Wind Tunnel Tests for the Validation of the New Brazilian Wind Code
Wind Tunnel Tests for the Validation of the New Brazilian Wind CodeAcir M. Loredo-Souza1, Marcelo M. Rocha1, Gustavo J. Z. Nez2, Mario G. Klaus Oliveira21Prof. of Civil Engineering, Univ. Fed. do Rio Grande do Sul, Porto Alegre, Brazil, firstname.lastname@example.orgResearch Associate, Lab. de Aerodinmica das Construes, UFRGS, Porto Alegre, BrazilABSTRACTThis paper describes the experimental program, through wind tunnel testing, established for thevalidation of the new improvements in the Brazilian Wind Code NBR-6123 , which its currentedition is from 1988. The wind tunnel facility, as well as some of the models tested, are describedtogether with the presentation of the comparisons, main modifications and improvements for thenew edition of the Brazilian Wind Code. Emphasis is given for torsion in buildings, the findingsfor transmission towers and the new chapter on vortex shedding. The new improvements in theBrazilian Wind Code lead to a more comprehensive and safe code, certainly conducting to amore cost efficient and safe design.INTRODUCTIONThe current Brazilian Wind Code  has its main format based in a quasi-steady approach, withthe forces being determined by the product of a dynamic pressure q, an aerodynamic coefficientC and a reference area A, as shown in Equations 1, 2 and 3. In these, Vk is the design wind speed,V0 is the basic (or reference) wind speed, S1 the topographic factor, S2 the factor taking intoaccount terrain roughness, building dimensions and height above terrain, and S3 the statisticalfactor.F = q (Ce Ci) A (1)q = 0.613 Vk2 (2)Vk = V0 S1 S2 S3 (3)The code presents a wind map (Figure 1) with the reference wind speeds based on 3-second gust wind speed at 10m height in open terrain, with an annual probability of exceeding0.02. There is not a separation of the type of climatological event which generated eachregistered velocity. Therefore, a thunderstorm (TS), an extratropical pressure system (EPS) oreven a tropical cyclone (TC) are treated the same and its resulting velocities absorbed withoutdifferentiation. There are a lot of meteorological stations and a constant work for updating therecords; however, the correspondent updating of the wind code do not have the deservedattention, mainly due to the lack of human and financial resources to deal with this task.Exception is the wind research team of the Universidade Federal do Rio Grande do Sul(UFRGS), in the city of Porto Alegre, which is constantly working for the improvement of theBrazilian Wind Code, but that also face great difficulties with regard to public funding. Most ofthe funding comes from commissioned wind tunnel tests, even though the results benefit thegeneral public. The code also brings a series of aerodynamic coefficients related to severalmailto:email@example.com, as well as indications for the determination of topographical effects, vicinity effects,internal pressures, torsion and others. There is a chapter dedicated for the determination of thedynamic effects due to atmospheric turbulence and an indication on how to verify if the structurewill have problems regarding to vortex shedding.Figure 1: Map with the reference wind speeds, in m/s, 3-s gust, at 10m height (NBR-6123)THE NEW DATA AND ITS VALIDATION THROUGH AN EXPERIMENTAL PROGRAMIn order to validate the new modifications, procedures and data, an experimental program was setto accomplish this task. The tests were performed at the Prof. Joaquim Blessmann boundarylayer wind tunnel , located in the Laboratrio de Aerodinmica das Construes (LAC) of theUniversidade Federal do Rio Grande do Sul (UFRGS), in Porto Alegre, Brazil, shown in Figure(2). Due to limitations is paper size, the focus will be on the torsion for buildings andtransmission tower data. More information is available with the authors.Figure 2: Prof. Joaquim Blessmann boundary layer wind tunnelTORSION AND FORCE COEFFICIENTS FOR BUILDINGSIn the current Brazilian wind code, the torsional moment on prismatic buildings of rectangularsection a x b (a b) is calculated considering