Download - Return on Investment Analysis T - CONCO Systems · 2011-06-10 · behavior of steam surface condensers and their impact on your plant’s bottom line. ... (HEI(4)) and uses the “cleanliness

W H I T E P A P E R

CONCO SYSTEMS, INC. • 530 Jones Street • Verona, PA 15147 USA • 1-800-345-3476

Return on Investment Analysis:The Economics of RegularCondenser Maintenance

• Mechanical Tube Cleaning• Air Inleakage Detection

Richard E. PutmanTechnical DirectorConco Consulting Corp.Verona, PA 15147 USA

Absolutely the Best.

Getting To The Bottom LineYou may already suspect that your plantis not operating at peak performance.You want to incorporate the latest tech-nologies to improve efficiency becauseyour “gut feeling“ tells you that improv-ing your systems will have a payoff, yetyou need a concrete process to justify aninvestment.

This White Paper summarizes a decadeof research by Conco Systems, Inc. toobtain a better understanding of thebehavior of steam surface condensersand their impact on your plant’s bottomline. This research was followed by thedevelopment of techniques to model condenser performance, with a view to converting deviations from proper behavior to the equivalent avoided costs. Mathematical techniques were also developed to ascertain the optimumfrequency of condenser cleaning to minimize these avoided costs. The fulldetails of this work are published in abook by ASME Press, New York(1)

entitled, Steam Surface Condensers:Basic Principles, Performance Monitoring and Maintenance.

Minimizing turbogenerator unit heat rateand maximizing generation capacity areamong the guiding principles in powerplant operation strategy. In fossil plants,heat rate is greatly affected by the efficiency with which the fuel is burnt.However, the heat rate of fossil plants,along with nuclear and combined cycleplants, may also be affected by the stateof the condenser. Its condition is also afactor in determining whether the designMW load can be achieved.

Chapter 8 of Putman(1) outlines the varietyof operating problems to which steamsurface condensers are prone, principalamong these being the fouling of con-denser tubes with its negative impact onheat transfer rate and turbine back pres-sure. The condenser condition can alsoaffect generation capacity. Air ingressinto the turbine/condenser subsystem has similarly negative effects.

In this White Paper, we will discuss thethermodynamic considerations that makethe condition of the condenser such animportant element in determining the performance of the turbogenerator. Wewill also outline condenser maintenancestrategies that can be adopted to mini-mize annual unit operating costs and,ultimately, maximize plant revenue.

Condenser operation also has a directimpact on power plant emissions. Theheat rejected by the condenser is asource of thermal pollution; however, if it can be reduced by improved con-denser maintenance, then the emission of carbon dioxide and other atmosphericpollutants is also reduced due to thelower consumption of fuel.

Another important condenser mainte-nance problem is the inleakage of cooling water into the condensate,contaminating it chemically (EPRI(2)).However, since it principally affects thecorrosion of boilers and heat exchangers— and since the main topic of this document is the economics and return on investment as a result of the effect of the condenser on heat rate and/orpower generation, we will not furtheraddress water inleakage here.

1


The Condenser/Turbine SubsystemAll nuclear and fossil plants are basedon some variation of the Rankine Cycle— the steam/water conditions throughthe low-pressure stage of the turbine for a fossil-fired plant (illustrated in Figure 1).

Here, expansion lines are shown for anumber of loads, steam from the outlet of the reheater passing through the intermediate and low pressure stages of the turbine and expanding finally into the condenser.

During the last stage of expansion, thevapor passes through the saturation lineso that the vapor entering the condenseris usually wet. As it becomes condensed,the conditions follow the constant pres-sure line (corresponding to the condenserback pressure) down to the liquid line(not shown). The wetness increases untilthe vapor becomes fully condensed.

The enthalpy of the vapor entering thecondenser depends on condenser backpressure. It should be understood that themagnitude of the enthalpy drop betweenthe reheater outlet and the condenserinlet determines the amount of energyconverted to mechanical work (in thiscase, the generation of power).

