An informed approach can extract more useful operation from components
Credit: Siemens

It is only through a full recognition of the way that various factors interact that the engineering lives of gas turbine components can be assessed, writes Michael Wood

Gas turbine hot section parts are high value components which have finite lives. Whether these lives are exactly as indicated by the original equipment manufacturers, or are amenable to informed review and revision by the owners and operators, is part of the dynamics of the interaction of these parties.

How this plays out depends on commercial arrangements – the presence of long-term service agreements or the state of the electricity market – and the technical issues underlying the subject. Nevertheless, gas turbines’ durability and repairability has a vital role in controlling maintenance intervals and through-life costs for a plant. In the last 10-15 years we have had the increasing preponderance of F class technology in the fleet mix. Additionally, due to changing circumstances, many machines nominally optimized for baseload are being required to cycle, sometimes quite frequently.

While these alter some of factors in the balance of considerations, the basic fundamentals remain: how long can the components be operated reliably, and how does one make well-informed decisions on this subject?

When considering the engineering aspects of this, it must be remembered that the integrity and durability of the component is dependent on a number of relatively independent factors. These include the adequacy of the basic design (choice of material, metal temperature and transient stress levels), the standard of manufacture, the capability of repair technology and the way in which the unit has been operated and maintained. It is only through a full recognition of the way that these factors can interact that the useful safe engineering lives of the various components can be assessed.

Component damage processes

There are a multitude of different metallurgical degradation processes. Some of these processes are more reversible during refurbishment than others. A general overview of the types of damage for rotating blades is given in Table 1 (a somewhat similar table could be given for the static nozzles). A qualitative ranking is given of their relative importance for units operating under conditions ranging from fairly steady state, through to cyclic (stop-start) conditions. Most of these processes are well known and have been discussed many times in the literature. However, the use of newer, lower chromium content superalloys have thrown up some new damage processes which need to be taken into account.

table 1

For units with many stop-start cycles the main life, or overhaul interval, limiting factor will probably be thermal fatigue, as seen in Figure 1. If the cracking is ‘minor’, confined mainly to the coating, it can be ‘removed’ during the replacement of the coating during refurbishment of the component. However, if it has grown into the blade then the amount of growth will govern the repairability of the blade, although this will still depend on where the cracks are in the blade. Those near a squealer tip can be repaired, whilst those on the leading/trailing edges transverse to the aerofoil are often unrepairable.

Figure 1: thermal fatigue cracking
Credit: ERA Technology
figure 1

In base/intermediate load units, the factors controlling the overhaul interval, and which ultimately are the life limiting factors, will generally be creep, microstructural instability effects and environmental attack. Creep certainly does take place in industrial gas turbine blades (usually at a low level), as do microstructural changes due to alloy instabilities or coating/substrate interdiffusion. The latter can be reversed,to a large degree, during the recoating/refurbishment cycle. Coating durability has a great influence on the overhaul intervals, but not necessarily on the component’s ultimate life.

Assessment methodologies

When considering the practical ways to assess the condition of components, one must be mindful of the changes in technical sophistication between E class and F class components. Reduced to a simplification, this amounts to the use of more advanced alloys and coatings, combined with improved cooling technology (more derived from aerospace practice) and smaller effective safety margins (as the usually more generous ones used in the older designs are not ‘required’).

There are two basic approaches which are currently employed in assessments. At first sight these may appear to be contradictory, but they are merely different aspects of the same issue. However, the balance between them, and their requirements, is influenced by the change in technological complexity in going from E to F class machines – that is, more engineering input is required as one moves towards the more complex F class type components.

The first approach is based around a ‘design type’ approach and involves three sequential phases.

Phase I: Calculate the consumed life. This requires a detailed model of the component along with material deformation and damage rules and temperature information;

Phase II: Non destructive examination using conventional techniques;

Phase III: Destructive evaluation of components using metallography and mechanical testing to further ‘quantify’ the damage.

Various organizations have followed this approach. Quite how ‘precisely’ the lives can be recalculated is open to debate, but it undoubtedly provides a far more quantitative indicator of ‘life consumption’ or damage accumulation than was possible previously.