the drag force, Fa (for wind perpendicular at eachfaade) acting, respectively, with the following eccentricities referred to the geometric verticalaxis: (1) without neighboring effects: ea = 0.075a and eb = 0.075b; (2) with neighboringeffects: ea = 0.15a and eb = 0.15b, where ea is measured along the length of the majorfaade of the building and eb along the width of the building (length of minor faade). Theexpressions above show that the Brazilian wind code considers a neighboring factor (NF) equalto 2. Fa is obtained through Equation 4, and the drag coefficient, Ca, is obtained from graphicalformat as shown in Figure 3, for high and low turbulence winds.Fa = q Ca A (4)Recent studies at LAC  indicate that a more representative value for the eccentricity isthe 15% of the faade, and thus this value is adopted solely. The conclusions were based on aseries of wind tunnel tests on tall building models, as well as with comparison with other codeswhich consider torsional effects [4, 5] and available papers [6, 7, 8]. Some of the buildingmodels are indicated in Table 1, and comparative graphs between the torsional momentsobtained in the wind tunnel and from the Brazilian Wind Code are shown in Figures 4 and 5.Also revised are the values of the force coefficients for tall buildings. The above mentionedwind tunnel tests and literature review also point that, for several cases, higher forces areobtained for oblique wind incidences than for perpendicular wind incidence. This reflects inforce coefficients higher than those drag coefficients presented in the current edition.Figure 3: Drag coefficient Ca for buildings in high (left) and low (right) turbulence winds (NBR-6123)Table 1: Examples of some building models studied for torsional effectsModel Picture Cross Section Height1Torre deMlaga118,00 m2LEssenceJardins120,10 m3RochaVer75,00 m4SP Wellness93,50 m5BrascanCentury108,60 m6CyrelaClassiqueKlabin72,52 mModel Picture Cross Section Height7GafisaEldorado142,50 m8e-Tower149,50 m9Mandarim63,50 m10Sundeckb = 22,08 ma = 43,82 m56,50 m11Sunset93,30 m12Estrela doAtlntico130,24 mLATTICED STRUCTURES AND TRANSMISSION TOWERSAn extensive wind tunnel program  was commissioned for the study of transmission towers.Examples of wind tunnel models are shown in Figure 7. They are representative of two typicaltowers from FURNAS, a Brazilian Power Company. The two main objectives of the study were: Verify the applicability of the criteria currently established in transmission linescodes and guidelines [1, 10, 11]; Determine drag and force coefficients on latticed transmission tower modelscompatible with the geometries most used in real towers, as a basis for therevision of the Brazilian Wind Code NBR-6123  and the BrazilianTransmission Lines Code NBR-5422 .MODEL 1 MODEL 2MODEL 3 MODEL 4MODEL 5 MODEL 6Figure 4: Comparative graphs of the torsional moments, Mt , obtained from the wind tunnel measurementsand from the Brazilian Wind Code for models 1 to 6MODEL 7 MODEL 8MODEL 9 MODEL 10MODEL 11 MODEL 12Figure 5: Comparative graphs of the torsional moments, Mt , obtained from the wind tunnel measurementsand from the Brazilian Wind Code for models 7 to 12The most employed method established in the codes, and used by engineers fordetermining the wind loading on latticed structures , consists in the determination of totalforces in modules of the lattice structure. Drag or force coefficients are determined through windtunnel tests in which the sectional tower models are put into force balances. The measured dragforce, Fa, is divided by the dynamic pressure, q, and a reference (effective) area, A, as shown inEquation (5).The drag coefficient is then associated with its corresponding solidity ratio, ,becoming adequate for codification purposes (Figure 6). The solidity ratio is an important factorthat influences the force coefficients for lattice truss structures, being defined as the ratiobetween the effective frontal area (area of all members in the windward face) of the structure, Ae,and the frontal area (area of the outline of the windward face) of the structure, A (Equation 