Gas turbine systems are widely used in onsite power installations for industrial plants, commercial office buildings, hospitals, shopping centers, high-rise apartments, and other distributed-energy applications.

The onsite gas turbines used can range in size from 30 kilowatts on up to 20 megawatts. Gas turbines in the 30- to 500-kilowatt range are generally referred to as microturbines; those in the 500-kilowatt to 20-megawatt range are known as industrial gas turbines.

This article, the first of two, deals with onsite gas turbine operations and maintenance, and focuses on the issues encountered with these industrial turbines. The second article will deal with the operations and maintenance (O&M) of microturbine systems.

To bring you the most current and authoritative information on gas turbine operations and maintenance, DISTRIBUTED ENERGY interviewed a leading expert on gas turbines, Nick Pozzi, manager of customer service for the Gas Turbine Division of Kawasaki Gas Turbines—Americas in Grand Rapids, MI.

In the following interview, we systematically walk through all the major components of a typical onsite gas turbine system in the 1.5-megawatt class (although the insights here are relevant to all sizes of gas turbines), including these:

• Air intake
• The compressor
• The combustor
• The turbine
• Air emissions
• The gearbox
• The electric generator
• Preventative and predictive maintenance
• Who should perform the maintenance (in-house versus outside experts)

DE: Are there any significant differences between a steam turbine and a gas turbine?
Pozzi: Yes indeed! It is true that there are strong similarities between a steam turbine and a gas turbine. The steam turbine uses steam to create motion in the turbine and the gas turbine uses a hot gas. Both use the energy of these fluids to turn a generator, which makes electricity. A gas turbine operates at a much higher temperature than a steam turbine (typically, at 2,000 degrees Fahrenheit versus 1,200 degrees Fahrenheit for a steam turbine).

Accordingly, a gas turbine must be designed to withstand these much higher temperatures. The components of the gas turbine (rotor, blades, vanes, etc.) are fabricated from much more expensive metal alloys—e.g., inconel. Further, the exhaust of a gas turbine can be used to make steam. And, when channeled into a steam turbine-generator, that steam can be used to make more electricity; or it could be used to operate a chiller to make cold water or cold air for air-conditioning a facility.

Because of the need to use these special metal alloys, a gas turbine is considerably more expensive than a steam turbine. A 1.5-megawatt gas turbine can cost in the neighborhood of $1.2 million—versus $750,000 for a comparable steam turbine.

DE: I noticed that your gas turbine system has a filtering unit on the air intake. Why is it necessary to filter air entering the turbine system?
Pozzi: The air has particles in it that are potentially damaging to the gas turbine system. Without an effective air-filtering system, these particles over time would reduce the turbine efficiency. Poor air filtration could cause what we call FOD—foreign object damage. Someone could also inadvertently drop a nut or bolt into the air intake duct, which could then damage compressor or turbine blades. As a specific example, we know of a case where snow found its way into the air intake and subsequently turned to ice. Some ice chunks were then drawn into the engine, caused vibrations, and, two weeks later, a turbine blade loosened up, requiring that the entire gas turbine be shut down for repair.

DE: Specifically, what particles are in the air that could potentially damage the compressor-turbine system?
Pozzi: There are many particles in ambient air, but most important are airborne salt (sodium chloride) particles, especially for sites along the coast. Such salt particles can be very damaging to a turbine. Salt is, of course, very corrosive to metals. But beyond that, it leads to a buildup of dirt on compressor and turbine blades. Once salt particles attach themselves to turbine blades, they apparently attract other particles—dirt. The net result can be a loss in the turbine's power output of up to 15%.

DE: How do you actually filter the air?
Pozzi:We draw intake air from the ambient through a special filter called a HEPA filter, not unlike the filters used in vacuum cleaners. It is most important to use this generic type of filter, as it removes 99.87% of the particles in the air. In sum, the HEPA filter keeps the turbine cleaner, resulting in higher power output.

DE: What if the operator of a turbine system has been negligent and has not done a very good job of filtering intake air? What to do then?
Pozzi: The solution is to water wash the engine periodically. Some operators do this once a month. This water wash is done with the turbine running. As an alternative, the turbine can be washed by the so-called cold-wash method, done at low revolutions per minute (rpm) when the turbine is at purge speed (30% of speed). A purge is required on all turbines to remove any potentially explosive vapors.

The point is that, with good intake-air filtration using the HEPA filters, there is no need to wash the compressor-turbine system as often as before. The turbine system stays cleaner for a longer time, thereby avoiding a degradation of performance.

DE: What do you mean by "purging" the turbine system? Why is it done and how?
Pozzi: Consider that the turbine system has been off. And now you wish to turn it back on again. There could be explosive unburned fuel vapors still lurking inside the turbine system—in the compressor, combustor, turbine, waste-heat recovery system, etc.