If the flow of steam is determined at thesteam generator or boiler(s), as occurs innuclear power or combined cycle unitsoperating in the boiler-follow mode, thenraising the back pressure will reduce theamount of power generated by that fixedamount of steam. This will negativelyaffect unit heat rate and, in some cases,will limit the amount of power that canbe generated. Conversely, a drop inback pressure will increase the amount of generated power.

In fossil-fired units, in which the load isdetermined by the setting of the turbinegovernor, an increase in back pressurewill result in more steam having to begenerated to support the load set on thegovernor. This, again, has a negativeeffect on heat rate. In severe cases, suchas the back pressure reaching the upperlimit recommended by the turbine manu-facturer, the power setting will have to be reduced.

The back pressure determines the amountof latent heat (also known as condenserduty) that has to be removed for thevapor to become condensed, recovered,and then recycled back to the boiler.Meanwhile, in a clean condenser, theback pressure is determined by the duty,and by the cooling water flow rate andits inlet temperature.

Figure 1

2


Figure 1 illustrates a design back pressure on which all the appropriateoriginal cycle calculations were based;but the actual back pressure experiencedcan be lower or higher than this. Forinstance, it can be lower if the coolingwater inlet temperature is low and thewater flow is at or higher than its designvalue.

If the tubes in the condenser becomefouled, then the thermal resistance causesthe effective heat transfer coefficient ofthe tubes to decrease to the point wherethe back pressure will rise for the sameduty, cooling water flow rate, and inlettemperature.

Similarly, air ingress increases the thermal resistance of the condensate film on the outside of the tubes, decreasingthe effective tube heat transfer coefficientand with the same effect on back pressure. The proper management oftube fouling and air inleakage has important consequences on condenserand unit operating costs, as well as onthe scheduling decisions made by themaintenance department.

Monitoring CondenserPerformanceThe basic principles used to calculate theperformance of a condenser based onboth design and current operating dataare detailed in Putman.(3) One methodhas been developed by the HeatExchange Institute (HEI(4)) and uses the“cleanliness factor“ as the performancecriterion.

The other method is that originally proposed in ASME PTC 12.2-1983 andupdated in the 1998 revision of that

standard(5). Based upon the sum of ther-mal resistances, the ASME method calcu-lates the clean single-tube heat transfercoefficient as a function of current operat-ing conditions and compares this withthe effective heat transfer coefficient cal-culated from present condenser duty.

One of the uncertainties in condenserperformance monitoring is the actualvalue of the cooling water flow rate. The design value can be affected by the aging of the pumps or by the effects of both tube and tube sheet fouling.Putman(6) shows that a reliable estimate of the actual cooling water flow rate can be obtained from the cooling watertemperature rise, together with the condenser duty calculated from the turbine thermal kit data.

The conditions that could be expected ifthe condenser were cleaned requires thesolution of a set of nonlinear equationsappropriate to the specific condenserconfiguration and constructing a Newton-Raphson matrix that reflects the details.

The cost that could be avoided if the condenser were to be cleaned can beestimated by using this simple equation:

Note that this avoided cost does not distinguish between the back pressurethat results from fouling or from that due to any air ingress that may be present.An estimate of the severity of air ingresscan be obtained by calculating the totalchange in heat transfer coefficient due toboth fouling and air inleakage. This canbe obtained by subtracting the present

3


Condenser dutywhen fouled

The condenserduty when clean

Fuelcost— x

single-tube U-coefficient value (Ueff) fromthat calculated if the condenser werecleaned (Uclean).

If this difference is small, all of the esti-mated avoided cost can be attributed tofouling and maintenance plans can bemade accordingly. However, if the differ-ence is substantial, steps should be takento determine the severity of the air inleak-age so the proper maintenance decisioncan be made.