6).The process automatically includes shielding effects.AqFC aa (5)AAe (6)Figure 6: Drag coefficients for square section latticed truss structures In general, towers used for transmission lines are built with sharp edge members and arenot Reynolds number, Re, sensitive. As indicated in Equation 7, Re is defined as the product ofthe wind velocity, V, and a characteristic dimension, l, divided by the kinematic viscosity of theair, . Therefore, the drag coefficients may be considered independent of the wind velocity forthese types of structures. However, when building a reduced model, it is necessary to check if theRe of the individual members are high enough to be velocity independent. Holdo  foundwind tunnel test results on model lattice structures below 2200 based on the lattice member crosssection to be Re sensitive, with differences of up to 40% in terms of drag coefficients. On theother hand, Vickery  states that data on sharp edged shapes suggests that local Reynoldsnumbers in excess of 1000 will be sufficient to prevent (or considerably diminish) model scaleeffects in lattice or framed towers.lVRe (7)From de design drawings, several reduced models of sections (modules) of transmissiontowers were built. The models were divided into three groups: (a) head and its component parts,(b) tower main body and extensions, (c) legs. Figure 7 show models for FURNAS Towers A33(for 345 kV) and A55 (for 500 kV). The selected modules were chosen as representative of theformation of the mentioned towers, following the recommendation of NBR-5422  and IEC-60826  to limit the division of the tower in segments with a maximum height of 10 meters.The tests were performed with different wind velocities and some of the models were tested indifferent scales for checking eventual Re sensitivity, with the main scales being around 1:11 to1:20. With regard to Figure 8, the tests angles of incidence were 0o, 45o and 90o for Tower A33,and 0o, 30o, 56o and 90o for Tower A55. Figure 9 shows an example o the results for Tower A33.Figure 7: Latticed tower head wind tunnel models: Tower A33 (left) and Tower A55 (right)Figure 8: Reference for the wind incidence for the latticed truss transmission tower models0,00,51,01,52,02,53,03,54,04,55,05,50,00 0,10 0,20 0,30 0,40 0,50 0,60 0,70 0,80 0,90 1,00CaNBR-6123/87 Vento 0NBR-6123/87 Vento 45NBR-5422/85 Vento 0IEC-60826/91 Vento 0Tor. A33-1Tor. A33-2Tor. A33-3Tor. A33-4Tor. A33-5Tor. A33-6Tor. A33-7Tor. A33-8Tor. A33-9Tor. A33-10aTor. A33-10bTor. A33-11Tor. A33-12Tor. A33-13Tor. A33-14Tor. A33-15aTor. A33-16Tor. A33-17Tor. A33-18aTor. A33-18cFigure 9: Drag coefficients for latticed tower parts Tower A33The main findings were: The calculation procedure and the drag coefficients from the NBR-6123, NBR-5422 and IEC-60826 are adequate for the determination of the wind forces actingon the tower main body and extensions; For all others elements, the application of this methodology and coefficients is notadequate due to the great difference which exists between the geometries mostused in real towers and those in which the codes were based; For all tested elements, the decomposition in partial modules is adequate; The wind force over the legs may be calculated independently for each leg, sincethe proper shielding effect is taken into account when appropriate (function of thewind angle of incidence) on the leeward legs.THE NEW CHAPTER ON VORTEX SHEDDINGIn the current edition, the procedure to evaluate a structure sensibility to vortex shedding islimited to the determination of the critical velocity, Vcr , through the basic Equation 8, where f isthe structure natural frequency, L a characteristic dimension and St the Strouhal number for eachspecific shape. This is given in an appendix of the code.StLfVcr (8)In the new edition, a special chapter is devoted to the subject, with much moreinformation and a detailed procedure to determine the vibration amplitude, Yo, given in Equation9, and the force per unit length perpendicular to the wind direction, qy(z), given in Equation 10,which must be added to the lateral force due to atmospheric