43. PGMON Meeting in Paris (October 6/7, 2011)

TOP 11: Degradation trajectories for components in power plants (Henk Wels, KEMA)

For components in E-production units the optimum time for overhaul is ill known. With a quantitative model the planned unavailability in relation to unplanned unavailability can be optimized as a function of operating conditions. Such a quantitative model should answer the question how much more or less failures are to be expected when the overhaul interval is lengthened. It can be expected that mechanical components show a degradation trajectory to failure. Therefore failure mechanisms and influence factors have been charted for such components. Use is made of FMECA results as well as KEMA failure investigations. The failure investigations have been used to obtain quantitative information for failure and degradation trajectories.

The failure investigations of power plants show that in practice there is a large amount of uncertainty in times to failure and degradation times. Even a failure database with 745 damages in some 30 types of components shows that there is still insufficient material to precisely estimate time to failure and degradation trajectories per failure mechanism. An uncertainty of a factor 2 in time to failure is not uncommon. Therefore expert judgment has been used to assess the probability of occurrence of a failure mechanism, the average time to failure and the degradation speed.

Degradation trajectories have been simulated with Monte Carlo analysis. The combination of expert judgment, FMECA, quantitative information from failure investigations and simulation is sufficient to calculate the optimum time between overhauls using Markov modeling of degradation trajectories.

Such modeling is easy to understand and elegant.

back...