To guard against the danger of explosion, it is standard practice to first purge the turbine system of these potentially explosive vapors. But one needs to purge safely. And this means one must proceed without activating the igniters (spark plugs) in the combustor (i.e., the turbine's combustion chamber), for the sparks could trigger an explosion. The control system does this automatically during a normal startup.

The starter motor (hydraulic or VFD) is used to rotate the turbine at 6,600 rpm, turning the turbine at about 30% of its usual rpm, enough to make possible the drawing in of fresh ambient air, which will then quickly displace (or purge) any potentially explosive gases or liquids from the system. That done, it is then safe to fire up the igniters, thereby setting the turbine into operation.

Incidentally, in carrying out this operation, it is important not only to purge the turbine, but also the waste heat recovery boiler. One does not want dangerous unburned fuel vapors lingering in the waste-heat recovery boiler. The Kawasaki 5-megawatt turbine system has a diverter downstream from the point where hot exhaust gases emerge from the turbine. This allows the hot turbine gases to either be sent out the stack or through the waste-heat recovery boiler.

By law, the operator must first purge the turbine with a minimum of six air changes before firing up the turbine.

DE: So filtering intake air then is very important? And it will greatly extend the life of a gas turbine system?
Pozzi: Yes! Changing the HEPA air filters periodically is very important. Indeed, it is one of the most important things a turbine operator can do to maintain his gas turbine system.

Nonetheless, the overall lifetime of a gas turbine system is greatly affected by the physical environment it is placed in. Is it located adjacent to a paper mill spewing sulfur compounds into the ambient? Such could lead to the formation of coatings on turbine blade surfaces or to premature bearing failures.

A gas turbine located at one site may last five years before it needs an engine change or a bearing replacement; and the same turbine at a different site may last only four years. It all depends on turbine usage and on location. Good air-filtration systems prevent damage from occurring, thereby increasing the lifetime of the equipment.

Where there are special environments—e.g., an offshore oil-drilling platform—it makes sense to increase inspection frequency. There, a gas turbine is used to drive pumps to move oil. In the air there are many salts and other contaminants. It is common practice to flare off hydrogen sulfide gas from the wells, some of which is drawn into the turbine system inlet, possibly causing premature damage to turbine blades.

Turbine field service

DE: What sort of maintenance is required of the gas turbine itself? How often is it done and by whom?
Pozzi: We recommend a maintenance service plan to our customers who call for our service technicians to perform three quarterly inspections and an annual inspection. In the second year, we recommend doing another three quarterly inspections and an annual inspection. At that 16,000-hour point, when this second annual inspection is being done, the technician will also do what's known as a hot-section inspection.

DE: Please expand on these quarterly inspections. What is involved in these quarterly inspections? Who does what?
Pozzi: During the quarterly inspection, the maintenance technician collects turbine system performance data—on such variables as vibration, pressures, temperatures, and outputs. The original equipment manufacturer (OEM) then analyzes the data and issues a serviceability report. This report compares the newly collected data to baseline data collected during startup of the new turbine. If we notice any deviations from the baseline, we flag the problem early and investigate it ASAP—meaning during the next available shutdown.

Such aggressive predictive maintenance helps eliminate downtime and possible expensive repair. This is one of the things we do to achieve over 98% reliability and availability (R&A). Indeed, some of the gas turbine systems that we have installed have achieved 98% R&A for the past 13 years. Preventive-maintenance and predictive-maintenance methods will vary some from one OEM to the next and the prospective turbine system buyer should inquire about an OEM's R&A track record.

What else is done during a quarterly inspection? The service technician will check lube oil and air filters for dirt. He will check magnetic pickups for metal deposits—for significant accumulated metal pieces could be a warning of early bearing failure. Even though most magnetic pickups are alarmed, we don't wait for the alarm to sound or the automatic shutdown to occur. Instead, we inspect each quarter for early warning signs.

DE: What about the annual gas turbine system inspection? What happens there and who does it?
Pozzi: The most important part of the annual inspection (as called for in the service agreement between OEM and customer) is the use of a borescope to examine the internals of the turbine. This borescope is very similar to the fiber-optic cable that a physician uses to examine a person's colon. The OEM technician snakes the borescope's fiber-optic cable inside the gas turbine so that he can inspect its internal components and take photos.

The technician is looking for cracks in the lining of the combustion chamber, erosion on turbine blade tips, loss of protective coating on blades, and signs of overheating—e.g., a blade tip may be melted off.

