Compared to a decade ago, pipeline natural gas is being delivered with more variable heating values and a greater tendency to form damaging hydrocarbon liquids.
By: Robert Swanekamp, P.E., Contributing Editor
The combined-cycle community understands all too well the adverse economic effects of soaring natural gas prices. Rising fuel costs have driven down the average combined-cycle capacity factor to around 30 percent, and have turned many such plants into poor financial performers-at least in the near term. But changes in the natural gas market also have caused adverse operating effects on gas-turbine-based facilities.
Looking to both minimize their production costs and maximize the heat content of their deregulated product-hence the price they can charge on a volumetric basis-many natural gas suppliers have cut back fuel-gas processing and scrubbing activities. This has resulted in greater concentrations of non-methane hydrocarbons in the gas mixture, compared to levels seen a decade or more ago. These Btu-rich hydrocarbons-ethane (C2), propane (C3), the butanes (C4s), etc.-which formerly were extracted from raw natural gas and sold as separate feedstocks, now are common in the pipeline gas that users once thought of as “only methane” (CH4).
The net result is natural gas with more variable heating values and a greater tendency to form damaging liquids. As Thomas Welker, a leader in the natural gas industry since the 1950s and the founder of Welker Engineering Co. (Sugarland, Texas), has famously said, “Clean dry gas is history.”
To fossil fuel power plants, Btus are valuable things, so what’s the problem if the supplier wants to send us some more of them in the form of ethane, propane, or butane? For starters, they cause variations in the heating value that affect the combustion process, potentially causing load swings, emissions excursions and flame instability.
Gas turbine (GT) manufacturers design and tune their combustion hardware-the supply valves, gas manifolds, fuel-nozzle orifices, and so on-for a given set of fuel characteristics. No single characteristic by itself provides an adequate measure for GT users, but a key one is the Wobbe Index (also called the Gas Index). This indicates the relationship between the fuel’s density and its energy content. If you want to do the math, the Wobbe Index is the ratio of the lower heating value to the square root of the specific gravity. If you want to do a little more math, there’s the Modified Wobbe Index, which is even more instructive because it takes into account the temperature of the fuel. The Modified Wobbe Index is the ratio of the lower heating value to the square root of the product of the specific gravity and the absolute gas temperature.
If you want to skip the math altogether, here’s the salient point: Any change in the fuel’s heating value will require a corresponding change in the fuel’s flow rate to the machine, in order to maintain the proper total energy input. For older, diffusion-type combustors, the GT control system likely can handle variations in the Modified Wobbe Index as large as ±15 percent. But for newer, dry low NOx (DLN) combustors, the control window is narrow. Variations in the Modified Wobbe Index of only ±3 percent are believed by many GT specialists to cause serious trouble. The corresponding velocity changes through a DLN system’s precisely sized fuel-nozzle orifices can cause flame instability, resulting in pressure pulsations-or “combustion dynamics.” At the low end of the trouble spectrum, Wobbe Index variations will force GT users to conduct more frequent tuning of their DLN combustors. At the high end of the trouble spectrum, the variations could be large enough and prolonged enough to cause pressure pulsations that catastrophically destroy the combustor and downstream hot gas path components.
Even if fed a consistent fuel composition, DLN combustors are prone to flame instability and combustor dynamics, because of their pre-mixed technology and ultra-lean fuel-air ratios. (See Power Engineering, May 2004, “On-Line Monitoring Systems Help Silence Combustor Humming”). But with changes in fuel composition, combustor dynamics in DLN systems can increase by as much as 50 percent, according to a paper presented at the 2003 International Joint Power Generation Conference by L.O. Nord and H.G. Anderson. In that 2003 study, changes in the fuel composition also changed the NOx emissions of the machine, causing excursions above the 15-ppm permitted level.
Those results do not surprise GT specialist Jeff Fassett, president of IEM Energy Consultants (Alexandria, Minn.), although he points out that hard data such as this are hard to come by. Quantifying the effect of Wobbe Index changes-and changes in other fuel indices-is still a developing science, particularly for DLN combustors that represent relatively new technology. “That requires real-time detailed analysis of the fuel, coupled with real-time continuous measurement of the combustor dynamics,” Fassett says. “Few GT plants are that well instrumented.” Nord and Andersen conducted their study using time-resolved data from a specially instrumented Alstom GT11N1-EV machine.
