8/3/2019 Mps Feb03 Article

http://slidepdf.com/reader/full/mps-feb03-article 1/2

February 2003 Modern Power Systems 25

S TE A M T U R B I N E TECHNOLOGY

B igge r bla de s c u t c o sts

The shape and design of a steamturbine blade determine howmuch of the energy of the steamis turned into work. A longer laststage blade increases the power

output capability of the steam tu rbine, which

in turn leads to improved power plant effi-ciency and a lower e lectricity production cost.In the 1980s most turbine manufacturers

had developed and p ut into service titaniumblades with similar annulus areas. Due to sub-sequent improvements in both designmethodology and material characteristics, GEPower Systems and the Toshiba Corporationmade the decision to develop new bladesusing steel as a blade material. The use of steelresults in a lower cost to the customer andavoids uncertainty in the supply, and there-fore price, of high quality titanium forgings.

GE and Toshiba recently completed th e de-velopment of new 40 in and 48 in steel last-stage blades for steam turbine applications

worldwide. In terms of annulus area, the 48in blade is the largest stee l 3000 rpm last stageblade in the world.

Increasing the annulus areaOne of the loss mechanisms in the steam tur-bine is the kinetic energy of the steam as itleaves the last stage blade. The lower the ki-netic energy, the higher the steam turbine ef-ficiency will be. The magnitude of loss isproportional to the square of the ratio of thevolume flow rate of the steam th rough the laststage of the steam turbine and th e annulus areaof the turb ine exit. To decrease th e loss, a larg-er turbine exit annulus area is needed.

An increase in the last stage blade annulusarea can be accomplished by either usingshorter blades mounted on a larger diameterrotor (larger “hub”) or by using longer bladesmounted on a smaller diameter rotor. Theseopp osing approach es yield different radius ra-tios, a key parameter in the aerodynamic andmechanical design of the last stage blade. Inthe GE–Toshiba development p roject, an op -timisation of the aerodynamic and mechani-cal considerations of the design resulted in:● A new last stage (L-0) blade length of 1016

mm (40 in) for 60 Hz application and 1219mm (48 in) for 50 Hz application.

● A hub diameter of 1565 mm (61.6 in) and1880 mm (74 in), for 60 Hz and 50 Hz L-0

blades, respectively.Steam turbine last stage blades typically areapplied to different machine configurationswith varying outputs and operating points.This approach requires a robust d esign for awide range of operating conditions. Typicalranges of operation for the last stage of steam

turbine are from 5 to 21 kg/s/m2 (3 500 to 15000 lbm/h /ft2) end load; from 150 to 300 m/ s(500 to 1000 ft/s) annulus velocity; and from

4 to 10 per cent exit moisture.Based on a review of market needs, GE and

Toshiba chose the primary design point forthe 40 in and 48 in last stage blades to be 11kg/s/m2 and 225 m/s annulus velocity (8000lbm/h/ft2, 750 ft/ s annulus velocity), and 8 percent exit moisture. In addition to this prima-ry design p oint, two secondary design pointswere selected: 8 kg/s/m2 and 200 m/s annu-lus velocity (6000 lbm/h/ft2, 650 ft/s annulusvelocity); and 20 kg/s/m2 and 275 m/s annu-lus velocity (15000 lbm/h/ft2, 900 ft/s annu-lus velocity). The design is optimised at theprimary design point wh ile the secondary de-sign points are monitored to ensure that no

performance problems are introduced atthese points.

Aerodynamic designThe last three stages work together as a sys-tem and are designed aerodynamically usinga combination of streamline curvature designmethods, two-dimensional cascade analysis,and state-of-the-art th ree-dimensional comp u-tational fluid dynamics analysis techniques.

This design employs advanced aerodynam-ic feature s including meridional flowp ath con-touring, axial and tangential compound leanof the L-0 nozzle, and tailored exit profilesfrom the L-1 stage to allow a radius ratio of 0.43 in the L-0 blade.

To further reduce the radial pressure gradi-ent at the L-0 nozzle exit, compound tangen-tial lean is employed in the last stage nozzle.This tangential lean has the effect of intro-ducing an inward radial force on the flow –forcing more flow into the hub region and in-creasing the pressure in t his region.

Combined with the flowpath contouring,the L-0 nozzle lean reduces th e radial pressuregradient at t he n ozzle exit, raises the root re-action, and allows for a lower hub /tip ratio.

Axial compound lean is applied to the L-0nozzle in addition to the compound lean inthe tangential direction. The p rimary purposeof the axial lean is to increase the nozzle-to-blade spacing at the tip while maintaining asmaller axial spacing over the remainder of the

Boiler

DynamometerB

LP turbine

CondenserFlow meter

Spray water

DynamometerA

HP turbine

Steam conversion valve

Figur e 1. Last

stage blade

Figur e 2. Test st eam tur bine system

This year will see the commercial

introduction of the world’s largest

steel last-stage blades for steam

turbines.

A m ir M u je z i n o v ic , GE,Schenectady, NY, USA

8/3/2019 Mps Feb03 Article

http://slidepdf.com/reader/full/mps-feb03-article 2/2

February 2003 Modern Power Systems 27

height. A larger axial spacing is desired at thetip to allow additional time for the waterdroplets torn from the trailing edge of the noz-zle to accelerate to flow velocity before en-tering the L-0 blade. Better matching of thedrop let and flow velocity reduces erosion andcontributes to the long-term reliability of thisdesign.

Mechanical design of stagesOn the last stage blade the most important me-chanical design features are the curved axialentry dovetail (a portion of the blade thatmates with the rotor), nub and sleeve as a partspan damper, and an integral blade cover (seeFigure 1).