If Rf = thermal resistance due to

foulingRa = thermal resistance due to

air ingress

Then

Clearly, if Rf, Ueff and Uclean are

known, then Ra can be calculated.

Method For Estimating Tube Fouling ResistanceThe principle of the method developed byBridger Scientific under EPRI sponsorshipfor estimating tube fouling resistance Rf isdescribed in EPRI(7) and illustrated inFigure 2 (where tube pairs are used).

One of the pair is a tube with blanked-offends through which no water flows. The other, the fouled tube, not only hassensitive temperature measuring devicesat both ends of the tube, but is also provided with a turbine-type flow meterfor accurate measurement of the waterflow rate through the tube.

The blanked-off tube is used to measurethe mean shell temperature in the vicinityof the fouled tube so that any vapor pressure loss through the tube bundlescan be accommodated. Several pairs oftubes are placed strategically in differentlocations within the tube bundle(s).

From the data obtained, the mean foulingresistance Rf can be estimated using themethod outlined in the EPRI(7) report. The value of Ra can then be calculatedusing equation (1) to the left.

(1)

Figure 2

4


Types Of Fouling DepositA tendency toward condenser tube fouling is a constant problem. Tube corro-sion is also a concern and is sometimesinitiated by fouling, as in under-depositcorrosion. The maintenance of unit performance, availability and lifeexpectancy depends on a multi-levelinteraction between tube metallurgy,water chemistry, and the fouling mechanisms involved, together with the thermodynamics and kinetics of thecondenser system.

In cleaning condensers over a long peri-od of time, at least 1,000 different typesof fouling deposits have been encoun-tered. Each plant has its own idiosyn-crasies and it is not unusual to find eachunit within a plant behaving differently.

Furthermore, different tube materials inthe same condenser also foul in differentways. For example, copper alloys in themain tube bundles foul differently thanthe stainless steel tubes in the air removalsection.

As such, fouling deposits typically fallinto five major categories:

• Sedimentary or silt formation• Particulate fouling• Deposits of organic or

inorganic salts• Microbiological fouling• Macrobiological fouling

Sedimentary DepositsSedimentary fouling occurs when particulates entrained in the coolingwater deposit on the tube walls arelower than the design value due to thewater velocity. The remedy is to raise thewater velocity closer to the design value.

Particulate FoulingFor particulate fouling to occur, theremust already be a deposit substrate to which particulates become attached. A common substrate is bacterial slime.Once the particulates become attached,the deposit layer accumulates rapidly.

Salt DepositionSalts of various kinds can becomedeposited due to changes in their concentrations. This is largely a result of the increase in water temperature as it passes through the tubes. Examples of crystalline fouling include calcium carbonate, silica, calcium sulfate, manganese, and calcium phosphate.Many of these types of fouling are difficult to remove once they havebecome deposited. Some salts alsoexhibit an inverse solubility. The solubilitydecreases while the temperature increas-es. This is especially true of calcium carbonate, the deposits of which oftenincrease toward the tube outlets.

Microbiological FoulingAll sources of power plant cooling watercontain bacteria, many of which exude a gelatinous substance that allows thebacteria to attach themselves to the tubewalls. Because these slimes contain some90 percent water, they have a high resist-ance to heat transfer. The organisms areoften rich in metal compounds, such asthose of manganese, which produce an impermeable layer and can causeunder-deposit corrosion, even with stain-less steels. Microbiological fouling is alsosensitive to both velocity and tempera-ture. The deposits tend to increase inthickness as the velocity falls.

5


Macrobilogical FoulingMacrobiological fouling occurs whenmollusks, zebra mussels or similar aquat-ic organisms pass through the screensand attach themselves either to the tubesheet or to the insides of the tube walls.The reduction in water velocity not onlyincreases the thermal resistance to heattransfer, but can also make the tubes susceptible to other types of fouling. Therate at which these fouling mechanismsdevelop is site-specific and depends onthe chemical treatment methods that areemployed. Aggressive cleaning of tubesand tubesheets is necessary to controlmacrofouling.

