Maximizing the productive lifetimes of gas-turbine assets
By Ron Munson, PE, Ron Munson Associates LLC
During the 1980s and 1990s, many electric utilities were challenged to extend the lives of their power generation assets. Some units were rapidly approaching, or even beyond, their design lifetimes of 30 to 40 years and economic necessity demanded that these assets continue to operate safely and with a high degree of reliability. EPRI and others adopted strategies for life extension based, in large part, on the results of nondestructive examination (NDE) and metallurgical evaluation of critical components.
However, the rapid acceptance of gas-turbine-based generation to meet increasingly strict environmental laws and other regulatory edicts, eventually de-emphasized life-extension initiatives for conventional fossil-fired steam/electric plants.
Today, GT owner/operators are the ones facing life-extension initiatives. Inexpensive gas has displaced coal generation on an economic basis in many areas of the country and aggressive environmental rules are forcing the shutdown of other coal-fired stations. Gas turbines are working harder and longer to meet electric demand.
At the same time, OEMs are telling GT users their prime movers have critical parts with finite lifetimes and that replacement of these parts is necessary to assure reliability and safety moving forward. It remains unclear at this time if the sudden emphasis on the lifetimes of parts is technically or commercially driven.
The technical discussion that follows reviews life-limiting deterioration mechanisms for some critical parts and offers owner/operators guidance in life-extension decision-making.
GT basics. Life-extension specifics depend in part on the type of gas turbine: frame or aeroderivative. The former are the most robust and shoulder the burden in terms of kilowatt-hours produced; the latter, lighter and typically smaller, are valued for their fast-start/ramp capability to meet peak demand and respond quickly to changing electrical requirements.
Life-extension issues are not as relevant for aeroderivatives as they are for frames because their modular construction permits routine replacement of critical engine sections over the life of the unit. Thus, the focus here is on frame engines.
When a generating company invests in a gas turbine, life expectancy usually is not high on the purchaser’s list of priorities. Delivery schedule, price, emissions, and heat rate typically are viewed as more important. For the most part, owners have assumed that GTs will be fit to operate well beyond the useful economic life of the facility in which they are installed.
Until recently, gas-turbine designs and materials capabilities had advanced the replacement of older frames on the basis of economics. Ever decreasing heat rates allowed relatively quick recovery of capital investment and provided the justification for ordering new rather than upgrading old. But the inability to recoup capital investments quickly today and thermodynamic limits on efficiency improvements have shifted emphasis once again to life extension.
GT users understand that many expensive hot-gas-path parts, such as combustion-system components and turbine airfoils, must be replaced or repaired regularly. With this a “given,” life-assessment issues focus on more durable components, such as casings and rotors, which are considered by most owners to have very long, if not infinite, lifetimes. They are costly to replace and lead times are long.
Casings for industrial gas turbines usually are robust, made of cast carbon or low-alloy steels. These components are quite large and are unique to a certain model, and often to a particular serial number, of a machine. Casings are characterized by many welded-on pieces—such as nozzles and flanges—which typically are added during manufacture with the attachment welds being appropriately heat-treated.
Replacement casings for legacy machines may no longer exist, the original fabricator having gone out of business and casting patterns destroyed. Repair of some old casings may require an expensive weld, heat treatment, and remachining—including line boring. The total cost could approach that for a new casing were one available. Replacement with a used casing may be an alternative, but its history is likely unknown and the pedigree questionable.
Guidance on casing repairs is offered in a companion article, “Four ‘knows’ help identify a viable approach for dealing with casing cracks,” p XX.
Rotors for most gas turbines are forged low-alloy steels, heat-treated to obtain a balance between strength and toughness. They can be constructed from a single forged ingot, end to end. But because of forging size constraints, the majority of GTs are built from multiple forgings that are mechanically bolted or fastened together. One OEM welds its forgings together rather than using a mechanical joint. A material exception is GE Energy, which fabricates the wheels for its F-frame turbine from a nickel-based superalloy.
There are many damage mechanisms conducive to the retirement or extensive repair of a GT casing or rotor. Listed below, they can be either short- or long-term in development. Life-extension activities are influenced primarily by the latter, especially “wear and tear” issues.
Short-term or random mechanisms include the following:
- Foreign-object damage (FOD).
- Domestic-object damage (DOD).
- Distortion or buckling from cool-vapor introduction.
- High-cycle fatigue (HCF).
- Environmentally induced cracking (stress corrosion cracking, SCC, or corrosion fatigue, CF).
Short-term mechanisms usually are not a major concern in evaluating the life exhaustion of either a casing or rotor. Periodic borescope examinations, scheduled inspections, and installed instrumentation usually will point to damage that has occurred or will be occurring.
These mechanisms typically are precipitated either by random or unpredictable events—such as DOD, FOD, trips, or cool-vapor introduction. HCF typically is related to design—improperly tuned blades, for example—and normally occurs early in the life of the component. Corrosion fatigue and SCC are rare, but usually predictable, if one understands the GT environment.
