The growing need for multi-starts could seriously injure your assets, says Gary Lock
As the energy industry increases its uptake of power from renewables, the level of demand on large frame gas turbines becomes less predictable, with start-ups triggered by the vagaries of the weather and the economic market.
Gas turbine operators are having to multi-start machines, sometimes as often as twice a day. Indeed, multi-starting has become the norm – it is common to hear of over 500 starts a year – with superfast ramp-up periods. What damage is this causing to machines that were designed, effectively, for continuous operation? And what might be the resulting risks and cost implications for the industry?
Gas turbines were initially designed to facilitate base loading with minimal starts but with industry changes, multi-starting is putting high-integrity components through multiple strain cycles for which they were not designed. Making assumptions that components can withstand these requirements – based on the generalization that a defined number of starts is equivalent to a particular number of operating hours – is at best optimistic, and at worst could be dangerous.
A key change in the impact on components is that, in traditional use, the dominant failure process for hot components was creep. They were thus designed using creep-resistant alloys and their stresses were controlled appropriately. In the new era of multi-starts, the components experience high thermal transients during startup, and this, together with ever-shortening ramp-up times and increasing operating cycles, has a significant detrimental effect on their integrity.
Multi-starts mean the stressed components in gas turbines are now subjected to low cycle fatigue and, for many, this becomes the dominant failure criterion. Hot components have the added problem of creep, as outlined above. This, when combined with low cycle fatigue, can reduce the component lifespan dramatically below that expected of each of these failure criteria in isolation. Historically, assessing low-cycle fatigue damage has been problematic. While inspection and evaluation of creep damage is fairly straightforward, for low-cycle fatigue analytical methods provide the most effective way of knowing the extent of life remaining in a component.
The difficulty in assessing low-cycle fatigue damage lies in the physical differences between the characteristics of creep failure and low-cycle fatigue during the strain cycle. In creep failure, materials develop defects in features such as grain boundaries, over time and under the application of steady stresses and temperatures, due to the gradual accumulation of strain by diffusional processes. Creep-resistant materials are developed with large-grain structures to minimize the potential for the onset and accumulation of damage, which generally appears at the grain boundaries.
Low-cycle fatigue, in comparison, is characterized by the accumulation of micro-cracks, which appear in regions of high strain. These micro-cracks then coalesce to form large cracks and these can propagate quickly through the grain structure. Low-cycle fatigue-resistant materials are typically developed with small grain structures to cut the potential for crack growth in any given cycle.
Analysis typically focuses on up to ten components where, should they fail prematurely, a safety-critical or financially unacceptable impact would result. Each component is modelled using a Finite Element Model (FEM), which accurately represents its geometry, its interaction with mating components, and its load cycles. The FEM model is then subjected to startup, steady state, and cool-down cycles of thermal and mechanical loads, and the results measured. Typically, the models are non-linear to enable them to accurately represent the gaps, friction and material yield and creep deformation that generally occur in hot gas path components. Post-processing of the resulting strain cycle produces a calculation of the damage accumulated during each cycle type.
Methods for calculating the damage differ depending on the material, and are based on either ductility exhaustion or strain energy density. For ductility exhaustion, it’s essential that we understand the dependence of the material’s ductility on complex variables, including strain rate and temperature.
Alloys, for example, will usually have greater ductility at high strain rates than at low strain rates. Thus, whether strains occur during long periods of constant load or during loading cycles will affect the damage calculation. The available ductility of the material is gradually used up as the strain accumulates during a constant load period or a start/stop cycle, until eventually it is fully exhausted.
Similarly, strain energy density-based methods will calculate the work undertaken by the material each time it goes through a load cycle. From this analysis we can develop a detailed understanding of how damage accumulates during varying load cycles and periods of constant load. This, in turn, provides us with significant insights into, and a deeper understanding of, the effects that result from different operating regimes.
With analysis demonstrating the low-cycle fatigue and creep stresses that multi-starts impose upon components, and the resulting impact that this has on their lifespan, this poses several issues for operators. The first is the need to be aware of the damage taking place, so that regular inspection, maintenance and replacement can be factored into asset management strategies and schedules.
Consideration of the ‘bigger picture’, however, poses a more complex question: Are multi-starts cost-effective – do the overall financial benefits really outweigh the costs of regularly ramping up the turbines so quickly? Taking the damage that is being caused to assets – together with associated costs such as time lost to equipment shutdowns and time spent in repair activity by staff – into consideration, an operator’s apparent profit could be significantly eroded.
Of course, apart from the financial aspects, operators must consider the potential health and safety implications of asset damage. As component stresses begin to occur, the possibility of a premature catastrophic equipment failure increases. These failures might involve blades, veins or even discs, and could cause both internal damage to the turbine and damage to surrounding areas.
Thus the problem facing industrial gas turbines is clear. Designed for creep resistance, using creep-resistant materials suitable for machines which base load with minimal start/stop cycles, the introduction of severe fatigue cycles into the mix poses a significant challenge to their long-term health.
It is imperative that operators recognize that gas turbine multi-starts may be resulting in more damage than they are aware of. This knowledge may then need to be taken into account when considering the electricity price at which generation is economical.
With the challenges already facing the energy market, the issue of asset damage due to multi-starts needs to be considered by both operators and governments to ensure that generators are not disadvantaged. With the need for flexibility in energy provision, gas turbine generation must remain viable.
Gary Lock is Senior Business Manager at Frazer-Nash Consultancy. www.fnc.co.uk