Optimizing Condenser CleaningFrequencyChapter 7 of Putman(1) discusses the technique for calculating the optimumcleaning frequency necessary to minimize the avoided costs incurred byfouling. The site-specific data requiredincludes knowledge of the fouling modelexperienced by that particular unit, aswell as understanding the historic monthly load profile and the coolingwater inlet temperature profile for the unit over the past 12 months.

Some typical fouling models are illustrated in Figure 3. The mathematicalrelationship between fouling resistanceand time must be generated. A typicalload and inlet temperature profile curveis shown in Figure 4.

Figure 3

Figure 4

6


Condenser duty is principally a function of generator load, cooling water inlet temperature, and its flow rate. It has beenfound that, at a given load, the avoidedlosses are a linear function of the foulingresistance vs. inlet water temperature, anda different linear function of the foulingresistance vs. load.

Given the load and inlet temperature profiles of Figure 4, and a fouling modelsimilar to one of those shown in Figure 3,it is possible to calculate the cost of lossesvs. fouling factor for each of the months in that profile. The initial data table for atypical case is shown in Table I.

An optimization program can be used totest the savings to be obtained with onlyone cleaning during the course of theyear, and even establishes the month inwhich it should occur. In then calculates

the savings if there are to be two clean-ings during the year, and establishes themonths in which they should occur. Theprocedure is then repeated for three andfour cleanings per year.

Return on InvestmentThe results of a typical case are summa-rized in Table II. Note that additionalcleanings always improve performance;but the amount of increase per cleaningdecreases.

The optimum frequency selected is thatwhich causes the incremental return oninvestment to match a threshold criterion(for example, an ROI of 2). In Table II, theincremental cost of cleaning is shown tobe $13,500, while the correspondingincremental savings (or return on cleaning)are shown in Column 5. The maximum frequency corresponding to an ROI = 2 is

Table 1 — Basic Data

431.41419.38433.13448.66385.47391.87365.69318.78307.65359.23430.85443.44

50.9750.0958.9762.9568.8074.6183.6084.4881.9174.7458.4851.84

1740.281691.221764.121835.211598.131638.881560.221373.711321.401508.631753.991789.46

1767.431716.961795.091869.921629.561674.791599.721407.651352.101540.791784.491818.04

27.1525.7430.9634.7131.4335.9139.5133.9430.7032.1630.5028.58

312831303130313130313031

2019.961729.732303.422499.122338.392585.522939.542525.142210.402392.702196.002126.35

2.5000E-052.2000E-045.0000E-047.5000E-049.0000E-049.8000E-041.0400E-031.0400E-031.0400E-031.0400E-031.0400E-031.0400E-03

580.744,376.21

13,244.6921,554.9124,202.3629,138.8135,156.9530,200.6326,436.3828,616.7426,264.1625,431.17

MW

CWtemp,

inDuty,clean

Duty,fouled

Foulinglosses

Daysin

monthSpecific

costFouling

resistance

Costwithoutcleaning

Yearly cost of losses — no cleanings $265,203.74

7


shown to be three cleanings per year,since four cleanings per year would offeran ROI of <2.

Mechanical Methods For Removing FoulingSpring-loaded carbon steel cleaners havelong been recognized as the most effec-tive tube cleaner available to the indus-try. They come in various forms designedfor multi-deposit removal, hard depositremoval, or thin and tenacious depositremoval. Since the blades of these clean-ers are spring-loaded with a determinedforce at the surface of contact, they canbe engineered to perform the intendedfunction while also improving the smooth-ness of the tube surface to restore thetube to its “as new“ condition.

