LIFE MANAGEMENT FOR TURBINE ROTORS

Published on: 

SHOULD USERS ACCEPT OEM ASSESSMENTS OF ROTOR LIFESPAN?

Many utilities have combustion turbines (CT) that are approaching their given original equipment manufacturer (OEM) life limit. There has been a chorus of proclamations from OEMs on what their respective life limits should be, and some modifications to OEM claims over the past year.

Initially, rotors reaching their expected life limits were supposed to be retired, but response from industry has been overwhelmingly negative. OEMs are now saying that after reaching their life limit, combustion turbines need full evaluations, including destacking (in the case of GE, Siemens, and MHI, which have stacked rotors).

GE, Siemens, Alstom and MHI have all limited their fired hours from just under 150,000 hours to less than 100,000 hours, depending on the specific OEM. Similarly, OEMs have limited starts to a range between 1,000 and 5,000.

The reason for OEM concern is not so heavily weighted on experienced-based failure data, but more on theoretical extrapolation of models. They have looked at the rotor and various forms of operation (i.e., cycled, load-following, base loaded) then speculated as to possible damage mechanisms that could occur.

Since they currently do not have adequate empirical data to provide examples of actual rotor failures or “near misses,” these conclusions are open to question. Keep in mind that we are not including failures of consumables, such as blades, vanes, seals, carriers and diaphragms, but the rotor proper, which in most cases is a low-alloy steel. Nevertheless, OEMs are typically conservative in operation, repair, and refurbishment, which have a safety, as well as a financial component.

All four OEMS have offered rotor life extensions based on the result of thorough inspection processes. The nature, scope and cost of the OEM inspection for rotor life extension varies widely, and depends on factors, such as construction, original forging data, expected degradation modes and operational modes.

Most OEMs require a full dismantle and destacking of the rotor into its individual components (discs, spacers, through bolts), except in the case of Alstom where most of its installed base in the U.S. has a welded rotor construction (this provides some benefit from a rotor inspection standpoint).

OEMs agree that rotor life can be extended if the turbine rotor is sent to the OEM rotor facility and the discs destacked and inspected along with other component inspections. One OEM currently has F-class turbine disc cracking issues, but the mechanism is slow growing, well characterized, can be found via eddy current inspection during normal major overhaul inspections, and does not require a destack.

Unit operating data

Based on combustion turbine experience, we have yet to be involved with, or hear about, a rotor or casing failure of a framesize CT caused by exceeding the life exhaustion point. We have heard of infant mortality failure for mechanisms that had not been anticipated. Much more often, we have heard of multiple cases where the utility had gone well past the life exhaustion date with no problems.

There are known problems unique to each OEM and frame type (fleet issues) with specific components issues and failure mechanisms. These mechanisms have caused major failures in the past. However, these mechanisms and components have been identified and are being covered by inspections per technical information letters (TILs) and are generally not life-limiting to the rotor.

For a given mobilization and audit, past overhaul inspection reports and operating data are reviewed. Plant personnel are interviewed with regard to historical operation (operating hours, starts, trips, load rejections, and so on), maintenance practices, upgrades, and past problems with the unit.

It is usually best to ask control room operators to provide trending of various parameters (vibration, disc cavity temperatures, and so on). Tabulated data may be presented as in Tables 1 and 2.

It is also advisable to review past overhaul inspection reports and OEM technical bulletins, upgrade reports, and incident reports that had been relevant to the rotor assembly. Plant personnel should be interviewed with regard to historical operation, maintenance practices, upgrades and any problems with the units.

It is typically found that most of the past inspections and overhauls have indicated that all the rotor assemblies did not show any signs of unexpected degradation. There were typical hot gas path component replacements along with compressor blade replacements and some modifications.

However, there generally have been no unusual rotor cracks, deformation, bowing and runout, or degradation found during the multiple inspections over the years. Probably the most common rotor issues have been wear on the disc steeples for units with high time on turning gear, failed disc through bolts and diaphragm contact with rotors and blades.