Besides doing the borescope inspection, the OEM maintenance technician during this annual inspection also checks every alarm and shutdown device to see if it is working properly. And he also checks all major fluid levels and filters.

DE: How often does he find something wrong? What happens if he finds a damaged blade?
Pozzi: If something is not working up to specifications, we replace it. Such is part of our typical maintenance service agreement. Gas turbine systems seem to have two lifecycle periods they go through: the first six months of operation, and after year 10. I am referring her mainly to the ancillary equipment provided with the turbine package—things such as transmitters, switches, automatic valves, human machine interfaces, and data loggers (computers that store historical data).

Remember that every minute that we are down counts against our availability. Accordingly we do everything in our power to make sure things will work correctly until the next inspection. If we think they will not, we change any suspect components.

Also bear in mind that even our scheduled downtime for the quarterly and the annual inspections, the hot-section inspection, and engine and gearbox change-out time—all this counts against our availability. Accordingly, we can't afford any unscheduled downtime. Our customers depend upon our equipment to run successfully all the time. So we do everything possible to make that happen.

DE: The compressor is quite clearly an important subsystem of an industrial gas turbine system. What sort of maintenance does that require? Who should do what when?
Pozzi: An important component it indeed is. Yet the compressor—which of course turns on the same shaft as the turbine—is very robust and only rarely needs any maintenance attention. Don't forget that the compressor is upstream from the combustor and, as such, is not subjected to hot combustion gases, but only to the ambient air drawn in to be compressed before it flows into the turbine's combustor. Essentially, the only maintenance needed for the compressor is that done once every 32,000 hours (about 4 years), as part of a total shop overhaul and rebuilding of the entire turbine.

DE: You mentioned that every 16,000 hours you do a hot-section inspection. Please expand. What is a hot-section inspection? Who does it, when, how, and how long does it take?
Pozzi: A hot-section inspection is an examination of those parts of a turbine system that are exposed to the hot gases created when compressed intake air is mixed with natural gas or other fuel inside the combustor and ignited by the igniters. In a word, the hot sections are mainly the combustor (i.e., the combustion chamber) and the turbine section—and any other components exposed directly to flame or to hot combustion gases.

This hot-section inspection is done every 16,000 hours (about every two years). It is crucial that the inspection be performed by an experienced and knowledgeable turbine technician, who does the inspection on-site.

The inspection involves opening up the combustor and turbine sections; carefully examining walls, linings, turbine blades, and vanes, etc.; replacing any worn or damaged components; then reassembling and starting up the system. On a 1.5-megawatt turbine system, this will usually take an experienced turbine technician about three days.

DE: There is a difference, then, between a borescope inspection of the turbine and a hot-section inspection?
Pozzi: Oh yes! Most definitely! A borescope inspection is done every year—that is, every 8,000 hours of operation—as part of the annual inspection. It is a way of examining the internals of the gas turbine—by snaking a fiber-optic cable inside the turbine casing—without going to all the time and trouble of opening up the turbine and looking at its internals directly.

By contrast, we do a hot-section inspection only once every 16,000 hours (i.e., once every two years). Such an inspection is much more thorough than a borescope inspection, for it involves opening up the turbine and examining it directly. Both the service technician and the turbine owner are able to see the internals of the turbine directly and what its actual condition is.

During a hot-section inspection, the turbine technician replaces damaged turbine blades, vanes, and any other components showing wear. Upon completion of the inspection and of any needed repairs, we guarantee to the customer that the gas turbine system will be good for another 16,000 hours of operation. We give him a serviceability report indicating bearing wear and clearances, and noting any components that were replaced. Our guarantee gives the customer a comfortable feeling about the equipment and its performance.

DE: Is that it then for the major maintenance on the turbine itself? A borescope inspection every year and a hot-section inspection every two years?
Pozzi: No! After all, no gas turbine will last forever. After the turbine has been operated for a total of 32,000 hours (i.e., four years), we will pull out the existing turbine and completely replace it with a factory-rebuilt turbine. The old turbine is then carted back to the OEM's factory, where it is used as the core for a rebuilt turbine. It will be completely overhauled and rebuilt. At this time, OEM technicians will replace virtually everything inside the turbine except for the case.

DE: During these major maintenance intervals—either the 16,000-hour or 32,000-hour service—are there any upgrades to the gas turbine system that the customer can opt to have—e.g., having turbine blades coated with protective coatings, etc.?
Pozzi: Advances in technology change the possible upgrades. Sometimes a customer will choose to have the latest technology installed in the gas turbine system during an upgrade.

Those customers with some of our older systems (over 13 years old) often consider these upgrades: a new control system with historical trending and/or management reports; remote monitoring; and predictive-maintenance software.

