Vous êtes sur la page 1sur 10

LIGHTWEIGHT STRUCTURES IN CIVIL ENGINEERING

PROCEEDINGS OF THE INTERNATIONAL SYMPOSIUM Warsaw, Poland, 24-28 June , 2002

NUMERICAL SIMULATION OF A TRANSMISSION TOWER FAILURE DUE TO MICROBURST WIND LOADS


G.A.R PARKE1, N. TOY2, P. DISNEY3 and E. SAVORY4
1 4

Reader, Professor, 3Lecturer, School of Engineering, University of Surrey, Guildford, Surrey GU2 7XH, UK Associate Professor, Department of Mechanical Engineering, University of Western Ontario, London, Canada

ABSTRACT: It has been suggested that many of the failures of transmission lines and towers are a direct result of high intensity winds (HIW) that are associated with tornadoes and microbursts. This work describes the results of a numerical simulation for the wind and loading time histories of a transient microburst as it passes a tower/line arrangement. In addition, a finite element analysis has also been performed on this tower for this event. Although the microburst did not produce failure of the tower, it did highlight the increase in loading suffered by the tower. It is anticipated that for failure to occur, the microburst is likely to be much more broad-fronted and produce larger contributions on the conductor loads, as well as providing some contributions from internal fluctuations that have frequencies close to the natural frequency of the structure. Key Words: Numerical simulation, transmission tower, microburst, wind loading 1. INTRODUCTION One of the most commonly used lightweight structures is that of the transmission tower whereby conventional geometries and conductor arrangements have been designed according to the many design codes and guides currently available, Refs 1 5. For these types of structures, not only is the dead-weight of the tower and conductors considered in the structural analysis but also the loading due to the wind. In this case, the mean wind loads on the tower are supplemented by the quasi-steady dynamic loads associated with the wind on the conductor lines themselves. Unfortunately, our understanding of the behaviour of transmission lines under the action of wind loading is still incomplete, Ref 6, especially for the off-design conditions associated with the action of High Intensity Winds [HIW], Ref 7. These particular HIWs are unpredictable, are often very localized, and are associated with downbursts and microbursts, as shown in Fig 1, as well as with winds of a tornadic nature. Figure 1 provides clear evidence of the production and dissipation of a microburst after touchdown, in which the resulting wall jet has disturbed and ingested dust particles into its structure. The occurrence of high intensity winds can be costly, and it has been reported that many of the utility organizations in the Americas, Australasia and South Africa have provided evidence that between 80 100% of all transmission tower failures were due to these types of wind. For normal wind conditions, even at high wind speeds, the design engineer determines both the wind load on the tower and, more importantly, the weight of the transmission conductors and the wind load on them. From these load sources, the resultant transverse forces may be determined allowing the tower to be designed such that the internal face members act to brace the structure but do not take significant loads themselves. In the case of HIWs, the resulting wind profile is far from uniform and may act such that the line of action of the wind loads may produce radically off-design conditions, thereby initiating collapse of the members of the windward faces of the tower. Such behaviour is unaccounted for by present design procedures and therefore new methodologies need to be developed to understand the loading mechanism associated with HIWs. At present, some numerical models are being developed in which data relating to a particular intensity and scale of a HIW and how it affects a specified length of conductor has been obtained from HIW occurrence and damage. Many of these models are based upon risk assessment for particular regions and for either tornadic winds, Refs 8, 9, or for downbursts, Refs 10 13. However, a number of questions relating to HIWs need to be resolved, for example, What are the time-histories of the wind profiles of a HIW? What are the associated loading distributions? How do these loading histories cause failure of the towers? What are the failure modes?

Fig 1 Although there is little data available concerning the make-up of HIWs, it may typically be described, from observation, as a rapid and intense downburst of air, Fig 2, forming a vortex ring of less than 4 km diameter that interacts with the ground and quickly spreads radially outwards, Ref 14. This process takes place within about 2-5 minutes, with the highest lateral wind speed appearing to be some 80 100 m above ground level directly below the ring vortex core. From these observations, a simplified numerical model has been developed for the loading time histories induced by a microburst on a transmission line. The generated data was then used as the input to a finite element structural analysis

CHAPTER 3 - Towers and masts

157

model in order to predict the expected failure modes and then to compare these modes with those observed in the field.