The second basic approach appears to be almost the reverse of this. It is: non-destructive inspection; destructive examination; calculation of condition, if required. This is actually the most common approach across the industry as few organizations are in a position to calculate life consumption. Whilst this approach may be perceived as not being that sophisticated, the use of contextual information, trending the behaviour of components so that the recommended lives can be aligned with actual field behaviour, has been used very successfully over a number of years, in the absence of detailed design information.

Apart from external cracking, most of the other degradation phenomena can only be adequately characterized through a destructive examination of the component. This may be seen as an expensive way of acquiring information, but it is unfortunately, at present, still the most informative approach. Non-destructive techniques, whilst improving in their ability to detect and size cracks and defects, have not been able to produce useful information on other aspects of the blade’s condition.

The only main exception concerns eddy current-related work which can now be used to characterize the chemical state of a coating (as opposed to the presence of cracks), and hence its useful remaining protective life.

As things stand at the moment, the most common examination technique is still metallography. I

n this, one is looking for changes in the alloy’s microstructure or the appearance of features such as creep void formation, cracking or excessive surface attack (oxidation/corrosion). In that sense the generality of the approach used for F class components differs little from that used successfully for E class components: ERA has used this type of approach on over 200 sets of E class components, in some cases allowing life extensions of 25-75 per cent of the originally indicated lives to be achieved (comparable work has also been carried out by ERA on F class components).

Deeper understanding

However, because of the increased use of directionally solidified or single crystal alloys, a deeper understanding of alloy degradation and ‘life consumption’ characteristics is required.

Further, there are ‘new’ issues, such as the risk of recrystallization during refurbishment, which need to be considered. Additionally, because of the increased complexity of the stress/temperature fields in the component, the regions meriting close examination can be best identified in conjunction with modelling work on the components, and in conjunction with the results from conventional non-destructive examinations.

The main strengthening phases in the alloy are fine precipitates and metal carbides, although carbides are less directly relevant in directionally solidified and single crystal alloys. Whilst changes in these features can be extremely informative in many ways, the information is generally used in only a semi-quantitative manner. There are very few quantitative relationships between what can be observed microstructurally and the level of ‘damage’ which has been sustained. Despite this, microstructural appearance is widely used to pass judgment on the serviceability of components.

Test pieces can be removed from components but there is no commonly accepted approach to determining the relevant residual mechanical properties. This problem arises partly out of a practical requirement to carry out a short duration test (10-100 hours) to assess a property which is of relevance over component service durations (25,000-75,000 hours). However, they are still useful in providing further information on the condition of the blade’s microstructure.

Destructive evaluations also permit an examination of the internal surfaces of the component. As has already been noted, some blade designs show cracking initiating at the inner surfaces, with these leading to blade failure in some cases.

Generally, although the observations are often not quantified in detail, they are exceptionally informative. Particular attention needs to be paid when examining low chromium content superalloys, since they can exhibit a greater sensitivity to environmentally influenced crack formation on the internal surfaces.

In all of this it must be appreciated that very few components are ever examined in such detail. When such observations are used to make an assessment of a set of components it therefore requires the degree of uncertainty to be carefully considered.

When one is assessing the tolerance of a structure to (engineering) cracks, then the more informed approach is to carry out modelling of the component (that is, stress analysis under transient conditions) to assess potential growth rates and defect tolerance. This applies to both cracked and repaired components (where one may be concerned with re-initiation in the repaired zone). Whilst this is necessarily more complicated and time consuming, it does allow a more thorough assessment.

Finally, one must take into full consideration the capabilities of repair technology. This is in the context of both removing cracks or repairing cracked regions and other mechanical damage, as well as replacing coatings, especially internal coatings. There is also the important effect of reheat treatment in reversing (most of) the microstructural changes resulting from service. Nevertheless, components cannot be repaired indefinitely: they do have real finite lives.

These points are, in part, a reflection of a greater need for the operator, possibly in conjunction with a third party or the OEM, to have a fuller view of the degradation of the components and their ‘true’ level of life consumption.

This ultimately involves the associated management of the risks of operation, particularly when considering operation beyond the original notional design envelope. However, the application of this type of approach allows a more informed approach to be taken with the potential of extracting more useful operation out of the components.

Michael Wood is gas turbine consultant at ERA Technology.

More Power Engineering International Issue Articles
Power Engineering International Archives
View Power Generation Articles on