As concern possible upgrades to the gas turbine itself, the latest technology uses ceramic materials for both turbine blades and vanes. Accordingly, this ceramic option is something a gas turbine system owner needs to consider for the 32,000-hour overhaul. Would it make sense for him to spend the extra money to install ceramic blades and scrolls (these distribute the hot gases to the blades)?

As for applying protective coatings to turbine blades, conventional metal-alloy turbine blades already have a protective coating on them that is very durable. Accordingly, there is no need to repaint those blades during a 16,000-hour turbine overhaul. Generally, it is not a good idea to change turbine blades in the field; for there is a need to properly balance the turbine rotor after installing new blades, an operation that needs to be done in the OEM's turbine shop, where there is the proper balancing equipment.

To overcome this obstacle, some manufacturers stamp each turbine blade with its exact weight. In this way, a given blade can be replaced with a new blade that is identical in weight, thereby eliminating the need to balance the turbine rotor after old blades have been removed and new ones added. A new blade is merely slid into a tapered slot on the rotor and secured with a bolt.

DE: Is the replacement of the turbine at 32,000 hours an absolute must? Or is there some way that this expensive task can be postponed?
Pozzi: No, it is not always an absolute must to replace all 1.5-megawatt gas turbines at 32,000 hours. But before we could recommend going beyond that, we would have to rigorously inspect the turbine. In some cases, we have been able to postpone replacement to the 40,000-hour mark. When extending the replacement time of a turbine in this way, it is also prudent to increase the frequency of inspections, to record predictive-maintenance data more frequently, and to be extra vigilant about watching for deviations of the data from the norm.

We have many of these 1.5-megawatt gas turbine systems operating in Mexico, mostly in industrial settings. Many of these units have 40,000 hours on them and yet have received no maintenance since being installed.

On these units, the cores of the turbines will be spent. These unmaintained turbines will not be suitable for use as cores to tear down and rebuild. Such is the consequence of doing no maintenance—of running a unit for 40,000 hours without changing air filters. Such can cause severe abrasion of the turbine shells. Accordingly, these turbine owners will not be able to receive a credit for their turbines when they are finished with them, for they will not be rebuildable.

DE: Are there certain ways to operate a gas turbine that will help extend its useful life? Perhaps, for instance, by not operating it at full load?
Pozzi: Concerning turbine operating strategies, a cardinal rule is to always operate the turbine at full load. Why? Because the turbine operates much less efficiently when operated at partial load.

Said another way, an onsite gas turbine system should be used wherever possible to meet a facility's baseload power requirements. And any power needs above this baseload should be met by purchasing power from the electric-utility grid.

DE: What about the combustor? Could you please explain: How important is it? Its major functions? And its most common maintenance problems and what to do about them?
Pozzi: The combustor is the heart of the gas turbine system. Our control system keeps the temperature of the combustor stages within normal operating limits, a measure that extends the life of the hot-section inspection components. The combustor is essentially a trouble-free component.

DE: What about gas turbine emissions to the atmosphere? Are they much of a problem?
Pozzi: Kawasaki guarantees that emissions from the gas turbine will meet air-quality emissions standards. With our 1.5-megawatt gas turbine system, we guarantee an emissions of 17% oxygen, less than 10 parts per million (ppm) of carbon monoxide, and NOx of less than 2.5 ppm. We use a catalyst to help achieve these low NOx levels.

Sometimes, a turbine does not burn its fuel as cleanly as it should. On large turbines (15 megawatts or above), there are usually ways to fine-tune the combustion. On the smaller 1.5-megawatt units, one can still tune the combustion—but it is a much simpler process than on the larger units (it takes 20 minutes versus three days).

Most gas turbine sites need to secure an emissions permit from the appropriate state environmental agency. Usually, there are heavy fines for non-compliance with emissions limits. Many states require that emissions samples be taken every quarter; and some states, like California, want the gas turbine owner to monitor 24 hours a day for carbon monoxide and NOx emissions.

But continuous monitoring equipment is very expensive and has a high O&M cost. Fortunately, for gas turbine operators using "proven" low-emissions technology, California is now providing an exemption to the costly continuous emissions-monitoring requirement; such operators need only take stack emissions samples quarterly and send them to the state. For example, Kawasaki's low-NOx-emissions Xonon technology meets the "proven" technology criteria, so can be used without costly continuous emissions monitoring. Other states with tough continuous monitoring requirements are beginning to follow California's lead and to provide exemptions for proven technology.

DE: What is actually adjusted in this tuning process? Are there ways to regulate both the air flow rate and the natural gas flow rate into the combustor? Is it possible to adjust the power output of the turbine? Or does it always have to be run at full load?
Pozzi: Combustion tuning in our equipment is very easy. We only have two adjustments and these involve fuel flow and bypass compressor discharge pressure airflow around the combustor. It takes a qualified field-service technician about 20 minutes to set this up.

