Ey 2610381046

7/30/2019 Ey 2610381046

http://slidepdf.com/reader/full/ey-2610381046 1/9

A.V.Pradeep, Kona Ram Prasad, T.Victor Babu / International Journal of Engineering

Research and Applications (IJERA) ISSN: 2248-9622 www.ijera.com

Vol. 2, Issue 6, November- December 2012, pp.1038-1046

1038 | P a g e

Design And Analysis Of Wind Turbine Blade Design System

(Aerodynamic)

A.V.Pradeep*, Kona Ram Prasad**, T.Victor Babu***

* (Department of Mechanical Engineering , S.V.P.Engineering College, Visakhapatnam)**(Department of Mechanical Engineering , S.V.P.Engineering College, Visakhapatnam)***(Department of Mechanical Engineering , S.V.P.Engineering College, Visakhapatnam)

ABSTRACTThe ever increasing need for energy and

the depletion of non-renewable energy resources

has led to more advancement in the "Green

Energy" field, including wind energy. An

improvement in performance of a Wind Turbine

will enhance its economic viability, which can be

achieved by better aerodynamic designs. In the

present study, a design system that has beenunder development for gas turbine turbo

machinery has been modified for designing wind

turbine blades. This is a very different approach

for wind turbine blade design, but will allow it to

benefit from the features inherent in the

geometry flexibility and broad design space of the

presented system. It starts with key overall

design parameters and a low-fidelity model that

is used to create the initial geometry parameters.

The low-fidelity system includes the

axisymmetric solver with loss models, T-Axi(Turbomachinery-AXIsymmetric), MISES blade-

to-blade solver and 2D wing analysis codeXFLR5. The geometry parameters are used to

define sections along the span of the blade and

connected to the CAD model of the wind turbine

blade through CAPRI (Computational Analysis

Programming Interface), a CAD neutral API

that facilitates the use of parametric geometry

definition with CAD. Either the sections or the

CAD geometry is then available for CFD and

Finite Element Analysis.The GE 1.5sle MW wind turbine and

NERL NASA Phase VI wind turbine have been

used as test cases. Details of the design system

application are described, and the resulting windturbine geometry and conditions are compared

to the published results of the GE and NREL

wind turbines. A 2D wing analysis code XFLR5,

is used for to compare results from 2D analysis to

blade-to-blade analysis and the 3D CFD analysis.

This kind of comparison concludes that, from

hub to 25% of the span blade to blade effects or

the cascade effect has to be considered, from25% to 75%, the blade acts as a 2d wing and

from 75% to the tip 3D and tip effects have to be

taken into account for design considerations. In

addition, the benefits of this approach for wind

turbine design and future efforts are discussed.

I. INTRODUCTIONBecause of the increasing need for energy

and to reduce the need for non-renewable energyresources, efforts are being made to utilize

renewable energy to a great extent. Wind energy isone such abundant resource, and huge efforts areunder way to make the available wind turbines more

efficient and more economical to operate. This thesis presents a turbomachinery

approach for wind turbine blade design forhorizontal axis wind turbines. Discussions on the

process, shown in Fig. 1.1, is detailed in Chapter 2.This chapter also includes the details for usage of T-AXI as a design tool for wind turbine blade design.

Comparison of geometry available in literature isalso presented in the same chapter. A brief discussion on T-AXI as a turbine design tool isincluded in Chapter 2. In Chapter 3, the use of MISES to analyze the blade profiles is detailed, anda comparison of aerodynamic data available is

made, to show where cascade effects matters in suchkind of machines. In Chapter 4, the use of a winganalysis code, XFLR5 for wind turbine blade isexplained and a 2D application of wind turbines is

explored. Chapter 5 explains the 3D CFD analysisdetails using Fine/Turbo. Chapter 6 discusses modalanalysis of wind turbine blades and FEA results forthe blade model generated through 3DBGB. Thefinal Chapter 7 includes a summary and futuredirections.

Figure 1.1: Process Flowchart for wind turbinedesign system.