The earliest mechanical tube cleanerswere designed for cleaning boiler tubes.It was not until 1923 that two brothers,Cecil M. Griffin and Vivian Griffin,invented the first cleaner for condensersand shell-and-tube heat exchangers.Figure 5(a), C4S, shows a current version of this cleaner, which consist of several U-shaped tempered steel stripsarranged to form pairs of spring-loadedblades. These strips are mounted on aspindle. At one end of the spindle is aserrated rubber or plastic disk (devel-oped by George Saxon), which allows a

jet of water, delivered by a pump operat-ing at 300 psig, to propel the cleanersthrough a tube with greater hydraulic effi-ciency. Figure 6(a), Model 200B PumpSystem, shows a typical pump, mountedon a portable skid, complete with thehandheld-triggered device (Figure 6(b),also known as a water gun), which permits the operator to direct the waterto the tube being cleaned.

No CleaningsOne CleaningTwo CleaningsThree CleaningsFour Cleanings

265,204180,469130,214

92,13369,522

013,50027,00040,50054,000

265,204193,969157,214132,633123,522

071,235

107,990132,571141,682

084,73550,25538,08122,611

1Cost oflosses

2Cost of

cleaning

3Total cost of fouling

4Annualsavings

5Return oncleaning

Table II — Return on Investment Analysis: Mean Loads

Figure 5(a) C4S

Figure 5(b) Hex or C2X/C3X

8


Since the earlier cleaner designsbehaved as stiff springs, loading thecleaners into the tubes was sometimesrather tedious. To speed up this opera-tion, while also providing the blades with more circumferential coverage of the tube surface, the hexagonal or Hexcleaner shown in Figure 5(b) was devel-oped by Saxon and Krysicki. This designnot only reduced the cleaning time for1,000 tubes, but was also found to bemore efficient in removing tenaciousdeposits like those consisting of variousforms of manganese.

A later development by Gregory Saxoninvolved a tool for removing hard calcite deposits, which were found to bedifficult to remove even by acid cleaning.This tool is also shown in Figure 5(c),CB, and consists of a Teflon body mount-ed by a number of rotary cutters placedat different angles, and provided with a plastic disk similar to those used to propel other cleaners through tubes. As part of a case study, cleaners of thistype used on a condenser removed 80tons of calcite material. The tool has nowbecome standard whenever hard andbrittle deposits are encountered.

Mechanical cleaners offer the most effec-tive off-line tube cleaning method. Strongenough to remove hard deposits, theycan cut the peaks surrounding pits andflush out the residue at the same time,

thus retarding underdeposit corrosion.Each cleaning tool is custom-made to fit snugly inside a tube having a stateddiameter and gage or wall thickness.This allows the blade contact pressure to be controlled within tolerances andalso ensures that as much of the tube circumference as possible is covered asthe cleaner passes down a tube of thecorresponding size.

A special rig has also been developedfor the heat transfer testing of tubes. In the development of these variouscleaners, extensive use of this rig wasmade to determine the effectiveness ofdifferent cleaner designs. Not only couldthe incremental material removed beestablished for each successive pass ofthe tool, but the cleaner’s effect on heattransfer could also be quantified for agiven kind of deposit.

In some plants, it has been commonpractice to use compressed air withwater to propel the cleaners. However,the rapid expansion of the air causes the cleaner to become a dangerous pro-jectile when the cleaner exits the tube.The use of water alone never allows the cleaner exit velocity to rise above asafe level. Similarly, water supplied at a pressure of 300 psig is much safer topersonnel than at up to 10,000 psig,which is sometimes attained when usinghigh-pressure hydrolasing techniques.

Figure 5(c) CB

Figure 5(d) Brush

9


Another advantage of mechanicallycleaning condenser tubes using water asthe cleaner propellant is that the materialremoved can be collected in a plasticcontainer for later drying and weighingto establish the deposit density (in gramsper square foot). In many cases, X-ray fluorescent analysis of the deposit cake isperformed. Experience has also shownthat properly designed cleaners will notbecome stuck inside tubes unless the tubeis damaged or severely obstructed.