The mechanical design consists of severalinterconnected steps: transforming an aero-dynamic shape of the airfoil into an as-ma-chined shape; static design; dynamic design

(or aeromechanics design); and erosion de-sign.

The airfoil shape used in the aerodynamiccalculations is the one that the airfoil will as-sume at the running speed. The low-pressureblades untwist during acceleration from restto runn ing speed. A necessary step during themechanical design is to follow a process thatdetermines the zero-speed airfoil shape thatyields the desired aerodynamic shape at op-erating speed.

The basic premise in the static design of theairfoil and its accompanying rotor wheel isthat both the maximum average stress in allthe w heel and blade sections – and the maxi-mum local stress throughout the blade and th e

wheel – are kep t under a certain level, estab-lished to ensure substantial margin to failureunder all operating conditions and extremefaulted conditions.

The maximum average stress (assumingfixed material properties) determines theoverspeed margin: when compared to theyield strength of materials (blade and wheel)it determines the overspeed at which sect iongross yielding would occur; and when com-pared to the ultimate strength, it determinesthe overspeed at wh ich ductile failure wouldoccur.

In contrast, the maximum local stress de-termines the low cycle fatigue life (number of 

cycles to a crack initiation) and also plays arole in the level of stress corrosion cracking(SCC) risk. The 40 in and 48 in last stage bladeswere designed to con servative stress and over-speed criteria in order to ensure safe and reli-able service for the life of the machine.

The basic premise in the dynamic design

(aeromechanical design) of the airfoil is thatthe blade natural frequencies – those fre-quencies at which vibration will take place inthe absence of any continuing excitation atrunning speed – must have a margin from arange of multiples of engine speeds (so-calledper-rev lines). The per-rev excitation comesfrom different sources, most of which are as-sociated with the non-uniform flow in thesteam path, either upstream or downstreamfrom the blade row.

In addition to the per-rev excitation, longlow-pressure blades (in particular last stageblades in some of the off-design operatingregimes) can experience vibration induced byflow (aeroelastic instability). Design solutionsdealing with aeroelastic instability have beenprimarily empirical in nature and specific toapplication.

The risk of aeroelastic instability is greatest

in the region of low flow (low steam axial exitvelocity) and high condenser pressure (a far-off-design condition). In such conditions amassive flow separation from the hub of thelast stage blade forces the migration of thesteam flow to the upper portion of the laststage blade. Flow separation from the airfoilthen can cause blade stall flutter, and flow in-stability can cause buffeting of the blades.

GE and Toshiba took a three -fold approachto guard against such instability:

1.Use of a last stage blade design th at emp loys

integral covers and part span dampers. Thisconstruction increases the frequency of therow of blades (when comp ared to free-stand-ing blades) and is a well recognised mitigationagainst instability. In addition, this construc-tion provides additional damping at the loca-tions of contact between adjacent blades.

2. Analysis using a reduced frequency analy-sis calibrated to empirical data.3. Validation testing in a subscale test rig,where the last three rotating and stationarystages were tested in actual steam conditions,at a variety of operating conditions (combi-nations of the steam axial exit velocity andcondenser pressure). Blade mounted dynam-

ic strain gauges were used to measure bladevibration.

TestingThe experimental verification of the overallstage efficiency of the new steel last stageblade was performed in an exp erimental low-pressure model turbine. A schematic diagramof the 10 MW model steam turbine facilityused is shown in Figure 2, and a photographof the model turbine train used in this systemis shown in Figure 3.

Model turbine tests have been performedwith several different end load conditions.The LP end exit axial flow velocity conditionsvaried from 94 to 212 m/s (310 to 700 ft/s).

The results show that the n ewly developedsteel 48/40 in last stage blade has exce llent ef-ficiency at the design condition and also in thepartial load conditions.

Mechanical testing confirmed the mechan-ical design calculations described above. Therunning speed to zero speed airfoil transfor-mation, static design and the dynamic design(blade natural frequencies) were validated ina wheel box test.

The test set-up consisted of a full-size singleflow rotor w ith the last th ree stages of blades(also full size) assembled on it (Figure 4). Theblades are instrumented with strain gaugesand the entire assembly is put into a spin cell

– basically, a bunker from which air can beevacuated. The test rotors are spun to th e de-sired speed using a drive turbine, the bladesexcited and blade frequencies measured atvarious speeds.

In addition to the blade frequency mea-surement, a rotor end mounted torsional shak-er was used to measure coupled rotor bladetorsional frequencies. To confirm the calcu-lated blade un twist, a strobe light was used tovisualise the cover un twist. Results confirmedthe rotating speed at which adjacent coversengage and from th at, it was inferred that theanalytical transformation of the airfoil fromrunning airfoil to the zero speed airfoil wascorrect.

During wheel box testing, both static anddynamic strain gauges were used for blade in-strumentation. Static gauges were mountedboth on the airfoil and in the w heel and bladedovetail to measure static stress level. The re-sults correlated very well with the analyticalresults. The dynamic strain gauges measuredblade-vibrating frequency and coupled rotorblade torsional vibration. The measured bladefrequencies confirmed the analytical predic-tions and the measured coupled blade rotortorsional vibrations confirmed the predictedmargins from two times electrical frequency.

Market introduction

Overall, the newly developed 40/48 in laststage group is a significant contribution tosteam turbine performance and reliability.

Beginning in 2003, GE and Toshiba exp ectto offer the new blades on their steam turbinesfor both combined cycle and fossil powerplant applications.

Figur e 3. Test model tur bine

Figur e 4. New blades installed on a r otor

S TE A M T U R B I N E TECHNOLOGY

MPS