7/30/2019 Ey 2610381046





1039 | P a g e

Figure 1.2: Lift and Drag on a wind-turbine blade profile

2. Wind Turbine Design Using T-Axi A wind turbine blade can be considered a

huge turbine blade without the casing. Thus, the’TT-Des’ module of T-AXI is used to initializeturbine blade flow parameters. First, TT-Des isexecuted with ‘INIT’ and ’Stage’ files. These aretext format files, where flow parameters can be input

or changed to get desired results. This allowsbootstrapping the calculations initially and values

for T-Axi execution is obtained. The details of theflow parameters as published in the technicalspecification hand book of G.E1.5sle MW [17] and

NREL phase VI blade [31] are used as test cases toreverse engineer the blade shape from theseparameters.

FIg 2.1: CAD Blade Design (NREL Phase VI blade)

Fig 2.2: NREL reverse engineered Wind turbineBlade.

Fig 2.3: G.E Wind turbine Blade [17].

Fig 2.4: G.E reverse engineered Wind turbine Blade

3. Cascade Analysis using MISES

MISES is a viscous/inviscid cascade solver

and design system. The program is a complete CFD procedure from geometry definition to post

processing tools. It is a quasi-3D computationalmethod used for design and analysis of airfoils foraxial turbo machinery designs. It has a finite volumeapproach to flow discretization. The inviscid flow isdescribed by Euler’s equations and viscous effects

are modeled using integral boundary layerequations. The cou- pled system of the nonlinear

equations is solved by a Newton-Raphsontechnique. MISES also uses the Abu-Ghannam/Shaw (AGS) for transition prediction.

7/30/2019 Ey 2610381046





1040 | P a g e

3.1Wind Turbine Design and Analysis using

MISES The 3DBGB code generates the ’blade’

and ’ises’ files that form the input to the MISESanalysis and redesign. As discussed earlier, MISES

is a cascade solver with the ability to have

boundary layer coupling included during execution.This is achieved by a Reynolds number input forthe blade section in the ’ises’ file. The inputs used

for the 3DBGB- NREL blade that is ’blade.case’

file and corresponding ’ises’ file is attached inAppendix L. The blade coordinates are the m’, θ

points on the blade surface, that starts fromattached in Appendix L. The blade coordinates arethe m’, θ points on the blade surface, that starts

from the trailing edge and then goes round theleading edge back to the trailing edge, but is notclosed, so that a blunt trailing edge is achieved.This is done to incorporate the Kutta condition over

finite thickness. Fig. 3.1(a) shows the cascadearrangement for the MISES setup. The pitch

between the blade sections forms thecircumferential separation of the cascade. The pitchvalue is set in the blade file. The ‘iset’ commandalong with the case extension sets up the case to

run in MISES. This creates the grid file for the

cascade as shown in Fig. 3.1(b).

a) Blade section in a cascade arrangement.

(b) Grid for the cascade arrangement.

Figure 3.1: MISES initial settings.

Figs. 3.2, 3.3, and 3.4 shows the MISES outputplots of shape factor, momentum thicknessReynolds number, and skin friction coefficient forall the three profiles.

(a) H plot - NUMECA-3D

(b) H plot - NREL-S809.

7/30/2019 Ey 2610381046





1041 | P a g e

(c) H plot - 3DBGB-NREL. Figure 3.2: Shape factor plot from MISES at midspan.

(a) Reθ plot - NUMECA - 3D.

(b) Reθ plot - NREL- S809.

(c) Reθ plot - 3DBGB-NREL. Figure 3.3: Reθ plots at mid span.

(a) Cf plot - NUMECA - 3D.

(b) Cf plot- NREL- S809.

7/30/2019 Ey 2610381046





1042 | P a g e

(c) Cf plot - 3DBGB-NREL.

Figure 3.4: Co-efficient of friction plots at mid

span.

4. Wing Analysis using XFLR5 XFLR5 is an open source code, used for

analysis of wings and airfoils, and is based on

XFOIL. XFLR5 is easy to use and no backgroundon how to run XFOIL is needed. This code wasused to see the correlation between the MISES

analysis and 2D wing analysis. Thus, all the profiles were analyzed using XFLR5, with the

same conditions as analyzed in MISES. XFLR5uses XFOIL code as its base, thus the blade files

which were used for analyzing in MISES, could bereused. The Cp plot for each profile was generated

from this code. Fig. 4.1 shows a comparison of theCp thus generated and plotted against m′, for thethree profiles analyzed through XFLR5 and itscomparison to 3D-CFD result (described in the

next chapter). The analysis shows the correlation of a 2D wing airfoil analysis to a 3D-CFD analysis forthe wind-turbine blade, showing the 2D nature of such kind of machines. This fact is further

discussed with the help of 3D-CFD results in detail,in chapter 5.