On larger turbine systems, combustion tuning can sometimes take days. The system performance is mapped over the full operating range. To do this, the fuel input is increased incrementally, and performance data such as emissions and fuel efficiency recorded. Why does the whole tuning process take days? Because each time the fuel input is ratcheted up a notch, it takes 20 minutes for the system to stabilize at the new level, so that meaningful emissions data can be recorded.

DE: Do you monitor for sulfur dioxide emissions?
Pozzi: This is not usually required, for one cannot readily control SOx emissions. Such emissions depend entirely on the amount of sulfur in the incoming fuel. The most common fuel used in gas turbines is natural gas. And usually the amount of sulfur in natural gas is quite low and controlled by the gas utility.

By contrast, the oxides of nitrogen (NOx) that are in gas turbine emissions are created by the combustion process itself going on inside the turbine's combustor. How much NOx is created depends upon the design of the particular combustor, its operating temperature (the higher the operating temperature, the more NOx created), and whether or not catalysts are used to facilitate the combustion process.

DE: Could you please expand further on catalytic combustion? What it is? How recent? How often used? What the pros and cons are? What maintenance is entailed?
Pozzi: The combustor is the heart of the gas turbine system. And the use of a catalyst is the key to the turbine's low emissions. Kawasaki is not the only gas turbine manufacturer that uses catalytic combustion; but in the 1.5-megawatt size category, it is currently the only one.

The catalyst facilitates combustion of the incoming fuel-air mixture. By operating the combustor's preburner below 900 degrees Fahrenheit, we eliminate formation of NOx. The remainder of the fuel is added and ignited in the combustor's burnout zone, the reaction being facilitated by a catalyst. In the main part of the combustor itself, there is no flame—just hot gases and unburned fuel. The mixture actually burns once it flows into the catalyzed burnout zone. Here, there is some flame, but little NOx is formed, yielding a very clean emissions.

This Kawasaki Xonon technology has been tested now for over five years and has been commercially available for over two. What are its pros and cons? On the pro side, the emissions are very low—10 ppm carbon monoxide and less than 2.5 ppm NOx. On the con side is the need to change the catalysis module every year (8,400 hours). The maintenance technician changes it as part of the turbine system's annual inspection, a procedure that takes a day. Overall, the catalysis module adds a little bit to gas turbine system operating costs. But the reduced emissions can make that added expense money well spent.

DE: Do you expect to see catalytic combustion spread to larger gas turbine systems, those larger than 1.5 megawatts? Is such feasible?
Pozzi: Yes. I am quite confident that catalytic combustion will advance into larger gas turbines—first the 7-megawatt units and later the 20-, 50-, and 100-megawatt units. And this will happen first in California. Incidentally, Kawasaki does not sell its Xonon catalytic-combustion technology to competing manufacturers, but there are other competing technologies out there.

DE: OK. You have told us much about the gas turbine, the air going in, the emissions coming out, and the periodic maintenance required. What, then, is the lifetime of a gas turbine system? Is it basically the 32,000 hours (four years), after which the turbine itself must be completely replaced?
Pozzi: No! It is indeed true that the turbine itself has to be completely replaced after 32,000 hours—although sometimes that time can be stretched to 40,000 hours. But there are other major components in a gas turbine system that will last much longer than the turbine itself without replacement. The gearbox, for instance, will typically last for about 50,000 hours (six years) before it needs replacement. And an electric generator typically lasts for 22 to 30 years. So, depending upon how you view a turbine system, its lifetime could be considered to be in the 20- to 30-year time frame.

DE: What is the function of the gearbox? And what sort of maintenance does it require?
Pozzi: Mechanical power in the form of a rotating shaft flows from the gas turbine to the gearbox's input shaft, then from the gearbox's output shaft to the electric generator.

The electric generator in this 1.5-megawatt turbine system needs to rotate at 1,800 rpm. Yet, the 1.5-megawatt turbine is rotating much faster—at about 22,000 rpm. The gearbox is needed to reduce the high rpm provided by the turbine shaft to the much lower rpm (1,800 rpm) needed by the generator shaft. Incidentally, most gas turbine system OEMs do not make their own gearboxes; such is done by specialty manufacturers.

As far as needed maintenance for a typical gearbox, there is not much that needs doing. The gearbox is a very robust unit and there is rarely a problem with it. Typically, they will last for 50,000 hours. After that, the manufacturer usually recommends replacing the gearbox with a rebuilt unit.