While real-time data from individual power plants are scarce, variations in Wobbe Index are known to be significant for U.S. pipelines on a regional basis. Consider the recent increase in coal-bed methane production in the Rocky Mountains and Appalachian Basin. Gas from these regions is nearly pure methane, and therefore has a lower heating value than historical pipeline averages. Meanwhile, the heating value of gas produced in other regions is rising, due to the reduced level of processing mentioned above (Figure 1).
In addition to variations by region, there can be significant variations in Wobbe Index for an individual power plant purchasing its gas from different pipeline suppliers, as corporate fuel managers try to exploit daily or weekly changes in gas price.
Steaming toward our shores
In the future, U.S. plants could face even greater variability in their fuel’s Wobbe Index as more pipelines are filled by imported liquefied natural gas (LNG). The power industry’s increasing demand for natural gas, coupled with soaring prices, have made LNG transportation economically viable on a continent that long subsisted on its domestic supplies. Four LNG import terminals currently serve the United States: Elba Island, Ga., Cove Point, Md., Everett, Mass., and Lake Charles, La. Note for history buffs: Lake Charles is the birthplace of the LNG industry. In January 1959, a converted World War II Liberty ship carried the world’s first load of LNG from Lake Charles to Canvey Island in the United Kingdom.
Today, some two dozen import terminals are under development in North America, according to the Energy Information Administration (EIA), and shipbuilders report more than 60 orders for new LNG tankers, further bolstering the outlook for trans-ocean trade (Figure 2). According to the EIA’s June 2004 report, LNG imports are projected to rise from around 900 billion scf/yr in 2003 to 1700 billion scf/yr in 2008. By 2025, analysts expect the fuel’s contribution to U.S. pipelines to rise from its current level of less than 2 percent to more than 15 percent.
What does this mean for GT users? Economics dictate that LNG shippers will purchase product from a wide range of gas fields and processing plants around the world. So an LNG tanker from Qatar offloaded into the pipeline system one week could deliver gas that’s quite different in composition than a tanker from Australia offloaded into the same pipeline the next week.
According to the Natural Gas Council-an industry working group studying LNG and other gas issues-impurities such as water and carbon dioxide along with hydrocarbons C5 and above (the pentanes, hexane, etc.) must be removed from the raw gas to prevent them from forming solids during liquefaction (the cooling process that lowers the LNG to approximately -260F prior to transport). But the resulting product still contains appreciable-and varying-concentrations of ethane (C2), propane (C3), and some butane (C4), which changes LNG’s heating value compared to traditional North American supplies.
The development of integrated gasification combined cycle (IGCC) plants will introduce additional issues for GT users because the heating values of gasified fuel can be as low as one-third that of pipeline gas (see sidebar, page 54).
While variations in heating value are a serious concern to GT users, the more-prickly problem caused by the changing composition of natural gas is the change that higher hydrocarbons produce in the fuel’s dewpoint.
All GT OEMs have the same specification for the amount of hydrocarbon liquids allowed into their gas-fired combustors: zero! The potential consequences for operators who violate this unambiguous spec include coking deposits on fuel injectors, auto-ignition of the fuel/air mixture upstream of the burner, flame instability if a small amount of the hydrocarbon liquids reach the combustor, and dangerous over-fueling if the hydrocarbon liquids have condensed and collected over time and are then “slugged” into the combustor (Figure 3).
Hydrocarbon liquids have even raised issues with the duct burners installed in the heat-recovery steam generators at some combined-cycle plants, reports John Koza, duct burner technical advisor for Forney Corp. (Carrollton, Tex), a leading supplier of supplemental firing systems.
