Modern industrial gas turbines must be capable of operational flexibility to fulfill the requirements of a changing power industry. At one extreme, turbines are required to run for extended periods of time at full load conditions to satisfy base load applications. At the other extreme, turbines are exposed to high cycles to satisfy peaking applications. Industrial gas turbine components that operate in the hot gas path are therefore subjected to loads that can lead to damage from both creep and fatigue.
The traditional design approach is based on worst case operational scenarios. These worst case assumptions compound and typically result in a single damage mechanism and bounding location, which represents the minimum life of the turbine. In contrast, condition based assessments tend to consider multiple operating scenarios, which may result in different damage mechanisms and potentially different bounding locations. Condition based assessments of gas turbines therefore require sophisticated lifing routines that can identify these peak damage locations for a range of damage mechanisms and load states, to ensure reliability and availability of industrial gas turbines for flexible operation.
Digital assets are a key element of modern condition based maintenance. However, a large fleet of diversely operated gas turbines results in an equally large number of digital assets. This produces specific challenges around data processing and storage of large machine datasets. Full order methods, such as Computational Fluid Dynamics (CFD) and Finite Element Analysis (FEA) are traditional approaches to predicting temperatures and loads on key components in industrial gas turbines. These models are computationally expensive and are therefore suited to idealized engine cycles. However, digital assets are used to analyze actual operational data and therefore require Reduced Order Models (ROMs) to manage large data sets within a reasonable time frame for optimal asset management. Key to this is determining the bounding assessment locations.
Structural integrity design codes provide a number of methods for assessing key damage mechanisms. This paper presents a comprehensive framework to use elements from these various codes to identify life limiting locations. This is a critical aspect of developing effective digital assets for condition based asset management. The approach provides a qualitative assessment of damage based on simple elastic finite analyses for typical engine load cycles. The method identifies locations with a high potential for the fundamental damage mechanisms including: creep, fatigue and oxidation and it allows for the assessment of elastic shakedown and ratcheting. Creep assessments in design codes typically consider only stress rupture. However, for cooled gas path components the stresses are both thermally and mechanically driven. The stress rupture approach is therefore insufficient to capture the full effects of creep and may misidentify key locations. The approach presented here addresses this shortfall by additionally considering creep strain due to stress relaxation, in addition to other damage mechanisms.