Volume 17p66-71 (2017)
In the commonly used approach, the lifetime of a superheater is estimated by characteristic values of the production parameters and the operating conditions. In this approach, a lower bound for a superheater lifetime is based on some arbitrary safety factor that does not necessarily reflect real life, where unexpected failures do occur. The method proposed here suggests coping with this reality, by employing a techno-economic probabilistic approach. It comprises the following two models: • A probabilistic time to failure evaluation model that considers the variability of the lifetime determining parameters. • A model to optimise values of technical parameters and operating conditions and to determine a superheater's optimal replacement policy, based on life cycle cost considerations. The proposed probabilistic time to failure evaluation model can help to identify the most influential parameters for planning for a minimal probability of failure. It is applied to a unique problematic steel T22 superheater of rather specific parameters: corrosion rate, the Larson Miller Parameter (LMP), diameter and wall thickness. Sensitivity analysis has shown that the dominant factor affecting variation in superheater lifetime is the variation in the LMP, while the effect of the other parameters is quite marginal. Decreasing the standard deviation of the LMP (by keeping a more uniform material) lowered the probability of failure. This resulted in a practical recommendation to perform periodical checks of the parameter wall thickness. We also tested the effect of changing the nominal values of these parameters on the lifetime distribution. Hence, we suggest that the selection of the nominal values should be based on life cycle cost considerations; and propose a model to calculate, for any given combination, the average life cycle cost. The latter model, the optimal parameters combination model, optimises the combination of changes in all the superheater's parameters by minimising the average life cycle cost associated with the superheater. Demonstrating the usefulness of the proposed approach, in a problematic case, suggests that it can be beneficially employed in the more general case whenever the planned lifetime of a design is threatened.
Archived Files and Locations
|application/pdf 164.5 kB ||
access all versions, variants, and formats of this works (eg, pre-prints)