Figure 4.1: Cp comparisons at mid span from

XFLR5.

Figure 7.2: Cp comparisons at mid span betweenXFLR5 and MISES (en ).

5.3D-CFD Analysis using Fine Turbo Fine/Turbo has a post-processing module

called ’CFVIEW’. The ’.run’ file generated byEURANUS is loaded in CFVIEW. The desiredflow quantities that were selected to be output

during the flow solver execution shows up in thegraphics window and can be selected for contour

plots or line plots. If required, new quantities aredefined and the flow solver is executed with oneiteration. This calculates the new quantity andshows up in CFVIEW. Fig. 5.2 shows the Y+

values on the blade surface. The Y+ value wasguessed and the ywall value was input initiallyduring grid generation.

Figure 5.1: Cp comparison between ’000’ grid and

MISES. .

(a) Y+ Suction side. (b) Y+ Pressureside.Figure 5.2: Y+ values.

7/30/2019 Ey 2610381046





1043 | P a g e

Figure 5.3: Cp plot at 50% span.

Figure 5.4: Cp comparison at mid span.

(a) Contour plot of radial velocity.

(b) Contour plot range.

Figure 5.5: Radial Velocity plot for NUMECA-3D.

(a) Contour plot range.

(b) Contour plot of Phi angle.

Figure 5.6: Phi angle plot for NUMECA-3D.

Figure 5.7: 2D line plot of phi angles forNUMECA-3D.

7/30/2019 Ey 2610381046





1044 | P a g e

Figure 5.8: Wing tip vortex

Figure 5.9: Iso-surface of static-pressure forNUMECA-3D.

(a) Area averaged contour plot of static-pressure in meridional view.

(b) line plot of static-pressure.

(c) Non-dimensional plot of static-pressure.

Figure 5.10: Area averaged plot for static-pressure

for NUMECA-3D in meridional view

.(a) Mass averaged contour plot of rVθ meridional view.

(b) line plot of rVθ .

7/30/2019 Ey 2610381046





1045 | P a g e

(c) Non-dimensional plot of rVθ . Figure 5.11: Mass Averaged plot for rVθ for

NUMECA-3D in meridional view.

6. FEA stress and Modal Analysis The wind turbine blade thus was analyzed

for its various mode shapes to account for theflutter and acoustics [4], [38]. Sample modal output

from ANSYS for the continuous slope disk isattached in the Appendix F. A subroutine

ANSYS_WRITER D is written to output a’ANSYS.AIN’ file which gives an ANSYS Para -metric Design Language (APDL) script, which

opened in ANSYS, automatically generates meshed part file that is ready for any kind of FEA study to

be done in ANSYS. The Meshed Wind TurbineBlade, with 8 node hexahedron brick 185 elementare shown in Fig.9.1. The first five mode shapes for

the GE reverse engineered Wind Turbine blade isshown in Fig. 9.2. Table 9.1 shows the first five

natural frequencies of the GE reverse engineeredblade, when simulated as a cantilever beam whichis rotating. The material used for the test case was

Aluminum to demonstrate the capability, althoughmost of the wind turbines are made of fiberglass or

other types of composites. The GE 1.5sle windturbine is rated for a range of wind speeds (3.5 m/s

- 25 m/s). Thus, it will have different rpmsassociated with these wind speeds, as the angularvelocity is directly proportional to the wind speeds,

and is given by the following correlation : ω=60Vz λ/ Πd

The fundamental frequency calculated fromangular velocity, is given by the correlation:

f= ω/60 Thus, calculation the rotational frequency

using the Equation (9.2), yields a range (0.2≤ f ≤1.433). The value of first modal frequency, as

tabulated in Table 9.1 is well below the resonantfrequency ranges. Structural analysis for the above

case was executed and the Von-Mises plot isshown in Fig. 9.3. Von- mises stress is often usedto estimate the yield criteria of materials. The von-mises criterion states that, failure will occur, if the

von-mises stress reaches a critical limit or yieldstrength of the material. Thus, FEA analysisidentifies the areas where this value is attained, isanalyzed and avoided by design changes or

strengthen the areas of high stress. For the present

study aluminum was used as the material. The yieldstrength of aluminum is 414 Mpa. From Fig 9.3,the maximum value of the von-mises stress is