When possible, a meeting should be made with the OEM to determine its methodology for rotor life extension. Often the OEM will go over its methodology for rotor lifetime degradation, rotor monitoring and inspection, and finally rotor residual life capability. They will discuss degradation mechanisms for critical components.

Table 3 shows the possible degradation mechanisms and corresponding rotor assembly components suggested by various OEMs.

Advertisement

As can be seen from Table 3 there is a spectrum of damage mechanisms and potentially affected rotor sub-components.

Risk is the probability of failure multiplied by the consequence of failure. The probability is based on many factors, but will be compared for the following scenarios:

1. Operate as is with no rotor inspections (10,000 to 50,000 hours of extension)

2. Perform nondestructive tests in the field to inspect the rotor for the various suspected failure mechanisms (10,000 to 50,000 hours of extension)

3. Perform a rotor swap

Two other scenarios can be investigated: shipping the rotor to either a third-party service provider or the OEM for destack of the rotor discs and inspection. Because of the length of time required for these scenarios and the subsequent loss of revenue, these destack scenarios are not typically advised.

The risk assessment followed the probabilities of failure for the various components and damage mechanisms for each of the three scenarios above. Actual probabilities are not available even to the OEM as few turbines have exceeded their life to date.

Causes and drivers

As such, the Bayesian method for probability has been used. Our understanding of failure mechanisms as well as their causes and drivers allows us to logically address the current state of knowledge.

Having validated the Bayesian method on steam turbines with 4,000 turbine years of accumulated operation over multiple units, the same methodology has been applied to gas turbines for probabilities of specific failure mechanisms in the rotor.

After assigning a probability of failure for each damage mechanism for each of the three scenarios, the next step was to assign a consequence. In the case of the utility, the consequence of failure was financially addressed.

Per data provided by the utility, the impact to the bottom line had been often on the order of $100,000/day when a unit was down. Thus, if the turbine had experienced a rotor assembly failure the costs to the company’s bottom line would be $100,000/day the unit is down.

Since the consequence will not change with any of the three scenarios, the risk can be calculated for a daily basis comparison; and a total risk would be determined by the time required to fix the damage and get the plant up and running again.

The cost of turbine repair or replacement after a failure is not used in the model, as insurance will cover most of this less the deductable; and the deductibles are often the same for each site within a large utility portfolio.

With the probabilities and consequences gathered, the risk for each scenario can be determined by adding the risk for all the rotor components and their potential damage mechanisms. Note that the risk will not include Hot Gas Path (HGP) consumables, such as, blades, vanes, diaphragms, and so on, but just the rotor assembly components.

Chart 1 shows a risk summary of a typical utility. In this scenario, the cost to perform a field inspection is in the tens of thousands of dollars. A rotor swap further reduces the risk, but the estimated cost is in the millions.

A simplistic return on investment (ROI) can be calculated using total risk reduction per dollar invested to reduce the risk. For example, a typical failure of a rotor disc would likely cause the turbine to be down for up to six months. When the numbers are evaluated, it is often found that the best risk reduction per dollar spent is field inspection. The added risk reduction of a rotor swap is usually not a good payback.

One concern that several OEMs have about their rotors is related to original manufacturing and inspection of the disc bores. Assuming the original boresonic data is available and does not show critical flaws based on modern fracture mechanics modeling, subsequent boresonic analysis may not be required.

Based on data gathered, it is recommended that utilities consider performing field inspections of rotors nearing OEM dictated life exhaustion during the next major outage. Unless something significant is found, most turbine rotors could go at least one more major overhaul cycle beyond the typical life expectancy before additional inspection is required.

Authors

John P. Molloy, P.E. and G. Mark Tanner, P.E are senior engineers at M&M Engineering, a company which has performed rotor life extension management consultation, unit audits and evaluations on rotors of all types, as well as rotor inspections. This includes over 185 combustion turbine failure investigations. For more information, visit: www.mmengineering.com