Combined-cycle plants began seeing substantial GT damage caused by hydrocarbon liquids in the mid-1990s. Failures in the published record include Progress Energy Florida’s Tiger Bay station (then owned by Destec), Constellation Generation Group LLC’s Perryman station, and an EPON station in the Netherlands. “Those were just the ones reported,” Fassett states. “There have been many, many more, on turbines of virtually all makes and models.” At the time, Fassett was an engineer at Tiger Bay who worked closely with GE Energy’s Colin Wilkes on the failure investigation of that plant’s 7FA machine. Their efforts led to a surprising conclusion about hydrocarbon liquids.
Margin of superheat
For years, GT designers and users understood the need to eliminate liquids, which can come not only from higher hydrocarbons in the pipeline gas, but also from residue of glycol used in the construction of pipelines, entrained oils from the pipeline’s or the power plant’s fuel-gas compressors, even carryover of water from the GT compressor-wash system.
The effective preventative measure is to keep the fuel gas safely above its dewpoint. This requires, for starters, an appreciation for the Joule-Thompson effect, which states that for every 100-psi drop in gas pressure there will be a corresponding drop in gas temperature of approximately 7 F. As Wilkes explains in his paper “Gas Fuel Clean-Up System Design Considerations for GE Heavy-Duty Gas Turbines,” gas enters the typical 7FA site at 900 psia and is reduced to 450 psia before it enters the turbine, thus the Joule-Thompson effect produces a 31.5 F drop in temperature.
To make sure that this temperature drop does not cause hydrocarbon liquids to condense and damage the GT, GE and the other manufacturers began to emphasize the need for fuel-gas heaters, sized and controlled to maintain a specified margin above the fuel mixture’s dewpoint. For its V84.2 and V84.3 machines, for instance, Siemens specified a margin of 18 F. For its heavy-duty Frame machines, GE specified a much more conservative 50 F margin.
But margin above what? The surprising lesson from the Tiger Bay investigation, Fassett reports, was in the determination of the fuel mixture’s actual dewpoint. The standard practice to determine this critical parameter is laboratory analysis of a gas sample, using the ASTM D1945 method. This method determines the percentage by weight of each individual hydrocarbon in the fuel up to pentane (C5), but all higher hydrocarbons are lumped together and reported only as one entry labeled “C6+.” In a presentation to the 7FA Users Conference a few years ago, Wilkes revealed the staggering difference that can exist between the standard method and the more accurate “extended” method, which analyzes the individual hydrocarbons all the way up to C14. For one gas sample at 400 psia, the standard analysis calculated a hydrocarbon dewpoint of -22F, Wilkes reported. For that same gas at the same pressure, the extended analysis calculated a hydrocarbon dewpoint of +89 F-a difference of over 100 F! Clearly, even the conservative 50 F margin of superheat, if added to an artificially low dewpoint calculation, will provide inadequate protection against hydrocarbon liquids.
C6+ method is inadequate
By now, most users of heavy-duty Frame GTs have installed a dewpoint heater, Fassett says. And most of them clearly understand the need to keep those heaters working to provide the OEM-specified margin to dewpoint. Unfortunately, he reports, most users are still calculating dewpoint using the standard or “C6+” method. “That just doesn’t cut it anymore,” Fassett asserts. He strongly recommends that users perform an extended C14 fuel analysis, as is recommended by GE in its fuel-gas specifications, to determine the proper setpoint of their dewpoint heaters. The extended analysis is more complicated and more expensive, Fassett concedes, and not all laboratories can perform this service. “But it’s the only analysis that will give you an accurate dewpoint.”
Another solution to the dewpoint dilemma that some GT users are employing is to go way beyond the OEM-specified margin and simply heat all fuel gas to approximately 360 F. Some of the OEMs are pushing this idea, not only to eliminate hydrocarbon liquids, but also, they say, to improve turbine efficiency. These heaters-referred to as “performance heaters” to distinguish them from the “dewpoint heaters” that provide only the 50 F margin of superheat-typically utilize intermediate-pressure steam extracted from the plant’s Rankine cycle.