Mechanical cleaners travel through thetubes at a velocity of 10 to 20 ft/s andare propelled by water delivered at 300psig. Some of the members of a well-known family of tube cleaners have thefollowing features:

• C4S. The C4S cleaner is general-purpose and may be used to remove all types of obstructions, deposit corrosion, and pitting.

• C3S. The C3S cleaner is designed for heavy duty use and is very effective in removing all kinds of tenacious deposits. Its reinforced construction also allows it to removehard deposits, corrosion deposits,and obstructions.

• C2X, C3X. These types of cleaner consist of two or three hexagonal-bladed cleaner elements, having six arcs of contact per blade. They are effective on all types of deposits, but are especially suitable for removing the thin tenacious deposits of iron, manganese, or silica found on stainless steel, titanium, or copper-based tube materials.

• C4SS. While the C4SS stainless steelcleaners can be used with all types of tube materials, they were originally developed for applications using AL-6XN stainless steel condenser tubes; however, they have also been used for cleaning tubes in highly corrosive environments.

• CB. This tube cleaner was specificallydeveloped to remove hard calcium carbonate deposits and is designed to break the eggshell-like crystalline form characteristic of these tube-scal-ing deposits. This cleaner has been found exceptionally helpful in avoiding the need for alternative and envi-ronmentally harmful chemical cleaning methods, which were all that were previously available for removing these types of hard deposits.

BrushesBrushes are principally intended toremove light organic deposits, such assilt or mud. They are also useful forcleaning tubes with enhanced surfaces

Figure 6(a) Model 200B Pump System

Figure 6(b) Water Gun

10


e.g., spirally indented or finned, or thosewith thin-wall metal inserts or tube coat-ings. The brush (Figure 5(d), Brush)length can be increased for more effective removal of lighter deposits.

Cleaning Productivity in the FieldUsing the methods outlined earlier in thissection, it is possible for 5,000 tubes tobe cleaned during a 12-hour shift, utiliz-ing a crew of four operators with twowater guns supplied from one pump.Clearly, an increase in crew size, in the number of water guns available, or in the number of mechanical cleanerssupplied for the project can increase thenumber of tubes which can be cleanedduring a shift, provided there is adequatespace in the waterbox(es) for the crew towork effectively.

Some Limitations of MechanicalCleaning MethodsEach type of off-line cleaning device hasits own limitations. For instance, brushesmay be effective only with the softest fouling deposits, whereas metal cleanersare more effective against tenaciousfoulants.

However, all methods may need assistance where the deposits have beenallowed to build up and become hard. In such cases, it may still be necessary to acid-clean, followed by cleaning withmechanical cleaners to remove anyremaining debris.

AIR INLEAKAGE DETECTIONAir inleakage can be inferred from anincrease in the air concentration in thegases drawn off by the air-ejector system.It is often associated with an increase incondenser back pressure. Note that thereis always a minimum air inleakage thatcannot be eliminated.

Westinghouse Electric Corp. used to recommend that air inleakage levels beheld to 1 ft3/min per 100 MW of gener-ation capacity, but other minima aregiven in the ASME and HEI standards.An increase in the dissolved oxygen concentration in the condensate can also indicate that air is leaking into the suction of the condensate pumpsbelow the condenser hot well.

Tracer gases are used to locate thesource of air inleakage, two of the most common being helium and sulfurHexafluoride (SF6). SF6 was first usedeffectively as an airborne tracer in atmos-pheric research and demonstrated itsgreater sensitivity as a tracer gas.

Sulfur hexafluoride, discovered in 1900,is a colorless, tasteless, and incom-bustible gas which is practically inertfrom a chemical and biological stand-point. It does not react with water, caus-tic potash, or strong acids and can beheated to 500°C without decomposing.

Two of its common uses within the utilityindustry are for arc suppression in high-voltage circuit breakers and the insula-tion of electric cables. SF6 also has manyother uses, such as etching silicon in thesemiconductor industry, increasing thewet strength of kraft paper and protect-ing molten magnesium from oxidation inthe magnesium industry.

