Dynamic Reliability Modeling

Dynamic simulation allows you to develop a representation of the system whose reliability is to be determined, and then observe that system’s performance over a specified period of time.

The primary advantages of dynamic simulation are:

   The system can evolve into any feasible state and its properties can change suddenly or gradually as the simulation progresses; and

   The system can be affected by random (stochastic) processes, which may be either internal (e.g., failure modes) or external.

In order to run a dynamic simulation, you must specify the duration of the simulation (e.g., 1 month, 1 year) and the length of the timesteps that you will use (i.e., the degree to which time will be discretized).

The decision on the simulation duration should be driven by the system variability you wish to capture with your model, since the duration is the amount of simulated time over which the system’s components will be allowed to interact and evolve as the real system would.  Typically you would use something such as the system’s serviceable life as the simulation duration.

Note, however, that depending on the system being simulated, there are several ways to approach the duration.  If you have a system that has a fixed service life (and the failure mechanisms have similar time scales to the service life), a dynamic Monte Carlo simulation in which the duration is the service life of the system would be the appropriate way to capture the behavior of the system.  However, if you have a system that operates for a very long time period with repeated failures and repairs (such that it effectively reaches a "steady-state" condition), you may want to run a single realization of the system with a very long duration.  Over a long duration, the variability in the failures and repairs will be adequately represented without running Monte Carlo simulation.

The specified number of timesteps is the minimum number of times that GoldSim will recalculate and update all of the elements in the model. The number of timesteps required to accurately model a system depends to a large extent on how you’ve built your model. 

Note that GoldSim will interrupt the simulation and force it to update when a repair or failure occurs, even if that failure or repair does not occur on a fixed timestep.

Nevertheless, some models may require a significant number of timesteps in order to generate valid results.  For example, if your reliability elements have failure and repair mode parameters that vary dynamically, have acceleration factors which vary dynamically, utilize user-defined failure modes, or have logic trees linked to standard GoldSim elements such as Reservoirs or Pools, you need to ensure that the number of timesteps is sufficiently large that these dynamic changes are represented accurately.

   Note: Details of GoldSim's dynamic timestepping algorithm, including a discussion of selecting the proper timestep for a model, are presented in Appendix F of the GoldSim User's Guide.

Related Topics…

Learn more about: