Steam Turbine Bypass Systems

H R S G S • GAS TURBINES • STEAM TURBINES • A U X I L I A R I E S

Fall 2003

www.psimedia.info/CCJ.htm

3 SPECIAL REPORTSteam turbine bypass systemsRobert W Anderson and Henk van Ballegooyen

10 GENERATION ECONOMICSTo maximize profit, keep one eye on the market, the other on your plantJason Kram and Doug Logan

13 PROFILE: RAVENSWOODHow to shoehorn 250 MW into a parking lot

1 7 Predicting gas prices more art than scienceKen Walsh

19 OPINIONWe’re government, we’re here to help!James F Wood

21 POWER MARKETINGHow deregulated energy markets impact plant revenue, operationKristin Domanski and Paul Flemming

23 Assessing the true cost of cycling is challengingFrank J Berte, David S Moelling, Craig A Udy

25 Condition monitors warn of impending generator failure

27 LONG-TERM SERVICE AGREEMENTSUnderstanding ‘fine print’ can mean the d i f f e r e n c e between financial success, failureJeff Fassett and Richard E Thompson II

30 User Group Activities33 Design/Operating Ideas35 Business Partners36 Professional Services

How to avoid serious andcostly O&M problems. . .3

When a combined-cycle plant is operat-ing at steady-state, 100% load, it’s anelegant thing to behold. Well, maybenot to the artist or the fashion design-

er, but it is to the engineer who can appreciate itsunmatched thermal efficiency, small footprint, lowwater usage, and single-digit air emissions. Unfor-tunately, during transient and low-load conditions,the marriage of the gas-tur-bine cycle with the steam-tur-bine cycle begins to look a bitawkward.

During these conditions—the ones of greatest concernare startup, shutdown, and asteam-turbine trip—the gasturbine (GT) is producing somuch exhaust heat at suchrapid rates of temperaturechange that, if it were imposeduncontrolled onto the steamside, thermal ramp-rate limitswould be violated in thick-walled components of theheat-recovery steam generator(HRSG) and/or the steam tur-bine (ST). Such transients arebeing encountered more fre-quently today than in the pastbecause of the need to cyclecombined-cycle units designedfor base-load service. In addi-tion, the transients are moresevere in magnitude because the latest combined-cycle plants typically:

■ Feature large GTs with high exhaust flowsand temperatures.

■ Operate at high steam temperatures andpressures.

■ Incorporate reheat steam, which significantlyincreases control complexity during plant startup.

■ Include two or three GT/HRSG blocks supply-ing a single ST, which complicates both startupand shutdown procedures.

■ Demand more aggressive startup, shutdown,and load-response times.

One solution to the imbalance between GTexhaust heat and the steam system’s thermal tran-sient limits is to install a bypass damper and abypass stack upstream of the HRSG, enabling theGT exhaust gas to be vented directly to atmos-phere. This permits simple-cycle operation, but can

still allow serious thermaltransients in the HRSGbecause it is difficult to pre-cisely modulate the bulkybypass damper and complywith National Fire ProtectionAssociation pre-start purgerequirements when transi-tioning from simple- to com-bined-cycle operation. Thissolution also is capital- andmaintenance-intensive, so itis installed in very fewp l a n t s .

Another solution is toallow the HRSG to generatesteam, but to vent it directlyto atmosphere until thesteam-side metal is properlywarmed up. However, theroutine use of so-called “skyvents” during every plantstartup incurs costly losses ofdemineralized water, not tomention problems with envi-

ronmental regulators and plant neighbors whoobject to the imposing noise and plume.

Cascading bypass systemThe most common method to manage the thermalimbalance during times when the steam turbinecannot use all the steam produced is the “cascad-ing” bypass system (Fig 1). In this scheme, thehigh-pressure (HP) superheated steam generated

COMBINED CYCLE JOURNAL, Fourth Quarter 2003 3

1. Steam-turbine bypass system isbeing installed at a 250-MW, 1-by-1combined-cycle plant. HP (A) and IP (C)valve bodies await receipt of actuatorsand valve gear. B is spray watermanifold for HP desuperheater.Desuperheated steam is routed tosparger in steam-turbine exhaust duct

SPECIAL REPORT

S t e a m Tu r b i n e Bypass SystemsA robust condenser dump system is critical to a combined-cycle plant’s ability torespond to transients. Inattention to detail at the design stage can lead to serious andcostly O&M problems

By Robert W Anderson, Progress Energy Inc, and Henk van Ballegooyen, Gryphon International Engineering Services Inc

during startup is diverted around the HP section ofthe turbine, through a pressure-control valve (PCV)and an attemperator, and into the cold reheat line.The primary job of the PCV is to control HP d r u mpressure, thereby limiting thermal stresses on theH P drum—one of the most vulnerable componentsof the HRSG because of its thick walls.

The attemperated steam then mixes withsteam from the intermediate-pressure (IP)drum and is directed through the reheater.

After passing through the reheater, the steam isdiverted around the IP and low-pressure (LP)steam turbines as it travels through a second pres-sure-control/attemperating station (hence thename “cascading”) located off the hot reheat line,before it is directed through a dump tube (or“sparger”) into the condenser. This second PCV’sprimary job is to control reheat steam pressure,thereby limiting IP drum level swings. The secondpressure-control/attemperating station often isreferred to as the “hot reheat bypass” or the “LPbypass” (Fig 2).

As the ST satisfies its warm-up requirements,the control valves in both the HP bypass and thehot reheat bypass progressively close and the STpicks up steam load.

Note that a similar solution applied at somecombined-cycle units is referred to as “parallelbypass.” In this design, the steam generated atstartup in the HP and IP drums is attemperatedand sent directly to the condenser, without flowingthrough the reheater. Another feature of the paral-lel bypass system is that the spray water for theattemperation process comes only from the con-densate pumps. By contrast, spray water in thecascading bypass system is taken from the dis-

charge of both the condensate pumps and the feed-water pumps, hence there is a small reduction inthe efficiency loss with the parallel design.

But in the parallel design, the reheater remains“dry” during startup and receives no cooling fromsteam flow. This forces HRSG designers to signifi-cantly enhance reheater tube metallurgy, andmakes the cascading bypass system the less expen-sive option.

Other advantages of the cascading bypasssystem include the following:

■ Steam flow through the reheater limits theheat input of the HP evaporator (which limits theH P drum transient) during both plant start-upand ST trip.

■ When properly designed and tuned, the cas-cading bypass system’s transient characteristicsclosely resemble the ST characteristics, whichensures the smallest discontinuities during tripand switch-over situations.

■ The steam produced is used to warm up thereheater headers and main steam lines, instead ofbeing dumped immediately after it has left the HPsystem.

A risk introduced by the cascading bypass sys-tem is its potential to cause “windage” overheatingof the HP turbine during startup and shutdown ifPCVs fail to precisely control HP and reheat pres-sure. Windage occurs when the pressure ratiothrough a turbine decreases to a point wheresteam begins to recirculate internally. This recircu-lation causes a very rapid increase in temperaturethat can easily exceed the material limits of the STblading. In certain conditions, windage can occurwithout any indication of trouble to the operator.

The worst case is a warm start during low-flowconditions. (A hot start allows faster ramp rates,

4 COMBINED CYCLE JOURNAL, Fourth Quarter 2003

SPECIAL REPORT

2. In a cascading bypass system, the HPsuperheated steam generated during startup isdiverted around the HP section of the turbine,through a pressure-control valve and an attempera-tor, and into the cold reheat line. After mixing with IPsteam and passing through the reheater, the dumpsteam is diverted around the IP and LP turbinesthrough a second pressure-control/attemperatingstation, before it is directed through a dump tubeinto the condenser

3. A risk introduced by the cascading bypass sys-tem is its potential to cause “windage” overheatingof the high-pressure steam turbine during startupand shutdown if the pressure-control valves fail toprecisely control HP and reheat pressure. Toaddress the problem, designs for some of the com-bined-cycle powerplants recently installed alsoinclude a “startup bypass,” which connects the HPturbine vent directly to the condenser, upstream of acheck valve in the cold reheat line

thus higher steam flows, which protect against theoverheating effect.)

To address this windage problem, some cascad-ing bypass systems also include a “startupbypass,” which connects the HP turbine ventdirectly to the condenser, upstream of a checkvalve in the cold reheat line (Fig 3).

Another risk with cascading-bypass systems isthe potential to motor the ST if the cold reheatcheck valves do not seat properly following an STtrip.

Tough on valves. Cascading bypass systemshandle an immense amount of energy at ele-vated temperatures, pressures, and velocities.

They also experience harsh temperature and flowchanges as PCVs and attemperation valves opensuddenly in response to plant transients. The resultis severe service for the valvesand steam-conditioning equip-ment, which has led to a host ofoperations and maintenance(O&M) problems in the field.

At HRSG User’s Group meet-ings, for example, many plantmanagers report poor reliabilityof their desuperheater sprayvalves. A common problem is thatthese valves have been over- s i z e dor utilize a “linear” trim, leadingto erosion, leak-through, and poorcontrol response. Re-trimmingwith a properly sized “constantpercentage” trim can address thisproblem. At a minimum, an effec-tive preventive-maintenance pro-gram for a cycling HRSG willinclude annual overhaul of thesespray valves, and inspection ofthe block valves, spray nozzles,and thermal liners, if installed.

Members of the HRSG User’s Group also reportgreat difficulty attaining stable, precise, reliablecontrol from their steam-bypass control valves.T h a t ’s a particularly critical problem for cyclingplants because the bypass system is expected toperform during every plant startup and shutdown.In addition to poor process control, the controlvalves in turbine-bypass systems often experiencethe following:

■ Premature trim and body failure caused by alack of control of fluid velocity along the flow path,and internal vibration.

■ Sluggish operation and sticking because ofdifferential expansion between trim and body.

■ Poor shutoff capability. Tight shutoff capabili-ty in the HP-bypass control valve must be pre-served to minimize HP pressure decay duringovernight and weekend shutdowns, and to avoidefficiency losses during normal operation. Ti g h tshutoff capability also is essential in the hotreheat bypass control valve to prevent high-tem-perature steam from leaking through, unattem-perated, into the condenser.

D o n ’t feel like the Lone Ranger if these prob-lems sound familiar. Many owner/operators whowere left with low-bid, poor-performing bypassPCVs by their architect/engineer have been com-pelled to replace these valves with higher qualityequipment. Some have even done so within theoriginal valve warranty period.

Valves need warming, too. Many control-valve problems can be attributed to thermalshock, because the valves are designed with-

out any provision for warming. Think about it: Thebypass control valves are there to warm up HRSGand ST metal, because we understand the damagethat can occur if we shock the HP drum, HP super-heater outlet header, or the turbine’s steam chest,casing, or rotor—the most vulnerable, thick-walledcomponents. Yet we often make no allowance for

the thermal shock that occursto the bypass control valvesthemselves.

During normal operationwith the ST loaded, the HPbypass system’s valve bodiescool to saturation tempera-ture—approximately 600F. Cer-t a i n l y, some amount of bypass-system warming occurs becauseof heat conduction. But in manyplant designs the HP b y p a s svalve is at the end of a longpipe run, perhaps 12 ft or more,arranged with the valve abovethe HP steam pipe and withbottom entry of the pipe. Thisarrangement precludes conduc-tion alone from providing ade-quate valve-body warming.

During both normal shut-downs and trips of the ST, the600F bypass valve bodies are


SPECIAL REPORT

5. At some plants, the maximumvolumetric flows of dump steamoccur early in a cold start, duringthe soaking period, when the GT isat minimum load. But because thearchitect/engineers based theirdesign on the condition of full GTload/full HRSG pressure, the dumpsystems end up being undersized

4. Control-valve reliability and service life can begreatly enhanced by installing a warmingconnection that, during normal combined-cycleoperation, limits the difference between steam tem-perature entering the respective bypass station andthe valve-metal temperature

suddenly subjected to high flows of steam at1 0 5 0 F, a temperature differential of around 450deg F. By comparison, the allowable temperaturedifferentials set by HRSG and ST manufacturersbetween steam and metal temperatures rangefrom 145 to 200 deg F.

S i m i l a r l y, when the plant is off-line, the bypassvalve bodies cool, all the way down to ambient tem-perature during a long shutdown. On the subse-quent startup, when the bypass valves open, thecold valve bodies experience a high flow rate of hots t e a m .

Control-valve reliability and service life can begreatly enhanced by installing a warming connec-tion that, during normal combined-cycle operation,limits the difference between steam temperatureentering the respective bypass station and thevalve-metal temperature (Fig 4). We recommend adifferential of 160 deg F before the control valvesare allowed to swing wide open.

Bypass valve warming during normal opera-tion can be effectively accomplished without lossof system efficiency or great capital cost. A s m a l l -bore pipe connected at one end to the bypassvalve just upstream of the seat and the other endconnected to the valve’s respective main or hotreheat pipe near the ST will supply sufficientsteam flow to keep the valves warm, while stilldelivering the warming steam to where it wasgoing in the first place. The additional warmingconnections still require the installation of start-up and shutdown drain lines to be installed inthe bypass piping.

Other design improvements. Valves arenot the only design weakness found in thecondenser dump systems at many com-

bined-cycle plants. To be sure, the design chal-lenges are formidable, and somewhat unique tothis plant type. Condenser dump systems havebeen operating successfully in North America formany years at nuclear plants, particularly at boil-ing-water reactor plants, where routine venting ofpotentially radioactive steam is not permitted. Buta nuclear plant’s use of saturated steam avoidsmany of the design challenges found in the super-heated, reheat steam systems of a combined-cycleplant. Similarly, many large fossil-fueled NorthAmerican steam plants have experienced years ofsuccessful condenser-dump operation. But theirbypass systems typically are sized for less thanthe full-load condition—a usual value is 30%dump capacity—so the demands on their equip-ment are less severe.

There has been substantial experience withhigh-capacity dump systems at steam plants out-side of North America. In Europe, South A f r i c a ,and Japan, for instance, the majority of steamplants installed over the past 50 years featuredBenson-type once-through boilers which required100%-capacity dump systems. Several usefullessons can be taken from that experience andapplied to today’s challenges at North A m e r i c a ncombined-cycle plants.

Design, O&M lessons to learnDump undersized for ST trip. Many NorthAmerican condenser dump systems are under-sized, which forces operators to supplement theircapacity by opening the sky vents and therebywaste valuable demineralized water. Use of skyvents also makes pressure control less precise.

When designing condenser dump systems, engi-neers typically consider the case where the STtrips from full load. Some assume, logically at firstglance, that the mass flow will be 100% of the flowthrough the fully loaded ST. But when the STtrips, the HP bypass valve opens and attemperat-ing spray is injected to desuperheat the steamfrom approximately 1050F to 600F. This attemper-ating spray adds mass flow to the fluid stream,perhaps as much as 20% more than the flowthrough the fully loaded ST.

The fluid stream then flows through thereheater and the second, hot reheat attemperationstation, which further reduces steam temperatureto approximately 350F. The hot reheat bypassspray flow adds yet more mass to the fluid stream,approximately another 20%.

In sum, the total mass flow that the condenser-dump system must be sized to handle for the caseof an ST trip is not 100% of the turbine’s full-loadsteam flow, but as much as 140%. In addition tosuitably sizing the piping, valves, steam inlets tothe condenser, and condenser-tube nest, this 140%mass-flow condition requires adequately sized con-densate pumps that can remove the larger volumeof water from the condenser.

Dump undersized for cold start, LP c o n-d i t i o n s . Another common cause of under-sizing condenser dump systems lies in the

analysis of startup transients. When evaluatingstartup scenarios, designers often assume that thehighest steam flow will occur during the latterpart of a startup, when the GT is at full load andthe HRSG is at 100% rated pressure. But a char-acteristic of GTs is that they produce a high per-centage of their full-load exhaust energy duringboth ramp-up and operating at low load—perhapsup to 70%. This means that the HRSG is capableof achieving near rated steam production, often atlower-than-design steam pressures.

During a cold start, the GT is quickly brought toand held at minimum load for heat soaking of theHRSG (Fig 5). During this soak period, steam tem-perature and pressure are low, therefore its specificvolume is very high (the steam is less dense). A tseveral plants analyzed, the maximum volumetricflows occur at these conditions—early in a coldstart when the GT is at minimum load. But becausethe designers of these plants sized the condenserdump system for the condition of full GT load/fullHRSG pressure, the operators must open the skyvents to supplement their undersized dump system.Realize that the operators must open these vents


SPECIAL REPORT

during e v e r y cold start, so the cumulative waste ofdemineralized water during a cycling plant’s 20-yrservice life is going to be enormous.

If the sky vents are not opened, then pressureramp-rate limits in the HP drum will be violated.So the Hobson’s Choice is to waste demineralizedwater, or shorten HP drum life.

Location, location. . .saturation. A n o t h e rdesign lesson that can be learned from over-seas experience concerns the location of the

bypass station, and the quality of steam that isdumped into the condenser.

The common practice overseas is to locate thehot reheat (HRH) bypass station close to—typical-ly just a few feet away from—the condenser neck.The North American practice, by contrast, hasbeen to locate the HRH bypass station closer tothe boiler (that is, more remote from the con-denser) to minimize the use of expensive P91 pip-ing. But the resulting longer piping runs of HRHdischarge piping, which could be several hundredfeet in length, necessitate either the use of largerdiameter piping or higher steam pressures at theHRH bypass to yield the correct steam pressureentering the condenser.

Closely related to the location issue, is the issueof steam quality as it is dumped into the con-denser. The industry standard in other parts of theworld is to dump saturated steam of approximately90% quality. This is to emulate the fluid that isnormally discharged into the condenser from theST exhaust. Makes sense.

In North America, however, recommended guide-lines published by the Heat Exchange Institute(HEI) and the Electric Power Research Institute(EPRI) require that steam dumped into the con-denser from a turbine bypass system must be s u p e r -h e a t e d . HEI, whose members include condensermanufacturers, is concerned about erosion that mayoccur from “wet” steam. In our experience, this con-cern is overblown, and is the cause of considerablemechanical damage and process-control problems atNorth American combined-cycle plants.

The specification of superheated steam leads toseveral problems. For starters, superheated steamhas a lower heat-transfer coefficient, so the con-d e n s e r, which was sized for 100% ST load, oftenends up with too little surface area in its cooling

tubes to handle the condenser-dump flow. Theresult, during condenser-dump operation, is a lossof condenser vacuum. In a 2-by-1 combined-cycleplant, if one power block is operating at load andthe other is dumping steam through the bypass sys-tem, there is a significant risk of tripping the ST onhigh condenser pressure (inadequate vacuum).

Another potential problem with the NorthAmerican practice is that superheated steam has ahigher specific volume, and therefore a greatervelocity than 90% quality steam, for the samemass flow rate. The tube-support spacing in con-densers is generally designed to handle the lowervelocities of 90% quality steam exiting from anoperating ST, hence the supports often are unableto prevent the condenser tubes from hammeringagainst each other when they encounter high-velocity steam from the dump system. In a mar-ginal case, damaging steam velocities may onlyoccur during cold-weather operation, when lowercondenser pressures further increase the steamvelocity into the tube nests.

When the condenser at a North American com-bined-cycle plant suffered tube failures shortly aftercommissioning, a root-cause analysis showed thattubes hammering against each other, particularlythe first four tubes closest to the dump-tube inlet,was the cause. Extra supporting elements had to beretrofit to this unit to dampen the tube movementcaused by the condenser dump system (Fig 6).

It appears that many North American designersare not even adhering to the HEI and EPRI guide-lines when it comes to the steam enthalpies thatthey are allowing to enter the condenser. HEI 5.4.2states: “Limit the enthalpy of entering steam to nomore than 1200-1225 Btu/lb except in the case ofhigh-flow steam dumps where the enthalpy shallbe limited to 1190 Btu/lb. Acceptance of flows withenthalpy higher than 1225 Btu/lb may be consid-ered depending upon specific conditions of service.”

The EPRI guidelines define “high flow” asgreater than 20,000 lb/hr. A typical F-class 2-by-1power block is configured with two dump systems of717,000 lb/hr each, one per HRSG, dumping into asingle condenser. During startups and shutdowns,dump-steam flow rates from each HRSG typicallyrange from 250,000 to 500,000 lb/hr, or more.

Cl e a r l y, these flow rates far exceed EPRI’s20,000-lb/hr threshold, thus the HEI limit of1190 Btu/lb for “high-flow” systems must

a p p l y. However many designers are allowingenthalpies of 1250 Btu/lb. They rationalize this byciting the HEI exception, which “may be consideredupon specific conditions of service.” But themechanical damage and process-control problemsbeing experienced in North American plantsstrongly suggest that this exception does not applyto F-class combined-cycle condenser dump systems.

At a minimum, then, owner/operators need toensure that their designers are adhering to the1190-Btu/lb enthalpy limit set by HEI and EPRI.To further enhance condenser-dump design, wepropose that the requirement for superheated


SPECIAL REPORT

6. NorthAmerican guide-lines, in contrastto European prac-tice, require thatsteam dumpinginto a condenser must be superheated. But thetube-support spacing in condensers generally isdesigned to handle the lower velocities of 90%quality steam exiting from an operating turbine.Tube damage, shown here on shiny area, wascaused by tubes hammering against each otherduring turbine-bypass operation

steam published in the HEI and EPRI guidelinesbe re-evaluated, and that the use of saturatedsteam, commonly practiced in other parts of theworld, be considered. Of course, effectively man-aging the effects of high-velocity water impinge-ment within the condenser remains important, butwe think this is a less troublesome problem thanmanaging highly superheated steam.

HEI published its latest edition (the ninth) of“Standards for Steam Surface Condensers” in1995, plus an addendum that included new mater-ial on turbine bypass in December 2002. The relat-ed EPRI publication is “Recommended Guidelinesfor the Admission ofHigh-Energy Fluidsto Steam SurfaceCondensers,” whichwas last updated in1982.

HRH attem-peration con-t r o l . The con-

trol scheme for thedesuperheating sprayvalve is another com-mon design weaknessat combined-cycleplants. Most controlschemes use a simplefeedback loop w i t hsteam temperature asthe single variable.This temperature-con-trol method worksfine for the HP b y p a s sstation, but not forthe HRH bypass sta-tion, where the steamis close to the saturation point. The results ofclosed-loop temperature control during transientsin HRH bypass stations are excessive hunting ofthe desuperheat-spray valve, and frequent controlexcursions. If the excursion is large enough, it cancause the condenser dump system to trip on high-temperature limit, and force the sky vents to open;or excessive spray to be emitted, leading to pipeerosion and condenser damage.

A more precise, reliable method of attempera-tion control for HRH bypass transient conditionsis a f e e d - f o r w a r d scheme that measures pressureand temperature upstream of the PCV, and pres-sure downstream of the PCV. From these para-meters and an understanding of the sparget u b e ’s pressure/flow characteristic, we calculatedump-system inlet enthalpy and dump-steamflow rate, which enables the precise determina-tion of required attemperation spray. Thisenthalpy-control method, extensively applied inEuropean, South African, and Japanese designs,has been retrofit to some North American plantsduring routine outages, and has eliminated thehunting problems that previously plagued thesef a c i l i t i e s .

Sometimes a combination of control schemes canbe used, with the temperature-control scheme han-dling steady-state conditions, while transients arehandled by the enthalpy-control which more rapid-ly positions both the PCV and attemperation valveto yield the correct enthalpy in the bypass. After ashort time delay, the valves are released back tosteady-state, closed-loop temperature control.

Sparge tube design. Several cases have sur-faced where condenser tube damage wascaused by the design of the sparge tube that

distributes dump steam into the condenser. Con-denser designerstypically assumethat dump steamwill be evenly dis-tributed across theentire area of thetube nests. Unlessthe sparge tube iscarefully designedand located, this willnot occur, resultingin higher localizedsteam velocitiesthan anticipated. Itis best if spargetubes extend the fullwidth of the con-denser and are ori-ented perpendicularto the direction ofthe condenser tubes,so that the dump-steam jets from thesparger are parallelto the condensertubes.

One style of sparge tube that can be particular-ly troublesome is represented in Fig 7. This tubehas relatively large holes arranged between the11:00 and 1:00 o’clock positions, with a “hood”positioned above these holes to deflect the steamto each side of the tube. Depending upon hole size,hole location, and hood height, this style of tubecan produce very powerful, potentially supersonicjets of steam that impinge on localized areas ofcondenser tubes, support structures, and sidewalls—causing localized heating, excessive vibra-tion, and other mechanical damage.

Noise problems. Excessive noise is a finaldesign weakness that we’ll discuss in thisarticle. Excessive noise can be generated in

the vicinity of the condenser because the con-denser casing is typically constructed of thermallyuninsulated, thin materials compared to that ofthe turbine-bypass valves and piping. For themajority of the applications, the noise is addressedusing noise attenuating trim in the turbine bypassvalve and a downstream dump tube (or sparger)inserted into the condenser.

The increasing use of air-cooled condensers fur-8 COMBINED CYCLE JOURNAL, Fourth Quarter 2003

SPECIAL REPORT

7. Condenser designers typically assume that dumpsteam will be evenly distributed across the entire area ofthe tube nests. But unless the sparge tube is care f u l l ydesigned and located, this will not occur. The result ishigher localized steam velocities that cause localized heat-ing, excessive vibration, and other mechanical damage

ther aggravates the noise problems. In an air-cooled condenser arrangement, the steam isdumped into a main steam duct that then distrib-utes the steam to finned tube banks located overthe forced-draft fans. This steam duct is a verylarge, thin-walled (typically 0.5-in.-thick wall)piece of pipe that runs along the outside of thecondenser support structure. The noise generatedin such thin-walled ducts can be enough to cause aplant to exceed noise-permit limits.

To eliminate or reduce the noise, several factorsmust be considered. First, is the noise generatedby the control valve itself? Noise is generated herebecause the majority of the system pressure dropoccurs inside the control valve. The next noise for-mation mechanism that must be accounted for isthe sparger. A smaller pressure drop occurs at thispoint, but the noise must be tightly controlled asthe flow dumps into the condenser duct. Finally, asmultiple turbine bypass valves are used in themajority of condenser-dump designs, the noisefrom each separate valve and sparger combinationmust be considered.

For one application with a direct air-cooled con-denser and multiple turbine bypass valves wherenoise from the condenser ducting and air- c o o l e dcondenser was of particular concern, we designeda solution that included four stages of multi-portpressure-reduction plates, with two stages ofattemperation, and lead-lined acoustic insulationon the bypass valves and piping.

Andrew Johnson and James McLeish of GryphonInternational Engineering Services Inc, and MikePearson of J Michael Pearson & Associates con-tributed significantly to this report.


SPECIAL REPORT

R o b e rt W Anderson is Manager of Com-bined Cycle Services—CT Operation forP ro g ress Energy Inc, Raleigh, NC. He is amechanical engineer who has managed,operated, and maintained a variety of equip-ment, systems, and powerplants for over 30years—including steam, nuclear, and gas-turbine facilities. Bob also serves as chair-man of the HRSG User’s Group, an educa-

tional organization that fosters collaboration among users,OEMs, and service providers for the advancement of HRSGdesign, operation, and maintenance.

Henk van Ballegooyen is President ofG ryphon International Engineering Serv i c e sInc (GIE), a multi-disciplinary consultingengineering firm based in St. Catharines,Ont, Canada. Services provided by GIEinclude conceptual and detail design ofsteam and power generating plants. Thiswork often includes detailed re v e r s e - e n g i-neering and testing of existing HRSG super-

heaters, drums and economizers; steam turbine/generator;condenser and cooling water systems; HP and LP bypassdump systems; steam vent systems; HP and LP piping sys-tems; fuel gas system; and water treatment systems to iden-tify and assess modifications required to allow a given plantto commence cycling duty. Prior to founding GIE in 1990,Henk had extensive powerplant experience in Euro p e ,including the design of plants for South Africa.

Astrategy aimed at maximizing profitsthrough market-driven plant operationsdemands that generation asset ownerssimultaneously balance plant optimiza-

tion against market opportunities. Though thismay sound logical and relatively straightfor-ward, the volatile nature of energy markets andthe often inflexible design of some generationassets can create competing interests within gen-erating companies (Gencos) that impact decision-m a k i n g .

Traditionally, Gencos operated under cost-mini-mization and reliability objectives. Fulfillingnative load obligations and minimizing productioncosts were the primary measures of their operat-ing strategies. Flexible generators were used toaccommodate changing demand, replace genera-tors that were forced out of ser-vice, and serve peak-load oblig-ations.

Today, Gencos must retool tooperate under a profit-maxi-mization strategy. The conver-gence of asset operations withenergy marketing and tradinghas created financial opportuni-ties for Gencos that are subjectto market fluctuations. Whilefulfilling load obligations andminimizing production costs arestill part of profit-maximiza-tion, profits from buying and selling excess gener-ation must now be added to the equation.

When economically viable, Gencos must be pre-pared to fulfill obligations from the energy marketand to shut down plants, remarket existing fuelcontracts, and curtail plants, or bring plantsonline to back up energy sales. Thus flexible gen-erators can be measured in value by their abilityto capture on-peak price spikes, replace firmpower obligations, and shut down under unprof-itable circumstances.

The purpose of the analysis that follows is toshow how market prices drive certain operationalprofiles in order to best capitalize on the opportu-

nities that exist. Its focus is on combined-cyclegeneration and how price-based commitment anddispatch, along with unit constraints, impact oper-ational characteristics.

To understand some of the dynamics that occur,we will examine forward power prices, forwardnatural gas prices, and variable operating costs fora natural-gas-fired combined-cycle unit consistingof two gas turbine/generators (GT), two heat-recovery steam generators (HRSG), and one steamturbine/generator (ST). The plant cannot operatein the simple-cycle mode but may run in a one-on-one configuration with one GT (and its associatedHRSG, together referred to as a “stage”) and theST producing power.

As shown in Fig 1, the combined cycle’s averageheat rate ranges from slightly over 8000 Btu/kWhat the lower operating limit to slightly below 7000Btu/kWh at the higher operating limit. For thisanalysis, the unit was assumed to have eight-houruptime and downtime constraints, a $3/MWh vari-able operating charge, and a $5000 startup costhurdle for each stage of the plant. The uptime con-straint means that once the plant starts operatingit must continue to do so for at least eight hours.Likewise, it must remain offline a minimum ofeight hours after each shutdown.

The energy price forecasts presented in Fig 2are modeled after actual hourly power prices andmonthly natural gas prices. In this example, on-peak power prices vary hourly and range from $35to $70/MWh in the summer (June-September) andfrom $35 to $60/MWh in all other months. Averageoff-peak prices vary hourly and average around

$30/MWh for the year.Under the assumed mar-

ket conditions and unit con-straints described above, weoptimally commit and dis-patch the unit to maximizeprofits. The optimizationwould be similar to one thatan energy manager or traderwould ask of a plant underthe assumed market condi-tions. The resultant operat-ing profile does not violate


GENERATION ECONOMICS

To maximize profit,keep one eye on themarket, the other onyour plant

By Jason Kram and Doug Logan, PCI

any of the previously men-tioned unit constraints.

The number of monthlyonline hours and number ofstarts are shown in Fig 3.On an annual basis, Gen-cos can assess whether theforecasted number of start-ups and online hours areconsistent with expectedmaintenance intervals andplanned maintenancecosts. For this example, themarket projections yieldroughly 7300 online hoursand 75 startups for theyear. By altering the mar-ket price forecast up anddown by 20% the opera-tional statistics can changeg r e a t l y. Specifically, fore-casted annual online hourscan vary from 3000 to 8300hours (Fig 4) , startupsfrom 40 to 110 (Fig 5).

Annual unit startups areof particular concern toasset owners because asthat number increases, theexpected life remaining incritical GT, HRSG, and STcomponents decreases. Howrapidly components agedepends in large measureon how conservatively thestartup process is managedand on whether the startsare hot, warm, or cold.

Plant managers typicallyare assigned the responsi-bility for calculating thecost of a startup, a numberthat is influenced signifi-cantly by the actual numberof starts the unit is expect-ed to experience in a giventime period—for example,one year. Obviously, thelower the startup cost themore often traders will beinclined to start the unit,taking advantage of marketopportunities that may beof relatively short duration.The higher the cost, the lesslikely it will be started.

It is important to not tooverlook the significance ofthe startup-cost calculation.It must include the obviousfuel and manpower costsprior to synchronization, aswell as a realistic estimatefor maintenance, repair,

and replacement of key compo-nents. The not-so-obvious compo-nent of the calculation resides inthe OEM-recommended mainte-nance costs (combustor inspec-tions, hot-gas-path inspections,and so-called major inspections).

Before finalizing the numbergiven to power marketers,the costs associated with

these maintenance inspectionsshould be built into the startupcost hurdle. The challenge in thiscalculation is that the startupcost depends on the number ofstarts a n d the number of startsdepends on the startup cost. Fur-thermore, the startup cost willvary over time and should berecalculated periodically.

Fig 6 illustrates the effect thatstartup cost has on both onlinehours and unit starts. You cansee that by increasing the start-up cost hurdle, the number ofunit startups decreases and thenumber of online hours increas-es. Important: This correlationdepends on the market priceforecast. As shown in Fig 7, anincrease in the startup cost hur-dle can reduce unit online hours.In some cases, when the startupcost is particularly high the unitwill not run at all because it can-not clear the financial hurdleimposed on it.


GENERATION ECONOMICS

KeySpan Corp’s (Brooklyn, NY) 250-MWRavenswood combined-cycle plant will bethe first major generating unit commis-sioned in New York City in more than a

decade when it begins startup testing at the end ofthe year. But it is just the first step in a buildingplan announced by Chairman andCEO Robert B Catell more than twoyears ago (see sidebar). He said,“KeySpan has a focused strategy toseek opportunities to develop gener-ating capacity in the New York metro-politan area. In so doing, we will con-tinue to be good neighbors and workclosely with local communities.”

Catell was convinced long before KeySpanclosed on the purchase from Consolidated EdisonCo of NY Inc (ConEd) of the existing 2160-MWRavenswood facility’s three steam units and peak-ing gas turbines, that additional generatingresources would be needed to help assure NewYork City’s economic growth. Mayor Michael RBloomberg echoed that need in a speech August 1when he called for the addition of 3000 MW by2008 to accommodate growing demand and replaceaging units. He encouraged the repowering andexpansion of existing facilities rather than con-struction of greenfield generating plants.

The blackout on August 14 put an exclamationmark on that plan.

Conceptual design of the combined-cycle unitbegan in the summer of1998, when KeySpan wasstill investigating the pur-chase of Ravenswood’s exist-ing generating assets,according to Howard AKosel, Jr, senior vice presi-dent, KeySpan EnergyDevelopment Corp, the busi-ness unit responsible forbuilding the company’s pow-erplants. The engineeringfirm, Burns & Roe Enter-prises Inc, Oradell, NJ, wascontracted to assist KeySpan in developing a con-

ceptual design, says Project Engineer Richard JPaccione. The assignment was challenging. Itincluded a modification of the designduring the licensing process to gofrom once-through cooling to an air-cooled condenser (ACC). The onlyspace on the site to build the com-bined cycle was a 2.4-acre parkinglot, and the only place available forthe ACC was above the plant.

The licensing process. KeySpanacquired the Ravenswood plant inmid June 1999 and filed a pre-application reportwith the NYS Board on Electric Generation andthe Environment—better known as the sitingboard—for the new unit a few weeks later. Orga-nized public opposition to the project appearedearly in 2000. Primary issues were related to airquality in northwestern Queens. KeySpan workedclosely with citizen groups to address their con-cerns. Kosel says a consistent theme of the publichearings was that “if you build a new plant youshould offset its emissions by shutting down anexisting facility or reducing pollutant dischargeselsewhere.”

Kosel challenged the company’s environmentalengineers to reduce further the emissions of nitro-gen oxides (NOx) from Ravenswood’s existingsteam generators so there would be no increase indischarges from the site when the combined-cycleunit was operating. The solution, announced inApril 2000, was the implementation of a $9-millionAir Quality Improvement Program (AQuIP) at thec i t y ’s largest powerplant despite the fact thatRavenswood was already operating well below per-mit limits.

Program involved burner modifications andinstallation of close-coupled overfire air, pluschanges on Unit 30, the well-known 1000-

MW “Big Allis” to permit full-load operation onnatural gas. The 40-year-old, 385-MW Units 10and 20 had been equipped to burn 100% gas previ-o u s l y. Bear in mind that the dual-fuel Unit 30 is


PROFILE

Richard J Paccione

Profile: KeySpan Corp’s RavenswoodCombined Cycle Powerplant,Queens, NY

How to shoehorn 250MW into a parking lot

The 170-in.-diam steamduct riser on the north sideof Ravenswood’s newcombined-cycle unitterminates in a header thatsupplies three 98-in.-diamrisers. Butterfly-type isolationvalves are installed in the twooutside risers to shut downone or two sections of theACC during periods ofreduced steam flow

Robert B Catell

required by the New York ISO (Independent Sys-tem Operator) to generate a certain percentage ofits electricity with low-sulfur oil—the exactamount depends on load—to ensure continuity ofoperation in the unlikely event that gas supply isinterrupted. NOx reduction from the upgrade pro-gram, about 750 tons/yr, is the equivalent of shut-ting down a 350-MW generating station.

The Ravenswood combined cycle also wasdesigned for cogeneration service. It can accommo-date the future implementation of an export steamsystem tied into ConEd’s Manhattan district heat-ing system and the replacement of 50-yr-old pack-aged boilers owned and operated by ConEd.

In New York State, all elements of powerplantlicensing are handled by the siting board’s Article Xprocess, explains Brian T McCabe, VP of GenerationDevelopment, including the public outreach effort. Soonce the public’s concerns were addressed by A Q u I P,KeySpan moved quicklythrough the licensingeffort. The companyreceived accolades for itse m i s s i o n s - r e d u c t i o neffort, including one fromthe Natural ResourcesDefense Council. SeniorEconomist Ashok Gupta,based in New York City,said, “KeySpan is to beapplauded for its commit-ment to integratingimprovements into theoperation of its existingfacilities at Ravenswood.These dramatic improve-ments show that energy-producing companiesreally can reduce emis-sions in response to theneeds of the environmentand the community. ”

Site preparation and construction. One ofthe first things KeySpan uncovered duringits survey work for the combined-cycle addi-

tion was that the site formerly was the location ofa manufactured gas plant and the soil was conta-minated. Such facilities were relatively commonwhen gas lighting was popular. After executing avoluntary cleanup agreement with the NYS Deptof Environmental Conservation, the companyshipped contaminated soil and excavation materi-als to a licensed remediation facility.

But that was not the last hurdle impacting sitepreparation. Underground infrastructure, bothlive and abandoned electrical cable and oil, gas,and water pipelines, crisscrossed the area and hadto be removed or relocated. This included thererouting—without disrupting service—of twohigh-pressure gas lines supplying the existing siteand relocating the fuel-oil supply line to ConEd’s74th St steam station in Manhattan. Mapping the

site was the first task and that waspainstakingly slow. It involved hand-digging a trench about 5 ft deeparound the construction area to seewhat facilities were either enteringor leaving the site.

KeySpan received approval fromthe siting board to begin constructionin September 2001 and spent thenext five months dealing with the undergroundinfrastructure. Next, says Project Manager JimMarzonie, came the installation of 300 caissons,drilled 12-15 ft into bedrock located 15-20 ft belowgrade, to support the major equipment, whichincludes the 3500-ton, roof-mounted ACC.

Construction at Ravenswood was particularlychallenging because the site offered virtually no lay-down space. To illustrate: Open space was at such apremium that there was only room for one step-up

transformer to serve boththe 18-kV gas turbine/gen-erator and the 13.8-kVsteam turbine/generatorinstead of the normal two.Even a standard open-air,138-kV electrical substa-tion wouldn’t fit, says Pro-ject Engineer Paccione,and an SF6 g a s - i n s u l a t e dsubstation, which has amuch smaller footprint,was installed instead.

E n g i n e e r / c o n s t r u c t o rStone & Webster Inc,Stoughton, Mass, a unitof Shaw Group Inc, man-aged the delivery andstorage of equipment, on-site and off. The companyarranged a just-in-timeinventory system withdeliveries by land and

water to support three days of construction activi-ty by General Contractor Slattery Skanska Inc,Whitestone, NY.

Stone & Webster stored equipment at variouslocations around the city, but its biggest supplydump was at an old navy pier facility in Bayonne,NJ, with both land and sea access. Some pre-assembly was done in Bayonne, says VP McCabe,such as the fan modules for the ACC. These, aswell as the turbines, generators, subassemblies forthe heat-recovery steam generator (HRSG), largepiping, and the ACC steam ducts, were deliveredto the plant dock on the East River by barge.

Combined-cycle design. R a v e n s w o o d ’s newunit is designed as a cogeneration facility, butinstalled as a conventional 1-by-1, two-shaft com-bined-cycle plant. It is capable of supplying up to 1million lb/hr of steam to ConEd’s extensive distribu-tion network in Manhattan, but no contract has beensigned to date and the auxiliary equipment neededto deliver that steam will not be installed now.


PROFILE

Operation of the ACC is illustrated above in func-tional diagram that does not accurately portray thearrangement of components in the system installedat Ravenswood

Brian TMcCabe

The combined-cycle unit is designed for highreliability. Only proven components were specified.The 171-MW gas turbine is the popular Model7 FA+e from GE Power Systems, Atlanta, Ga; theHRSG is a non-reheat unit built by KawasakiThermal Engineering Co, Kusatsu City, Japan,under a license from Vogt Power International,Louisville, Ky; and the steam/turbine generator isa 85-MW unit from Alstom, Midlothian, Va. Thecycle, optimized for cogeneration service, has anapproximate heat rate of 6500 Btu/kWh (based onthe fuel’s lower heating value) in combined-cycleservice as now arranged.

The supplementary-fired (duct burners fromCoen Company, Burlingame, Calif) HRSGwas sourced from Japan because of its unique

design to accommodate cogeneration service. It hassignificantly more surface area than the standardboilers for cookie-cutter type combined-cycleplants that filled domestic shops at the timeof order. The triple-pressure RavenswoodHRSG produces 1435-psig/1000F at the high-pressure (h-p) superheater outlet, 155-psigsteam at the intermediate-pressure super-heater outlet, and some 20-psig steam. Theselective catalytic reduction (SCR) systemprovided with the HRSG is designed to limitN Ox emissions to 2 ppm with 5 ppm ammonia slip.Aqueous ammonia is the reagent.

Ravenswood’s 20-in. h-p steam line has a 12-in.takeoff, currently blanked-off, to supply the kettleboilers (not installed) that would produce steamfor the ConEd main. Shell-and-tube kettle boilerswork much like the steam generators supplied fora conventional nuclear powerplant with a pressur-ized-water reactor. High-quality, h-p steam fromthe HRSG would be the heating medium to boiltreated city water for the steam distribution sys-tem. Since there is no return system on ConEd’ssteam distribution grid, the kettle boilers avoidthe expense of using demineralized water to pro-duce export steam.

Demineralized water is used only for normalHRSG makeup and for water injection into thegas-turbine combustor for NO x control whenkerosene is burned. The kettle boiler concept alsoavoided problems associated with ConEd’s chem-istry specifications for export steam which disal-low amine compounds. These specs could not havebeen met easily had an extraction steam systembeen selected.

One final point: When operating in cogenerationservice with maximum steam flow of 1 millionlb/hr to the district heating system, electrical out-put from the steam turbine/generator drops toabout 10 MW.

Control system. The plant’s digital controlsystem, like the HRSG, was of a custom design toaccommodate both the combined-cycle and cogen-eration configurations, says McCabe. The DCSselected, from Emerson Process Management’sPower & Water Solutions division, Pittsburgh, Pa,features the manufacturer’s recently enhanced

software platform for its Ovation expert controlsystem. Ovation is a key component of Emerson’sPlantWeb digital plant architecture for the powergeneration industry.

Emerson’s project manager for the Ravenswoodproject, Rick Marchionda, says that the systemcontrols process variables as required for safe, effi-cient, and reliable operation of the plant, its sys-tems, and individual components. Ovation isdesigned to safely bring the plant from cold start-up to the desired operating condition and thenback to cold shutdown. It operates on a Fast Eth-ernet network, relying on seven human/machineinterfaces (HMIs)—one engineer/database server,four operator stations, an asset-management-sys-tem station, and a data historian.

Four redundant controllers located in theRavenswood control room interface with remote I/Ocabinets located in the ACC electrical building,

HRSG, compressor electrical room, andswitchyard control house. Use of remote I/Osaved KeySpan a considerable amount ofmoney by eliminating field wiring from theselocations back to the control room.

Operator workstation graphics and fea-tures built into the historian and asset man-agement system enable the DCS to monitor,d i s p l a y, and record process data received

from hundreds of field sensors and communica-tions links. This information, says Marchionda, isused for general process supervision, performancecalculations, and record-keeping, includingsequence-of-events recording and diagnostics tofacilitate plant management and maintenancedecisions.

One of Ovation’s benefits is that its connectivityfeatures give operators a fully integrated consolewith the same look and feel as the standalone con-trol systems provided by the major equipmentmanufacturers. For example, the Ethernet inter-face to GE’s Mark VI gas-turbine control system isseamless, allowing operators access to all GT infor-mation on their Ovation workstations in real time.

The system’s ability to continually track perfor-mance is a particularly valuable management toolin an area like New York City where the cost offuel is higher than in most areas of the nation.Performance calculations are run periodically forthe gas turbine, HRSG, steam turbine, ACC, andintegrated plant to determine the following:

■ Equipment efficiency as it relates to equip-ment age, maintenance procedures, and systemupgrades.

■ Actual equipment performance versus guar-anteed performance.

■ Impact of changing conditions and/or methodsof operation.

A i r-cooled condenser. The most distinguish-ing feature of the Ravenswood combined cycle is itsBalcke-Durr ACC, manufactured by Marley Cool-ing Technology Inc, Overland Park, Kan, which ismounted on the roof of the totally enclosed, space-challenged site. The building’s large structuralmembers—including 10.5-ft-deep roof girders—to


PROFILE

Jim Marzonie

accommodate the A C C ’s 7-million-lb operatingweight, give Ravenswood the fortress-like appear-ance of powerplants built a half-century ago.

The rooftop location dictates that exhaust fromthe steam turbine be routed to a 14-ft-diam riserattached to the outside of the plant’s north wall.About three-quarters of the way to the roof, theriser connects to a header that distributes steamto the three 8-ft-diam risers serving individualsections of the ACC (photo).

Arrangement of the ACC is this way: Three A-shaped condensing rows operating in paral-lel each have six modules of the type shown

in the drawing. Each module, in turn, is served byeight finned-tube bundle assemblies and one vari-able-speed fan. Air is drawn in an upward direc-tion through louvers on the supporting structureand directed through the tube bundles. Theexhaust steam header at the top of the A - f r a m edistributes steam to the tube bundles and the con-densate produced drains to 2-ft-diam manifolds atthe bottom of the legs.

Note that five modules per row have parallel-flow tube bundles in which the steam and conden-sate flow downward. The sixth, called a reflux mod-ule, features a counter-flow arrangement whereuncondensed steam bubbles trapped in the collec-tion headers are vented upward while the conden-sate produced flows in the opposite direction.Reflux tube bundles increase condensing efficiency.

The main riser delivering steam to the ACC isequipped with spargers for receiving high-, inter-mediate, and low-pressure steam during turbinebypass at startup. The bypass system, which alsoincludes pressure-reduction/desuperheating sta-tions within the plant proper, was provided byControl Components Inc, Rancho Santa Margari-ta, Calif. CCI’s equipment was specified, says Pro-ject Engineer Paccione, because its so-called Dragtechnology eliminates problems of noise, erosion,

and vibration associated with some other types ofdump systems.

Noise, of course, is of major concern to owners ofpowerplants close to load centers. The CCI systemis designed to meet Ravenswood’s 55-dBA require-ment 1000-ft from the plant, thereby permittingnighttime operation of the bypass system. At otherACC-equipped plants, it is not unusual to findnoise levels of 110 dBA near the condenser ductingand 70 dBA three-quarters of a mile from the facil-i t y. Noise normally associated with the operationof ACCs is muted by the selection of variable-speed fan drives and the selection of speciallydesigned “quiet” fan blades.

I n t e r c o n n e c t i o n . During plant construction, adecision was made by KeySpan to petition the sitingboard and the New York ISO to change the electricalinterconnection point from ConEd’s 345-kV Raineysubstation north of the site to the 138-kV Ve r n o nsubstation on the south side. An order grantingamendment of the Certificate of EnvironmentalCompatibility and Public Need was issued in July2002 and an amended System Impact and Reliabili-ty Study was submitted by KeySpan and approvedby both the ISO and ConEd. The SIRS demonstrat-ed that fault-current impacts were less connectingto the 138-kV system than to the 345-kV system.

Engineers were time-challenged to redesignequipment—such as the main and auxiliary trans-formers, gas insulated substation, and the dielec-tric cable—to accommodate this change. Modifica-tions approved by ConEd included addition of anew breaker position and associated relay protec-tion house, as well as a fiberoptic communicationslink to the Vernon substation. Plus, an existingbreaker in the ring bus had to be replaced and ele-vated. Substation modifications were performed bycontractors under the direction of KeySpan per-sonnel who worked in close cooperation withConEd. The 138-kV cable was energized in Sep-tember 2003. CCJ


PROFILE

KeySpan, ANP form venture to build powerplants

KeySpan Corp, Brooklyn, NY, and AmericanNational Power Inc, Marlborough, Mass,announced in early September the formation

of a joint venture to build powerplants on LongIsland—this in response to the Long Island PowerAuthority’s (LIPA) RFP for new generating capacityissued last spring. Long-term power-purchaseagreements would be part of the deal.

The venture combines KeySpan’s proposed 250-MW Spagnoli Road project and ANP’s pro p o s e d250-MW Brookhaven facility, both of which havebeen approved by the NYS Board on Electric Gen-eration and the Environment—the so-called sitingb o a rd. ANP received certification for a 540-MWplant in Brookhaven in August 2002, KeySpan was

c l e a red to build its facility in Melville last May. Current plans are for Spagnoli Road to begin

commercial operation in 2006, the unit atBrookhaven a year later. A second unit atBrookhaven could be built for service in 2008 ifmarket conditions warrant.