The fundamental property of SF6 is that it can be detected in very low concentra-tions-as low as 1 part per 10 billion (0.1 ppb)-compared with the lowestdetectable concentration of helium of 1 part per million above background. It was later found that on-line injections utilizing SF6 also allowed leaks as smallas one gallon per day to be detected.

11


Analyzer to Detect SF6 inCondenser Off-GasIn the early 1980s, EPRI sponsored thedevelopment of an analyzer to detect thepresence of SF6 in condenser off-gas.Known as the FluorotracerTM Analyzer, itwas based on mass spectrometer technol-ogy. Figure 7(a), FluorotracerTM Analyzer,is a general view of the SF6 analyzer,while Figure 8 provides a schematic flowdiagram of a sampling system.

For the analyzer to detect the presence of tracer gas in the received sample, it is important that it be free from both mois-ture and free oxygen. This diagramshows how the off-gas sample is firstcooled, then shows how the moisture isremoved in a water trap and passedthrough a dessicant tower to remove anyresidual moisture. The diagrams alsoshow how the sample is finally receivedby the analyzer.

To remove any oxygen contained in thesample from the off-gas system, hydrogengas is introduced into the sample stream,entering with the sample into the catalyticreactor, where a chemical reactionoccurs between the oxygen and hydro-gen. The moisture is removed by anotherwater trap and dessicant tower.

The dried sample gas is then pumpedinto an electron capture cell, where itpasses between two electrodes and isionized by a radioactive foil. Ionizednitrogen in the sample supports a currentacross the electrodes, the current levelbeing reduced in proportion to the concentration of SF6.

Figure 7(a) FluorotracerTM Analyzer

Figure 8

Figure 7(b) SF6 Pak

12


An analyzer for use with SF6 as a tracergas is commercially available for use inboth fossil fuel and nuclear generatingstations. The analyzer is also providedwith an SF6 dispenser (shown in Figure7(b) SF6 Pak.) Weighing approximatelyeight pounds, the flow of the dispensercan be adjusted to obtain the SF6concentration needed to perform the leak detection task under current plantconditions.

Tracer gas leak detection involves a time delay between the injection of thegas and the response shown on the stripchart recorder connected to the massspectrometer. It should be noted that,while the indicators mounted on the caseof the mass spectrometer display theinstantaneous values of the measure-ments, they are difficult to interpret without the addition of a trace from astrip chart recorder. A typical example is shown in Figure 9.

The information available from therecorder chart not only provides a hardcopy for future use, but also tells techni-cians when they are getting close to aleak; when they have passed the leak;when they have located the leak; whetherthe gas is traveling to another leak; or,whether the leak is closer to the outletend. Whether a valve is leaking at thepacking, as opposed to the flange, canalso be determined.

SF6 can be used whenever helium isused; however, the reverse scenario isnot true. The following are among thefactors that go into the choice of tracergas for a given situation:

• Air inleakage into the unit.If the unit has more than 10 ft3/minof air inleakage, either tracer gas may be used. If the inleakage is less than 10 ft3/min, then SF6 shouldbe used.

• Dissolved oxygen (DO).The standard procedure for searchingfor the cause of DO leakage below the water line is to use SF6.

• Unit turbine power. If the unitis running at 20 percent or greater turbine power, either tracer gas may be used. If the unit has no turbine power and cannot be brought up to any level of turbine power, then helium should be selected.

• Unit size. Inspections of units of less than 50 MW capacity should always use helium.

Figure 9

13


Air inleakage inspections are best pre-ceded by a test shot to establish the levelof any background contamination and to ensure the instrumentation is function-ing correctly. It is recommended that allair inleakage inspections begin on theturbine deck, usually starting with the rupture disks. When on the mezzaninelevel, the tester should start spraying tracer gas at the condenser and workoutward along the hood, the expansionjoints, and other potential sources of airinleakage.

