Output power variation

During the winter of 2001-2002,several of the Madison turbineunits experienced inconsistentmeasurements from the chilledmirror dewpoint/humidity in-strumentation, resulting in errat-ic inlet bleed heat valve opera-tion for the anti-icing system.The output power from theunits was subsequently varyingby several MW, making it diffi-cult for the operators to main-tain hourly output accurately.Further, the cycling action onthe unit hardware would eventu-ally have an adverse impact oncomponent life. Also, the opera-tors of the electricity grid werenot pleased with the fluctuatingpower generated. A study wasundertaken to identify the causeof the erratic measurements and

to possibly recommend a re-placement dew point/humiditysensor.

Many humidity sensors usedon gas turbines determine dew-point using a chilled mirror hy-grometer. Chilled mirror hy-grometers are accurate and pro-vide repeatable measurements.However, contamination of themirror surface is a commonproblem and this requires prop-er maintenance. Contaminationcan result in substantial meas-urement error, which is reme-died with mirror cleaning. Con-sequently chilled mirror systemstend not to experience drift withtime but can be maintenance in-tensive. Other technologies,such as the polymer (capaci-tance) sensor, do have some mi-nor drift, but are lower cost andmuch lower maintenance.

The original factory suppliedhumidity sensor in the MadisonGE 7EA units used a chilled mir-ror device. The output is dew-point, which is fed into the GEcontrol system where the specif-ic humidity is then calculated.The device puts out a 4 – 20 mil-liamp signal. The GE interfacecan accept either a 4 – 20 mil-liamp or a 0 – 10 volt signal.

Study of technologiesAn investigative study was con-ducted to identify the technolo-gies used by different dew-point/humidity sensors andtheir suitability for a combus-tion turbine environment as-sessed. Other parameters consid-ered included accuracy, main-tenance, power supply, outputsignal and cost. Since the Madi-son units also have inlet fogging

systems for hot day operation,the impact of very high humidi-ty on the dewpoint/humiditysensor was another considera-tion These are some of the con-clusions of the study:

1. All the dewpoint/humidi-ty products evaluated in thisstudy were considered to be ca-pable of satisfying the need toprovide the humidity informa-tion necessary for determiningthe need to activate inlet bleedheat for anti-icing.

2. The chilled mirror sensors,while very accurate, seemed tohave a difficulty with contami-nation of the mirror and re-quired frequent maintenance(cleaning). The capacitance sys-tem appears to be more tolerantof contamination problems.

3. While the chilled mirrorsystems are the most accurate,

20 166/2004

The Cinergy Madison Generating station, locatednorth of Cincinnati, Ohio, came into commercial operation in July 2000. The plant has eight GE 7EAsimple cycle combustion turbines that are used togenerate electricity primarily for peaking duty. Total generating capacity is 576 megawatts. Initially,these units were not equipped with anti-icing protection for winter operation. However, duringthe first winter of operation, a sister site experiencedice damage to the compressor. Subsequently, themanufacturer was requested to install anti-icing protection which was accomplished before its second winter of operation.

The Challenge of Measuring Inlet Humidity onCombustion Turbines

Harold StockerCinergy EnergyUSA

the capacitance systems are cer-tainly acceptable for the 7EA in-let humidity measurement.

4. Costs vary sufficiently tobe a factor in which system to se-lect.

5. Several dewpoint/humidysensors have been certified bythe OEM.

6. Some dewpoint/humiditysensors are direct drop-in re-placements, while others will re-quire modifications to the exist-ing duct opening. Generally, thechilled mirror units have every-thing mounted on the inlet ductside wall, while the capacitancesensors have the probe mountedon the duct side wall with thetransmitter located separately.

An evaluation table was usedto rank and prioritize the variousparameters, mentioned earlier,

to be considered in the selectionof the replacement sensor instru-mentation. The Vaisala Humidi-ty and Temperature TransmitterHMP235 was initially deter-mined to be the best value forthe cost. The capacitance tech-nology avoids much of thechilled mirror maintenance andVaisala has an excellent reputa-tion for reliability. The HMP235is the same as the Vaisala Hu-midity and Temperature Trans-mitter HMP247 except it doesnot have the heated sensor headwhich eliminates dew formationin condensing environments(i.e. downstream of foggers). Forthis reason, the final selectionwas the HMP247 transmitter.Vaisala had also developed anadapter mounting system thatmade use of the existing instru-mentation location.

The Madison turbines arenow operating with theHMP247 transmitters installed.The instruments have been reli-

able and trouble-free, enablingus to improve the overall perfor-mance of the generatingstation.●

166/2004 21