Engineers point out that while they may in fact raise the thermal efficiency of the GT itself, they likely are lowering the overall efficiency of a combined-cycle plant, by robbing steam from the steam turbine. A report published in 2001 by Flowtronex International (now Integrated Flow Solutions Inc. (Houston, Texas), calculated that a fuel-gas heater driven by electric-resistance coils and providing 50 F superheat for a 7F machine consumes 740 kW of energy-or over 6 million kWh per year for a baseload plant. Note that many acceptance specifications for new projects or post-overhaul testing require efficiency tests only of the GT-not of the overall combined cycle. Bottom line: The GT OEM and the EPC contractor may benefit from these performance heaters, but in the long run the guy paying the fuel bill may not.
Given the problems with varying heating values and hydrocarbon liquids in pipeline gas, GT owners and OEMs have joined together to press for regulatory assistance. On a national level, the Federal Energy Regulatory Commission (FERC) regulates gas quality, through its authority to approve the specifications of each interstate pipeline’s tariff. However, the FERC-approved pipeline tariffs specify only a minimum heating value in the fuel-a simple criterion that may be good enough for residential heating systems, but certainly not for advanced gas turbines and sensitive DLN combustors.
In an effort to recommend more rigorous criteria to FERC, the Washington-based Natural Gas Council formed a working group in 2004. The working group is comprised of gas suppliers and distributors, but also power producers and equipment manufacturers, including GT OEM reps Bruce Rising of Siemens Westinghouse Power Corp. and Colin Wilkes of GE Energy. In February 2005, the group presented a white paper to FERC with their recommendations on gas quality and interchangeability. Note that “quality,” in this context, refers to the variations in hydrocarbon content of natural gas that can lead to liquids forming in the gas stream. “Interchangeability” is the ability to substitute one gaseous fuel for another in a combustion application without materially changing operational safety, efficiency, performance or materially increasing air pollutant emissions.
In its February publication, the Natural Gas Council urged FERC to establish the following specific criteria:
- Maximum Wobbe Index = 1400 (based on gross or higher heating value at standard conditions of 14.73 psia, 60 F)
- Maximum variation in Wobbe Index = ±4%
- Maximum mole percent of inert gas = 4%
- Maximum mole percent of butanes-plus = 1.5%
- Maximum volumetric content of C4+ = 1.5%
Unfortunately for GT operators, these recommendations, while a step in the right direction, do not meet all of the requirements specified by the OEMs for many of the machines currently operating in the U.S. fleet. For example, the variation in Wobbe Index and the volumetric content of C4+ exceed Siemens Westinghouse Power Corp.’s current fuel specifications. Calpine Corp’s Nicole Prudencio, who also participated in the working group and recently presented the results to the W501F Users Group, is troubled by the fact that the OEM specs were not adopted. Instead, the working group challenged the GT manufacturers and users to prove that the council’s recommended criteria were not adequate. To date, that challenge is not being undertaken by the Natural Gas Council working group, Prudencio reports.
Taking Action in the Plant
Gas turbine owners and operators cannot stand still while initiatives at the regulatory level are being bantered about. At the plant level, they are working to upgrade their fuel-gas monitoring and conditioning systems, to combat the problems with pipeline gas.
By now, most GT users have installed on-line instruments to monitor the heating value of their incoming pipeline gas-such as the microprocessor-based gas chromatograph supplied by Daniel Measurement and Control Inc. (Houston, Texas), a division of Emerson Process Management, installed at Tenaska Inc.’s Paris station in Texas.
Users of DLN combustors are installing even more sophisticated monitors to guard against flame instability. These on-line systems continuously measure dynamic pressure pulsations to provide early warning that a combustor is out of tune, which could be caused by heating-value fluctuations in the fuel gas. Based on data from the online system, GT users are able to re-tune their combustors as needed to ward off catastrophic failures. To date, most combustor dynamics monitoring systems have been installed by the OEMs. A handful of non-OEM companies also offer the equipment. These include Power Systems Manufacturing (Boca Raton, Fla.), Control Center LLC (Orlando, Fla.), and KEMA (Burlington, Mass.).