We use synthetic oil in the gearbox. Very durable and very resistant to thermal or mechanical breakdown, such oil typically lasts for several years. Most gas turbine system owners pay little attention to the gearbox. Occasionally it is possible for a bearing to fail or a gear to chip—failures that would usually cause the gearbox to rumble.

The cutaway view of the Kawasaki GPB15X Combustor: 1. Fuel Injector 2. Mixer 3. Preburner 4. Catalyst 5. Burnout Zone

DE: But doesn't the owner have to do some maintenance to the gearbox? Or at least to check out its proper operation from time to time, to monitor its performance—so-called predictive maintenance—for any warning signs that there might be trouble ahead?
Pozzi: Yes indeed! Such predictive maintenance is widely used in the electric-power generation field. Maintenance technicians typically collect baseline performance data on the gearbox unit and on the electric generator—recording periodically such things as vibration levels, temperatures, and pressures.

As equipment ages, there are changes in these variables; they tend to drift away from the baseline that was established when the equipment was new. An increase in the vibration levels of the generator or of the gearbox, for instance, very likely indicates a problem with the bearings.

Maintenance technicians will also periodically analyze oil circulating through bearings in the gearbox, in the generator, and in the turbine. They are looking for metal particles suspended in the oil and signs that the oil has been breaking down due to overheating—indications that something is wrong with the bearings. This is all part of predictive maintenance.

Currently I am working on developing some predictive-maintenance software. This will calculate the deviations of measured variables from their baseline values. If the measured variable values drift off too far, then the computer will sound an alarm to alert the operator. Or the computer could alert the gas turbine system OEM by calling him on the phone and conveying the abnormal information.

We like to sell gas turbine systems equipped with appropriate telemetry. In this way, Kawasaki can have operating data on its customers' gas turbine systems sent via telemetry to Kawasaki so experts there can monitor the operation of customer units.

Or conversely, Kawasaki can call the computer-monitoring system for a gas turbine system over a regular phone line and check on the current status of the system, comparing its current operation with baseline profiles, etc.

DE: Predictive maintenance, then, is already widely used for gas turbine systems being used in distributed energy applications?
Pozzi: Yes. The use of predictive maintenance is not confined to large central electric power plants. It is also used in distributed energy applications, where the electric power being generated is in the 1- to 20-megawatt range.

In these gas turbine distributed energy applications, it makes sense to monitor the vibration of major equipment, where any kind of unscheduled downtime can be very costly. A change in the vibration of a piece of equipment indicates that something has changed. There might be a shaft bearing going bad; or the lube oil may have lost its properties, etc.

Such changes are a warning that something is wrong and that corrective action is needed. For instance, we might have to change the bearings in the electric generator, or at least change the oil used to lubricate its bearings—steps if taken expeditiously may prevent the generator from failing entirely.

The cutaway view of the Kawasaki GPB180

Here, our predictive-maintenance system is suggesting that we take corrective action soon—but not necessarily immediately. We try to work the needed maintenance into the customer's schedule. We will ask our customer: When is your next scheduled shutdown? Can we come in at that time, then, and change the bearings on the electric generator?

In the old days, many maintenance people merely waited for the equipment to fail before taking action. Or, at most, they might give equipment an extra shot of grease. Today, the approach is much more scientific. Invoking the methods of predictive maintenance, we continuously monitor the condition of major equipment. In that way, we can observe changes over time. And when we see enough drift from initial baseline conditions, we schedule maintenance—e.g., a bearing change.

DE: How long does the oil last in these lubrication systems?
Pozzi: That depends on how hot the oil gets during service and how often lubrication oil filters are changed. We use synthetic oil in our gas turbine systems, and that can last for years.

DE: Could you please expand a bit more on the maintenance needs of the electric generator used with a 1.5-megawatt gas turbine system?
Pozzi: Certainly. First, let me say that most OEM gas turbine system manufacturers do not make their own electric generators but buy the appropriate generators from electric-generator manufacturers. Most commercial and industrial customers want an electric generator with an output of 480 volts, 60 cycles/sec, three-phase. But requirement can vary.

Usually, the customer will tell us his electrical requirements, what sort of output electrical power he needs. We will then select the appropriate electrical generator from the numerous manufacturers on the market and integrate it into the gas turbine package.

Yet, like gearboxes, electric generators are also units that are largely trouble-free. Their most common failures are shaft bearings. It is not uncommon for an electric generator to run trouble-free for 13 years or more. Perhaps after that it might need bearings changed. Occasionally, there might be a failure of diodes. I've seen this a few times after units were affected by lightning-induced voltage surges.

The only maintenance that needs to be done on these generators is the predictive maintenance previously mentioned. This involves taking periodic oil samples from the oil circulating through the generator's bearings and analyzing them for signs (metal particles, oil breakdown) of pending trouble. It also calls for monitoring generator vibrations for signs of deviations from the initial baseline.

