so that blocks at one level can be decomposed into an RBD at a lower level. and Fault treesA graphical method of modelling the reliability of the system as AND and OR gates and failure events..
The main disadvantage is that the models do not give exact results with the consequence that run times can be quite long for complex systems with high reliability. If the repair times do not vary much, the accuracy of the simulation result will be given by the number of failures (for the service or RBD) that occurred during the simulation. This will be given by a Poisson distributionA statistical distribution that gives the probability of obtaining exactly n successes in N when the probability of obtaining a successful trial is low. It is the limiting case of the Binomial distribution when the probability is low.. so that for 100 failures the results should be accurate to 17%.
Optimal Solutions have considerable experience in developing reliability simulations, and welcome enquiries on Reliability. We would be pleased to provide consultancy tailored to your requirements. You can get in touch by sending a message from our Contact Us page, or by calling us on the number below.