turbulence. The latter is given inanother special chapter dedicated to the dynamic response to atmospheric turbulence.oRlefcro MCCflqY1218 2121 (9) )/()()2()( 21 hzzmfYzq oy (10)In Equation 9, ef is the effective damping (structural plus/less aerodynamic), Cl is thelateral force coefficient, Mo is the equivalent mass per unit height (Equation 11), andqcr=0.613Vcr2 . The correlation coefficient CR is given by Equation 12, with = h / l1 (h is thestructure height and l1 is the dimension perpendicular to the wind incidence). The lateral forcecorrelation length, LR, is related with the amplitude of the transversal vibrations through theempirical expression given by Equation 13.dzhzdzhzzmM hho0202)/()/)(((11) 11 )/1(1lLC RR (12)1/41 1012/lYRoelL (13)After the determination of the amplitude of vibration at the top of the structure, the lateralequivalent force per unit height due to vortex shedding is given by Equation 10. The procedure isbased in the work by Paluch  and Vickery and Basu  and more detailed comments maybe found in Blessmann .CONCLUSIONSThe new improvements in the Brazilian Wind Code lead to a more comprehensive and safe code,certainly conducting to a more cost efficient and safe design.ACKNOWLEDGEMENTThe authors would like to thank FURNAS and its staff, in the person of Eng. Afonso de Oliveirae Silva.REFERENCES Associao Brasileira de Normas Tcnicas - ABNT, Rio de Janeiro. Norma Brasileira NBR-6123(NB-599): Foras devidas ao vento em edificaes. Edio 1988. J. Blessmann, The Boundary Layer Wind Tunnel of UFRGS, J. Wind Eng. Ind. Aerodyn. 10 (1982)231-248. E.A. Carpeggiani, Determinao dos Efeitos Estticos de Toro em Edifcios Altos Devidos Aodo Vento. Dissertao de Mestrado Programa de Ps-Graduao em Engenharia Civil, Escola deEngenharia, UFRGS. Porto Alegre, 2004. Deutsche Normen, Lastannahmen fr Bauten. DIN-1055. Teil 45, Seite 3. 1977. National Building Code of Canada, National Research Council of Canada, Associate Committee onthe National Building Code, Ottawa, NRCC No. 23178. 1990. N. Isyumov, M. Poole, Wind induced torque on square and rectangular building shapes, J. Wind Eng.Ind. Aerodyn. 13 (1983) 183-193. G.R. Lythe, D. Surry, Wind Induced Torsional Loads on Tall Buildings, J. Wind Eng. Ind. Aerodyn. 36(1990) 225-234. D.W. Boggs, N. Hosoya, L. Cochran, Sources of torsional wind on tall buildings: lessons from thewind tunnel, Advanced Technology in Structural Engineering (Proceedings of the 2000 Congress &Exposition), Philadelphia, May 2000, SEI/ASCE. A.M. Loredo-Souza, L.I. Rippel, M.M. Rocha, E.A. Carpeggiani, G.J.Z. Nez, A.O. Silva, J.C.P.Medeiros, Coeficientes de Arrasto Aerodinmico em Estruturas Treliadas de Linhas de Transmisso.EletroEvoluo (Rio de Janeiro), Rio de Janeiro, RJ, v. 41, p. 36-45, 2005. Associao Brasileira de Normas Tcnicas - ABNT, Rio de Janeiro. Norma Brasileira NBR-5422:Projeto de linhas areas de transmisso de energia eltrica. Edio 1985. International Electrotechnical Comission - IEC-60826: Design criteria of overhead transmissionlines. American Society of Civil Engineers, Guidelines for Electrical Transmission Line StructuralLoading. ASCE Manuals and Reports on Engineering Practice No. 74. New York. 1991. J. Blessmann, Tpicos de normas de vento. 2ed. Ed. Universidade/UFRGS. Porto Alegre, 1990. A.E. Holdo, Reynolds number effects on lattice structures forming part of a wind tunnel model. J.Wind Eng. Ind. Aerodyn. 45 (1993) 229-238. B.J. Vickery, The aeroelastic modelling of chimneys and towers. Workshop on Wind TunnelModeling,Gaithersburg, Maryland, April 14-16. Pp. IV-2. 1982. M.J. Paluch, Vibraes transversais por desprendimento de vrtices de torres e chamins deconcreto armado de seo transversal circular. Tese de Doutorado Programa de Ps-Graduao emEngenharia Civil, Escola de Engenharia, UFRGS. Porto Alegre, 1999. B.J. Vickery, R. Basu, The development of a codified approach to the determination of wind loads onchimneys. In: Canadian Wind engineering Workshop, 3. Vancouver, Canada. Paper V.1, 26 p. 1981. J. Blessmann, Introduo ao estudo das aes dinmicas do vento. Ed. Universidade/UFRGS. PortoAlegre, 1998.