All work on gas trubines, from installation to periodic maintenance, is performed by highly trained technicians.

DE: How long does an electric generator typically last?
Pozzi: The electric generator will last a very long time if properly maintained. Its shaft bearings will typically last between six and eight years, then need to be replaced. With periodic bearing replacement, an electric generator can typically last for 20 or 30 years—as problems in its windings and other electrical components are rare.

So the bearings, then, are the most vulnerable component of a generator. Actual bearing lifetime will depend upon how cool the oil circulating through them is kept. Some bearings are lubricated by a splash system, with oil in a reservoir splashing up to keep metal surfaces coated. In a better system, oil is actually pumped through the bearings, cooled, filtered, then recirculated through the bearings in a continuing cycle.

DE: Earlier, you talked about predictive maintenance and the desirability of being able to remotely monitor a gas turbine installation. But is such telemetry capability really all that important? Isn't it more like frosting on the cake?
Pozzi: No! It's a lot more than mere frosting! For it allows the turbine system OEM to see the same information as the customer. And this enables us, as the experts on the system, to be able to identify potential problem areas and to recommend solutions.

Most importantly, such remote capability saves the time and the expense of having one of our service technicians travel to the customer's site. Such can greatly reduce the customer's downtime—during which he could be losing millions of dollars. This remote monitoring and access capability also reduces the cost of maintenance, enabling the turbine system OEM to offer a maintenance service contract to customers at a lower price than otherwise.

Further, providing operations and maintenance services remotely like this is a way to get the customer more involved in the care and maintenance of his own equipment. Using the information gathered from the data telemetry system, the manufacturer can guide the customer's maintenance staff in what to look for and what corrective steps to take. In a word, the customer will be troubleshooting his own gas turbine system under the guidance of the OEM.

Using this remote diagnostic approach, Kawasaki has been able to achieve a 95% troubleshooting success rate. That is, we are able to resolve 95% of customers' operating problems without having to travel out to their sites.

DE: You speak of both "monitoring" and "access" of the remote turbine system. What is the distinction between "monitoring" and "access"?
Pozzi: By monitoring I mean merely collecting data and transmitting it via telemetry—data transmitted over ordinary telephone circuits—from the remote gas turbine system. By access, on the other hand, I mean being able to actually intervene and control the remote turbine system—e.g., making changes in software and even starting the turbine remotely.

This remote monitoring and access capability is a very powerful tool. Example: One of our customer's gas turbine systems was not starting. One thermocouple in the system was indicating a high temperature, sending the entire turbine system into alarm status. We suspected the thermocouple might be bad, indicating a dangerously high temperature when in fact conditions were normal. To check out this hunch, we looked at the reading from a nearby thermocouple that was monitoring the temperature of circulating oil. That temperature reading was normal, the proof we needed to confirm that the first thermocouple was indeed malfunctioning.

To get the turbine system up and running with minimal delay, we made changes—remotely—to the control-system's software. These changes in effect told the control computer to ignore the readings from the bad thermocouple. With that change, we were able to remotely restart the turbine. Subsequently we dispatched a repair technician to the site to replace the bad thermocouple—at a time of low-power usage for the customer.

DE: Does the owner of a gas turbine system in the 1.5-megawatt class size need to have his own in-house maintenance staff?
Pozzi: We have some gas turbine system owners who want to get deeply involved in the maintenance of their equipment. It is important that the owner's O&M staff work very closely with the gas turbine system OEM during the startup and commissioning phase. At that time, they will learn much about proper turbine system O&M—including how to properly test various components.

OEMs also offer customers formal training classes on equipment operation. Kawasaki sends its trainer directly to the customer's site: 50% of the training is in the classroom; the remainder is in the field with the equipment itself.

DE: Does that mean, then, that every company needs staff for maintaining its gas turbine system?
Pozzi: Certainly every company needs to have well-trained equipment operators. But not every company needs to have gas turbine maintenance staff per se.

In an attempt to save money, some customers take steps to train their boiler-room operators to maintain their gas turbine equipment. Others, with no in-house maintenance staff of their own, rely heavily on the gas turbine system OEM for continued maintenance.

For applications in oil fields, there are already many maintenance people on staff for maintaining oil-drilling equipment. Accordingly, such customers tend to get deeply involved in turbine system maintenance, replacing filters, troubleshooting, etc., thereby saving money on their service contract.