47.417 Mpa (SMX), which is less than 1/3 times

the yield strength. The importance of the aboveexercise was to show the FEA analysis part of theproposed design system. Also, the above case wasexecuted as a solid body, and simulated like a

cantilever beam problem, to make it easier and

show the capability of FEA coupling to the systemproposed.

Figure 6.1: Meshed GE reverse engineered wind-turbine Blade with 8 node Hexahedron Brick 185

element.

(a) Mode-1(flapwise).

(b) Mode-2(edgewise).

(c) Mode-3(flapwise).

7/30/2019 Ey 2610381046





1046 | P a g e

(d) Mode-4(mixed).

(e) Mode-5(mixed).

Figure 6.2: First five Modal solution of GE reverseengineered blade.

Figure 6.3: Von-Mises Stress plot on GE reverse

engineered blade.

7. Conclusion An approach to the wind turbine design

from a turbo machinery perspective is presentedthat can leverage many of the design codes and

processes developed for axial turbines, open rotors,axial compressors, and fans. The multi-disciplinary

design system has the ability for geometry creationand analysis for axial compressors. It is beingadapted for wind turbines which have their ownunique issues. The multi- disciplinary approachmakes it easy to address a vast number of

aerodynamic and structural issues. A parametricdesign tool for geometry has been developed thatwill help implement quick design changes from a

command line input. The system developed wasinvestigated with the in-house turbo machineryaxisymetric solver, T-Axi, as it was easy to changethe source code to suit our needs for wind turbine

design. For a conventional horizontal axis windturbine, the analysis shows: • The lower 25% of span should account for

cascade effects. • From 25% - 75% span, the wind turbine can beassumed 2D and isolated (wing theory applicable). • From 25% - 75% span, the wind turbine can beassumed 2D and isolated (wing theory applicable).

8. Future Work Capabilities from a turbomachinery design

system have been adapted for use for windturbines. This approach can add understanding of wind turbines from classical turbomachinery

methods. This design system is a foundation and isnow extendable. It will be easy to add unique tiptreatments, as well as new environment. The futurework should include the validation of the tool with

other available tools for design of wind turbineblade. The presented code should also be tied to anacoustic module for noise prediction from the

blades and ways for reduction through design

changes. Also other modifications to the blade design such as a parametric tip design

for reducing the tip noise effects and improvingeffeciency will be possible.

Now that a wind turbine blade designsystem has been established using axial

turbomachinery con- cepts, further analysis toinclude available blade profiles (NREL airfoils forHWATs) can be used in the code to take theadvantage of designing the wind turbine blade in amore realistic manner. Wind turbines must dealwith off-design and pitch changes which make

them different from axial machines. A

methodology that combines the wind turbinefeature using T-AXI and conventional Bladeelement method should be developed.

Distortion analysis to understand theearth’s boundary layer effects and the effects of thepylon on the rotating blades of the wind turbine

should also be developed. This should be possiblewith the Non- Linear Harmonic capability in the3D CFD code Fine/Turbo, but needs somedevelopment.

Reference[1] A.C.Hansen and C.P.Butterfield.

Aerodynamics of horizontal-axis windturbines. In Annual Rev.FluidMech.1993.25:115-49, 1993.

[2] G.G. Adkins-Jr and L.H. Smith-Jr.Spanwise mixing in axial-flowturbomachines. Journal of Engineering for

Power, 104:97 – 110, Jan,1982.[3] Dayton A.Griffin. Blade system design

studies volume ii : Preliminary bladedesigns and recom- mended test matrix.

Technical report, Sandia NationalLaboratory, California, USA, 2004.

[4] ANSYS. Websitehttp://www.ANSYS.com

Documents

Ey 2610381046