It is important to keep track of everythingthat is sprayed with the tracer gas.During a typical air inleakage inspection,it is not uncommon to spray tracer gas on literally hundreds of suspected leak-age paths within the condenser vacuumboundary.

In order to isolate a leak as quickly aspossible, technicians should know wherethey have been and what they haveseen. If a large leak is found on the man-way on, say, the west side of the turbine,an indication of this must be made on thestrip chart recorder. When the techniciangoes to the west side of the condenserand sprays a suspected penetration intothe condenser on the mezzanine level,the large leak could very well cause anerroneous indication of leakage there.Technicians can also waste a lot ofinspection time searching for a leak thatthey have already found on the turbinedeck. This is another reason the responsetime must be known.

Once the inspection results are in, it isimportant to understand what can andcannot be done with them. A leak detec-tion program is useful only if there is afollow-up repair program. Both SF6 and

helium detectors give readouts, one inmillivolts, the other in an arbitrary scale.

Plant personnel can determine a plan of action to repair the leaks after compar-ing millivolt or scale readouts. These arerelative values, not calibrated in engi-neering units such as ft3/min. It is of noimportance to have exact leakage values.

Generating stations already know whattheir total air inleakage is. Because of themargin of error due to all the variables ofa condenser under vacuum, quantifyingleaks does not add any information towhat the plant personnel already know.Therefore, it is not cost-effective. What isuseful is the exact location of each leakand its subsequent repair and retest.

When the turbine is not under power, itis very likely that the background concen-tration of the tracer gas will become sohigh that it eliminates any chance of isolating a leak. Both air inleakage andcondenser tube leakage inspectionsrequire vapor flow to carry the tracer gas out of the condenser with the rest ofthe non-condensibles.

If a sprayed tracer gas is sucked into thecondenser, it will begin to accumulate,and the background concentration willrise and saturate the detectors. There aresome occasions when, in an effort tobring a unit back up on-line as soon aspossible, a station has no choice but toattempt a tracer gas inspection with theunit shut down, and many have been successful in doing so. However, a minimum 20 percent turbine power isrecommended in order to perform a tracer gas inspection effectively.

14


CONCLUSIONThis White Paper has provided examplesand guidelines to help you understandthe factors that impact a return on invest-ment analysis. These illustrations focus onthe basic components necessary to under-stand a simple ROI analysis when consid-ering the payback for regular condensermaintenance within your plant.

In summary, it is about operating efficien-cy within your plant systems — doingmore with less. While it is difficult toassign a hard number to improving yourplant’s operating performance, you cancertainly calculate how much annual revenue is at risk should you fail to keepyour systems operating at peak performance.

When you are ready to investigate regular maintenance systems andoptions, we recommend that a moredetailed assessment be performed — specific to your plant’s unique requirements.

REFERENCES1. Putman, R.E., (2001), “Steam Surface

Condensers: Basic Principles, Performance Monitoring and Maintenance“, publ. May 2001, ASME Press, New York, NY.

2. Condenser Inleakage Guideline, EPRI Report TR-112819, publ. EPRI, January 2000.

3. Putman, R.E. (2001), Ibid, Chapters 2 thru 6.

4. HEI (1995), “Standards for Steam Surface Condensers“, 9th. edition, publ. Heat Exchange Institute, Cleveland.

5. ASME, (1998) “Performance Test Code for Steam Surface Condensers“, ASME PTC.12.2-1998,publ. American Society of Mechanical Engineers, New York, NY.

6. Putman, R.E. (2001), Ibid, pp. 75-76.

7. EPRI, (1994), Instrumentation of the On-Line Condenser Fouling Monitor, EPRI Technical Report TR-109232.

CONCO SYSTEMS, INC.530 Jones StreetVerona, PA 15147 USAPhone: 412-828-1166Fax: 412-826-8255www.concosystems.cominfo@ concosystems.com

15