DE: Should a company, then, with an already existing maintenance staff, perform all turbine system maintenance itself? Or are there certain tasks that should be done by in-house staff, other tasks by experts from the gas turbine system OEM?
Pozzi: As I see it, there are three levels of maintenance:

Level l—This has to do with understanding the principles and characteristics of gas turbine systems and how to operate them safely and efficiently.

Level 2—This has to do with operating the turbine system on a day-to-day basis, including the performance of routine troubleshooting, sometimes in telephone consultation with the OEM manufacturer. It also involves periodically performing certain basic yet important maintenance tasks: changing various oil filters; changing inlet-air filters; etc.

Level 3—These are the more advanced maintenance tasks that need to be performed by highly trained technicians—e.g., the 16,000-hour borescope inspection of the turbine, the 32,000-hour turbine replacement. Said another way, the turbine owner's maintenance staff could be trained to do the required quarterly inspections by people doing Level-2 maintenance. But the annual inspection, including borescope examination, the 16,000-hour hot-section inspections, and the 32,000-hour turbine changes need to be done by highly trained Level-3 maintenance people from the gas turbine system OEM.

DE: Is the major gas turbine maintenance you speak of almost always done by the OEM? Or are there some third-party maintenance companies out there that are doing some of this work?
Pozzi: Usually the major maintenance is done by the OEM. Nonetheless, there are third-party turbine-maintenance companies out there. And unless those companies have a partnership with the OEM, they will not fully guarantee their work. Typically, they will warranty their labor—but not their parts. By contrast, OEMs typically guarantee both labor and parts.

DE: Do many customers opt for service contracts with gas turbine system OEMs?
Pozzi: Yes. One reason is that the warranty on the gas turbine system is usually predicated upon having an extended service agreement.

Another major consideration is that many gas turbine systems are financed by a bank or other financial institution. And, typically, they want the gas turbine system to be protected by a long-term service agreement between the OEM and the customer. Often, these agreements are for 20 years, the typical lifetime of a gas turbine system. The financiers want to make sure that the turbine systems they have invested in are well-maintained. In that way, if a company using a turbine system were to default on its payments (payments for either buying or leasing the turbine system), the financial institution would still have some valuable collateral to protect itself against financial losses.

In effect, we do not sell turbine systems to commercial and industrial customers: we sell maintenance agreements. The financial institution is often the owner of the turbine system and leases it for use to the customer.

With this leasing approach, both the monthly lease payments and the monthly service-contract payments become, from an accounting perspective, part of the turbine-using company's operating expenses. This approach is a smart financial move. It is better for the company's bottom line than if it were to actually become the outright owner of the gas turbine system.

DE: Most interesting. What about replacement parts. Where does a company go for those?
Pozzi: Although there are third-party parts suppliers out there, the gas turbine system OEM is usually the least expensive place that the turbine operator can go to for replacement parts. For these third-party maintenance firms themselves often have to go back to the OEM to buy the parts.

Kawasaki provides its customers with a list of every part in the gas turbine system, its part number, and the original manufacturer. The only parts they have to buy through Kawasaki are parts for the turbine itself, because that is the only equipment that Kawasaki makes; the other parts of the gas turbine system (e.g., gearbox, electric generator, fans, filters, etc.) are made by other manufacturers; and Kawasaki buys those components from those suppliers when it is putting together the gas turbine system.

DE: Are there aftermarket manufacturers of gas turbine system components?
Pozzi: Most components are made only by the established equipment manufacturers. Consider the turbine itself. The various components that go into the turbine are either made by Kawasaki or by parts makers working closely with Kawasaki. Kawasaki does not make all parts that go into the turbine—e.g., it does not make bearings. But it does carefully specify to the bearing manufacturer what requirements that bearing must meet.

Accordingly, a turbine operator needs to go through Kawasaki to purchase any part having to do directly with the gas turbine itself. But for any components outside the turbine—the gearbox, the electric generator, etc.—the customer is free to purchase that component wherever he chooses. Nonetheless, we encourage customers to buy all components of the gas turbine system through their OEM because, we believe, they will get a better price there than if they were to run back to the original manufacturer.

If a turbine operator does buy a component from a third-party supplier and it does not meet Kawasaki specs for that component, the warranty becomes invalid. Suppose an operator, for example, replaces an oil-circulation pump. The original specs called for an 1,800-rpm pump, but the replacement pump is only 1,100 rpm. If there is a subsequent failure due to the oil-circulation system, then the warranty would be invalid.

What's the point? If a customer is going to venture out on his own to buy replacement parts rather than go through his OEM, he must be very careful to buy components that meet the OEM's specifications. If not, he will be invalidating his OEM warranty.

GENE DALLAIRE is a former feature writer for Chemical Engineering and Civil Engineering magazines. He currently teaches history at Lansing (MI) Community College.

DE - September/October 2004