Gas Turbine Maintenance Policy: A Statistical Methodology to Prove Interdependency Between Number of Starts and Running Hours

Abstract

In general, two approaches have been followed so far in gas turbine maintenance procedures to determine correct inspection intervals: “no interdependency” or “interdependency” between number of starts and number of running hours. The first approach is based on the assumption that starts and running hours induce different deterioration processes not correlated to each other. Accordingly, the number of starts defines the life limit for cyclic duty operation where low-cycle fatigue phenomena dominate, while the number of running hours define the continuous operation life limit for which erosion, corrosion and creep are controlling factors. The “interdependency” approach instead assumes that failure is produced by combination of low-cycle fatigue and continuous degradation mechanisms: in this scheme the frequency of starts becomes a fundamental parameter in order to determine the optimal maintenance interval. A statistical and reliability engineering methodology to validate the first or rather the second line of action is described in the paper. The population on which the study was conducted is made up by GE Oil & Gas PGT10 gas turbines that are in operation worldwide with fleet operation totaling 1,5 million hours. Most of the cases examined consist of mechanical drive applications for natural gas production, storage and transportation, with significant combination of both intermittent and continuous operation. Hot gas path components have been chosen for examination in consideration of their sensitivity to effects of both cyclic fatigue stress and wear mechanisms. The analysis concentrates on transition piece and combustion liner, both having scored a number of failures statistically significant for the purposes of this study. This analysis is considered the key to optimize inspection intervals and therefore achieve extended machine life. The methodology, based on Weibull data analysis, has been applied to a restricted sample of machines that operate in “standard” conditions, corresponding to gas fuel utilization, mechanical drive service with homogeneous load factor and very low number of trips. The study shows that interdependency between starts and running hours does exist and, given the number of starts, the corresponding running hours can be evaluated, and the inspection intervals appropriately predicted. Further developments of this study will be aimed at evaluating maintenance factors for “non standard” conditions such as dual fuel combustion systems, generator drive and operation with higher number of trips etc.