Long-term or progressive mechanisms are these:
- Thermal mechanical fatigue (TMF).
- Low-cycle fatigue (LCF).
- Hot corrosion.
- Wear/fretting loss of dimensional tolerances.
- Microstructural ageing and embrittlement.
Long-term mechanisms are the ones most responsible for casing and rotor life-exhaustion issues. Both TMF and LCF depend directly on cycling; magnitude of the damage is influenced more by ramp rate than by the number of cycles.
Creep is very temperature-dependent. In fact, there are temperatures below which a material is not going to suffer creep damage. Keep in mind that, while creep is related directly to firing temperature, casings and rotor, if properly cooled, may be operating below the creep initiation temperature for their alloys of construction. Fortunately, creep damage occurs with deformation, which is conducive to rubbing of components and does not lead to catastrophic failures.
Corrosion, hot corrosion, fouling, erosion, and loss of dimensional tolerances are slow to accumulate and quite progressive in nature. They are detectable by visual examination if their locations are accessible, and become problematic when they cause cooling passive disruption, or the corrosion products that form result in loss of clearances. Clearance loss can lead to rubbing or loosening of parts and contribute to fatigue or wear.
One particular issue is that corrosion products (rust) from the steel compressor case enter the hot section and cause a change in blade damping characteristics, thereby contributing to fatigue failures. Another example is that blades on turning gear will fret against adjacent blades and their attachment points, and change blade fit-up, resulting in HCF from poorly damped, harmonically induced vibration.
An important issue that must be faced in life extension is the long-term ageing of construction materials. Most steels, and even superalloys, are meta-stable in the condition they are used in gas turbines. (This means they want to be in a lower thermodynamic energy state than where they currently exist.) At room temperature, the time it takes to get to the thermodynamically stable state is very long, if not infinite.
However, the GT’s high-temperature/high-stress environment wants to push the meta-stable structure to stable structure in measurable times. This is analogous to in-service heat treatment. Such an alteration can have a dramatic impact on component performance. Examples are gas-phase embrittlement (hold-time cracking) in GE hot-section discs, graphitization of carbon steel in compressor casings, and temper embrittlement of low-alloy-steel rotors.
Drivers for damage. Each damage mechanism has a primary driver. The mechanism could be caused by cycling (such as TMF), be related to fired hours (creep and erosion), or attributed to random events. Last could be FOD from the inadvertent introduction of objects during outages. Machine operation, configuration, and maintenance also have a huge impact on susceptibility. Some factors to consider when doing life-extension studies are the following:
- Time on turning gear.
- Fuel quality.
- Inlet-air conditioning and filtration.
- Equipment environmental protection during outages.
- Procedure for compressor water washing.
- Environmental control equipment—such as low-NOx combustors.
- Firing temperature (design versus actual).
- Quality and knowledge of overhaul contractor.
Risk-based approach. There are many avenues you can take to assess GT condition. One is the “Chicken Little” approach: You accept the OEM’s recommendations without question and conduct the testing or replacements recommended at the intervals mandated. It is safe, but costly.
An alternative is to do nothing. While this approach has a low initial cost, it is likely to be very unpopular with your insurance carrier! Ultimately it could lead to severe financial penalties in property damage and lost power generation at a very inconvenient time.
A risk-based approach is more prudent: You consider each possible damage mechanism identified above and make an engineering judgment as to the probability of that mechanism being present in your engine. Judgment must consider operational factors, past inspections, and failures.
Risk has two parts: probability (likelihood) and consequence (severity). The consequence does not change with time alone. It depends considerably on contractual issues—such as power purchase agreements, which have significant variations in their terms. A part of risk is the consequence/severity of a failure that may occur. If the consequence part of the risk is low, it may be worth delaying assessment or replacement until conditions are financially favorable.
The probability of a failure does change over time. Example: The probability of failures resulting from certain damage mechanisms—such as wear, creep, microstructural ageing, LCF, and fouling—increase with time and/or more cycles.
However, there are relatively few data to use as a basis for quantifying the increase in probability over time. Many factors (called risk-modification factors) can greatly influence probability, which does not increase linearly with time or cycle accumulation. A prudent risk evaluation requires knowing what these factors are and how to apply them.
In sum, the key to a successful and cost-effective risk-based assessment of GT condition involves knowing where to look, what to look for, and what method to use to find the damage. In order to make these choices, one has to understand the history and features of the particular machine in question. Do not stereotype a machine based upon age, number of cycles, or model number. CCJ
Ron Munson, PE, is corporate engineer at Ron Munson Associates LLC, a metallurgical engineering and materials consulting firm headquartered in Round Rock, Tex. He has decades of experience in evaluating the condition of gas and steam turbines, reciprocating engines, high-energy piping systems, boilers, and related components and in recommending repair/replacement options where damage is in evidence.