In addition to better instrumentation and dewpoint heaters, upgraded fuel-gas conditioning systems are being retrofit to many GT plants. These systems strip out any hydrocarbon liquids that might form, as well as other liquid or particulate contaminants in the gas stream. It’s important to understand that the GT OEMs typically do not supply robust fuel-gas conditioning systems. However, their specifications make it clear that the GT user must. According to GE’s Colin Wilkes, a proper fuel-gas conditioning system contains:
- A dry scrubber (meaning one that functions without the use of oils or solutions) to remove the bulk of both liquid and solid contaminants. These typically feature multiple cyclones that use the dynamics of centrifugal force and gravity. The only routine maintenance required is blowdown of the collection sump.
- A coalescing filter comprised of a number of fibrous filter elements attached to a tube sheet, typically arranged in a vertical pressure vessel. The filter elements need to be changed when a predetermined pressure drop is reached for the specified volumetric flow rate of gas. For a baseload GT, Wilkes recommends two coalescing filters in a duplex arrangement, so that fouled filters can be replaced while the GT is operating.
- An inertial separator, in which the gas passes through an inlet baffle to remove large liquid droplets, and then through a series of angled vanes that impart inertial forces on the smaller liquid droplets.
Fassett points out that proper equipment arrangement is important too. He continues to see plants where the dewpoint heaters are incorrectly located upstream of the scrubbers and filter/separators. “The dewpoint heater needs to be the final defense, as close to the gas turbine as possible,” Fassett says.
One of the suppliers of fuel-gas conditioning systems busy with GT retrofits is Peerless Mfg Co. (Dallas, Texas), a supplier to the natural gas industry since 1937. That’s when the infamous New London School disaster occurred, which killed more than 300 children and teachers when an unrecognized gas leak from the boiler room was ignited by a spark in the wood shop. In response to that disaster, Donald Sillers, the founder of Peerless, developed devices for odorizing natural gas and helped to pass a new Federal law mandating the injection of mercaptan, which now gives pipeline gas its identifying smell.
Today, the Peerless “Absolute Separator” is a popular choice for GT users because it performs the three types of required gas conditioning, all in one package. As the company’s Bill Asbury, application engineer/fuel-gas conditioning systems, explains, fuel gas goes through primary filtration using an inlet sump for gravity dropout with an inlet tangential baffle to give the gas a spin, causing some coalescing on the vessel walls.
In the second stage, the fuel gas is dry scrubbed, entering a cyclone tube that sets up a swirling motion. Solid and liquid contaminants are thrown outward and drop from the cyclone tube. The swirling gas then reverses direction at a vortex and rises through the tube’s outlet.
Finally, the fuel gas passes through high-efficiency coalescing filters, from the inside of the elements to the outside, where contaminants diffuse and impinge on the closely spaced surfaces (Figure 4). Any remaining liquid droplets agglomerate into larger droplets and emerge on the outer surface of the coalescing element, before running down the element and collecting in a chamber. “The coalescing elements are there for final polishing only,” Asbury emphasizes. “They don’t see more than 40 ppm of contaminants, which enables long run-time for each element before it needs to be replaced.”
Overall, the Peerless Absolute Separator delivers separation efficiencies of 99.999 percent of droplets 0.3 to 0.6 microns in size, with a maximum carryover of 0.001 ppm by mass depending on the coalescing media required by the customer.
Cyclones meet hurricanes
One of the most recent installations of this cyclone-based system is at a cogeneration plant in Beaumont, Texas-the town readers may recognize as the landfall location for the eye of Hurricane Rita in September. Peerless was called by the customer months before that, Asbury explains, after the existing dewpoint heater and horizontal coalescer failed to remove hydrocarbon liquids from the fuel-gas stream. Plant operators reported that the liquids caused nearly $2 million of damage to hot gas path components.
Figure 5. The three-stage absolute separator recently installed at this cogen plant delivers a separation efficiency of 99.999 percent of droplets 0.3 to 0.6 microns in size. Photo courtesy of Peerless Mfg Co.
The operators solicited quotes from several suppliers for an upgraded fuel-gas conditioning system, but selected Peerless and its vertical, three-stage Absolute Separator (Figure 5). The new system was installed in early September, and will enter service when the cogen plant completes its recovery from Hurricane Rita.