KeySpan, the largest distributor of natural gas inthe Northeast with nearly 2.5 million customers, isalso the largest investor-owned electric generatorin New York State. In addition, the company isunder contract to LIPA to operate its electricsystem, which serves more than a millioncustomers. ANP, a subsidiary of UK-based Interna-tional Power plc, has a portfolio of generatingassets in excess of 4000 MW.

Executives responsible for the purchase offuel for gas-turbine-based powerplantsneed to know what the price of natural gaswill be tomorrow, next month, next year—

and why prices are moving in a given direction.Generating company profits depend on their abilityto understand the fundamentals of natural gas sup-ply and demand and to read accurately the “tealeaves” on market intangibles, such as “perception.”

Natural gas supply and demand are influencedby many factors, including these:

■ Timing effects of natural gas exploration andproduction. For example, when gas prices fall,exploration and production activity is reducedthereby paving the way for shortages and higherprices in the future, when demand increases.

■ The business climate—that is, increases ordecreases in commercial and industrial manufac-turing activities requiring gas as a fuel or feed-stock.

■ Weather. One impact of a colder-than-normalwinter would be to increase gas flow to storagecaverns during the refill period from April 1 toOctober 1, increasing both spot and longer- t e r mprices. A warm summer means increased gas con-sumption to support higher electrical demand,increasing spot prices.

■ Oil prices. An increase in oil prices tends topush gas prices higher because the two fuels com-pete with each other in most markets and gasprices can be raised without adversely impactingmarket share.

While the cause and effect relationship of sup-ply and demand fundamentals can be monitoredand also forecasted in the short term, market psy-che cannot. For instance, gas storage injectionsran about 1 billion ft3/day (Bcfd in gas-industryparlance) above “nominal” in July and August, andpossibly will continue to do so through September.H o w e v e r, if storage injections decline to “normal”fill levels, the market could perceive a short-termtightening in North American gas supply thusputting upward pressure on the price of natural-gas futures. Likewise, flat growth in US gas pro-duction in third quarter 2003 could generate con-cerns relative to future gas deliverability after thepreliminary third-quarter producer reportsbecome available. This also would work to drivenatural gas prices up at the beginning of the heat-ing season.

Demonstrating the influence of the market’s per-ception of where the price for natural gas shouldbe, the New York Mercantile Exchange (Nymex)natural-gas futures (from August 2003) are higherthan Platts Research & Consulting’s (PR&C) priceforecast until winter 2005 (Fig 1). PR&C’s view isthat the Nymex futures are out of synch with sup-ply/demand fundamentals. However, Platts expectsthe dynamics of the tangible fundamentals to ratio-nalize market behavior over the long term.

New gas. To d a y, enormous amounts of naturalgas sit untapped in the ground while traditional gasbasins in the US are reaching maturity or indecline. The increased supply necessary to meetgrowing demand will likely come from one or moreof the following:

■ Deep water sources in the Gulf of Mexico,■ Rocky Mountains,■ Eastern Canada (Sable Island),■ Western Canada,

■ Imported liquefied natural gas (LNG), and/or■ The Arctic region.The timing and competitive economics of each of

these incremental sources is at the core of the uncer-tainty in today’s gas markets. However, two sourcesstand out as having the potential to significantlyimpact the price on natural gas for years to come:LNG and the proposed Alaskan pipeline projects.

L N G. The high value the US places on naturalgas may provide the financial wherewithal to drivea substantial expansion of the international mar-ket for LNG, where gas is liquefied at remote gasfields and transported to consumers by ship. Thismay result both in a reduction in the cost of LNGexport and import facilities and the developmentof natural gas as a true commodity. The resultwould be a worldwide liquid and competitive LNGmarket, which should help mitigate fuel risk andfoster the development of natural-gas-fired power-plants worldwide.

The delivered cost of LNG is the sum of the well-head price, the cost of liquefaction, shipping charges,and regasification cost. For shipping distances lessthan 6000 miles, liquefaction is the largest componentin the LNG cost equation; beyond 6000 miles, ship-ping charges begin to exceed the liquefaction cost.Since the late 1980s, the liquefaction capital costs


NATURAL GAS

Predicting gas pricesmore art than science

By Ken Walsh, Consultant

declined from about $300–$400/ton per year to $200.Economies of scale and improved technology couldreduce costs to as little as $150 in the near future.

Alaskan gas. Three potential pipeline projectsbeing investigated would deliver natural gas fromthe Arctic supply region to exist-ing pipelines in Alberta, Cana-da. They are:

■ The Alaskan Natural GasTransportation System or “Alas-ka Highway” pipeline, 2 to 5Bcfd.

■ The Dempster Lateral of theBeaufort Sea and MackenzieValley pipeline, approximately 4B c f d .

■ The Mackenzie Va l l e ypipeline, 1.2 to 2 Bcfd.

Regardless of which pipeline is actually con-structed, natural gas flowing from the Arctic regionwill flood the North American natural gas marketand thus have a dramatic effect on gas prices. Fig 2forecasts the Henry Hub price based on an antici-pated 2005 build-out of the LNG market infrastruc-ture (Scenario 1) and with potential production vol-ume flowing to the US market through whicheverAlaskan pipeline is built (Scenario 2).

Although the region has become a political hotpotato, the National Petroleum Reserve—Alaska,the Arctic Outer Continental Shelf, and the A r c t i c

National Wildlife Refuge (ANWR)—are areas withtremendous potential for oil and gas development.Energy legislation introduced by the Republicanswould allow the development of the ANWR areaand therefore potentially bring additional natural

gas supplies through at least one ofthe proposed or existing A l a s k a npipelines to the lower 48 states.Again, it is apparent from Fig 2that gas supply deliveries from theArctic region will have a dramaticeffect on gas price futures.

In summary, if you want toknow where natural gas prices aregoing to go in the short-term andlong-term and why, keep and eyeon the factors that drive the

demand for and supply of natural gas. However,there are forces in the market that like to spookthe price of commodities such as natural gas fortheir own short-term benefit. These are the “mar-ket perceptions” or psychological mind games thatcause some of the volatility. But there are themore supply/demand fundamental or tangiblepieces of information that provide the underlyingprice support in the market, which requires con-stant vigilance. Market fundamentals should helpdetermine the overall direction of natural gasprices over the short-term and long-term depend-ing on the planning horizon. CCJ


NATURAL GAS

The group's annual meeting will address topics such as:

• Design aspects of the new breed of merchant plants • Construction techniques for new generation • Start up and commissioning issues • Operational considerations to maximize the re t u rn on investment while providing reliable power • Staffing ideas to keep your valued and best employees • Maximizing efficiency and output to respond to the day's demand re q u i rements • HAPS - Impact on Design and Operation • Considerations of cycling base load designed CCGT Plants • Start up Emission impacts • Gray market equipment • Topics from the Floor

The Combined Cycle Users Group is developing a data base of power plants, a library of articles and a chat room wheremembers can exchange experiences. Membership is free of charge until next April and is open to power plant owner/operators, engineering companies, OEMs and other relevant stakeholders.

The Combined Cycle User Group has been formed in cooperation with the ASMEPower Division Combined Cycles Committee, the ELECTRIC POWER Conference andother industry groups. It addresses issues concern i n g the interaction between all majorcomponents of the combined cycle power plant - the gas turbine, HRSG and steamturbine. These issues are critical at a time when many plants must operate on razorthin margins and in modes for which they were not designed.

Join today on the web site www.combinedcycleusers.org or contact:

Documents

Steam Turbine Bypass Systems