Gasified Fuels Spark Interest
For the most part, the coal sector and the gas turbine sector have been very distinct industries. But in the not-too-distant future, the two may be joined in lockstep, thanks to gasification.
Gasification is an endothermic process in which a solid fuel is thermo-chemically converted into a low-Btu gas. The resulting synthesis gas (syngas) can be used like natural gas to fuel a wide range of power plant technologies-boilers, spark-ignited reciprocating engines, fuel cells, gas turbines, etc. Syngas produced from coal or petroleum coke is sparking interest among developers of the next generation of combined-cycle projects, and among some owners of existing gas turbines as a replacement fuel for natural gas.
Whether they’re considering syngas for new IGCC projects or for existing plants looking to convert, owners need to understand that its composition and energy content are substantially lower than other fuel gases. Heating values of syngas produced by oxygen-blown gasification fall in the range of 200-400 Btu/scf, with hydrogen content of approximately 30 percent by volume. The syngas produced by air-blown gasifiers tend toward even lower heating values: 100-150 Btu/scf. They also produce lower hydrogen content, ranging from 8 to 20 percent by volume. Note that in between the low-Btu gasification fuels and high-Btu natural gas resides an array of medium-Btu byproduct fuels from chemical plants and refineries, which also can be fired in advanced gas turbines.
Because of the lower heating values, the volumetric flow rate of fuel in a syngas-fired turbine might be as much as four to five times that for a natural-gas-fired unit. In addition, the rate of diluent injection required for NOx control can be as high as the fuel flow rate. The increase in total mass flow through the machine, compared to when it was firing natural gas, would require virtually all systems in a retrofit unit-from the fuel stop valve through the first-stage nozzle-to be redesigned and replaced.
Also, low-Btu fuels dictate the use of diffusion-type combustion systems, because of their low heating values and high hydrogen content. So if an existing gas turbine is being retrofit from a DLN system, the site emissions limits may force the addition of an SCR system-adding further to the capital cost of a retrofit project.
Developers of new projects are studying the considerable record that gasification technology has amassed. There are three major suppliers with proven technology: Conoco/Philips, GE Energy, and Shell. According to the Gasification Technology Council (Arlington, Va.), there are more than 160 commercial gasification plants in operation, under construction, or in development in 28 countries.
Of three DOE-funded IGCC projects built in the US in the 1990s, Tampa Electric Co.’s Polk Power Station clearly has the most impressive record. Commissioned in 1996, the 260-MW facility has operated at high capacity factors throughout its history. Polk comprises an oxygen-blown coal gasifier based on Texaco technology (recently acquired by GE Energy) fueling a GE 7F gas turbine. Mark Hornick, general manager of the Polk Power Station, reports that syngas firing has produced no significant problems, and that the gas turbine reliably cranks out 192 MW. Unlike many natural-gas-fired gas turbines, output remains consistent, summer or winter, thanks to power augmentation from the injection of nitrogen – produced from the plant’s air separation unit.
Because the syngas is low in Btus and high in hydrogen content, Polk’s combustor employs diffusion technology, rather than DLN. “The diffusion burners along with nitrogen injection and syngas saturation can meet our 15-ppm NOx limit without an SCR,” Hornick explains. Emissions stay within the permitted limit with little trouble, and typically run in the range of 10-12 ppm.
The combustor and all other hot gas path components also have proved durable, Hornick reports. His team has conducted turbine maintenance at the regularly projected milestones-combustion inspection at 8000 hours, hot gas path inspection at 24,000 hours, and major overhaul at 48,000 hours.
Polk’s syngas does contain some contaminants, Hornick explains, chief among them sulfur. An amine-based scrubber removes more than 98 percent of the sulfur, converting it to sulfuric acid that can be recycled to the chemical industry. According to Hornick, the remaining sulfur that enters the gas turbine has caused no noticeable problems. He offers a note of caution, however, to operators considering a retrofit project: If Polk had an SCR, that remaining sulfur could combine with the injected ammonia to form deposits on downstream HRSG tubes.