Although there are a number of variations of how and where the microburst will impose loadings on a transmission tower, the case taken here represents the condition where the microburst has already reached the ground and is spreading out as a wall jet, and therefore the main loading components will be in the horizontal plane. This model has been adopted by previous workers, Refs 16, 17, and allows analysis of the horizontal wind speed vector at a specific location to be determined. In order to allow for different horizontal speeds, a vertical profile of the horizontal mean radial wind speed (U) is given as a function of different heights (Z) above the ground, Eqn 1, and shown graphically in Fig 4.

U =e U max

0.15

Z Z max

3.2175

Z Z max (1)

Where Umax is the maximum horizontal radial velocity within the microburst that occurs at a height Zmax above the ground (with Zmax being typically about 60 m). Fig 2 2. THE MICROBURST MODEL In the case of a downburst, two forms of a simplified model for the wind field may be considered, namely an impinging jet model, Refs 15 17, and a ring vortex model, Refs 15, 18, 19. In the first of these models, as depicted in Fig 1 and shown schematically in Fig 3, there is a radial outflow of air from the touch down position thereby providing a wall jet that dominates the wind field. The second type is described as a vortex ring since the descending column of air induces this vortex as it approaches the ground. However, laboratory simulations using large scale impinging jets have provided better agreement with some available full-scale data, Ref 20, 21, and therefore, the impinging jet model has been used in the following analysis.

5.0
Z / Zmax Height above ground

4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.0 0.2 0.4 0.6 0.8 1.0
U / Umax = e -0.15 Z/Zmax - e -3.2175 Z/Zmax

(a) Wall Jet Model

Horizontal velocity component U / Umax


Fig 4

In order to determine microburst induced loads an earlier mathematical model has been adopted, Refs 22, 23, that account for both the static drag forces and the inertia forces due to accelerations in the wind as the flow field passes the obstruction, in this case, a transmission tower. In this case, the overall force per unit height of the tower may be given, at any given time, F(t), by Eqn 2,

F (t ) =
(b) Vortex Ring Model Fig 3

1 1 dU C d DU U + C m D 2 2 4 dt

(2)

PART I Structures

158

where Cd and Cm are the drag and inertia coefficients respectively, is the air density, U is the horizontal velocity component, and D is the projected width of the tower arrangement. However, some small modifications have been introduced to this model that allows use of microburst wind field data as an input. In addition, since the microburst is only being considered as a wall jet, the vertical component of velocity and the spatial and temporal rates of change of this component are ignored, together with any smaller scale wind turbulence. In this case, the spatial and temporal rates of change of the horizontal wind speed component are effectively the mean windspeed gradients.

mass damping has also been incorporated into the numerical model in order to damp out an initial vibration that was found to occur as a result of the application of a small step load at the start of the time history. Since it was impractical to determine the wind loads in each individual member, the tower was divided into six vertical sections, approximately equal in height as shown in Fig 6, and the overall loadings on these "patches" computed in the longitudinal or x-direction (normal to the conductors) and the transverse or y-direction (parallel to the conductors and including the cross-arms of the tower). These loadings were arrived at by considering a drag coefficient of 3.0 for the area of one face of the tower, as specified in Fig 4.3(a) of the UK Code of Practice for a lattice tower with a solidity ratio of 18%, Ref 26, taking into account the shielding of the leeward members by the windward ones. In addition, since a tower is primarily constructed of right-angle members, a drag coefficient of 2.0 could be assigned to each, Ref 26, 27, thereby providing evidence that the leeward members would take about 1/3rd of the total load. Furthermore, the inertia coefficient was considered to be unity for this present model since previous work, Ref 23, had suggested that a value close to unity is appropriate for towers under dynamic wind loading conditions. The time history of the X and Y components of the wind load has been introduced into the model using the AMPLITUDE option in ABAQUS. At every time increment the incident wind load at each section has been applied at the corner nodes of the horizontal plan bracing existing at that level, as illustrated in Fig 6. This allows each front node to receive 1/3rd of the load acting at that level and each back node 1/6th of the load.

3. DETAILS OF THE MODELLED TRANSMISSION TOWER For the purpose of this analysis a CEGB type Blaw Knox L6 standard height lattice tower, Ref 24, has been adopted as a generic form of lattice tower. This particular tower has a height of 50.5 m with a square base area of 9.1 x 9.1 m, as illustrated in Fig 5. The conductors are approximately 30 m above ground level with a typical span of 340 m between adjacent towers. The tower has been modelled using ABAQUS three-dimensional beam elements type B31, with a relatively small value for the moment of inertia for these elements. This allows the bending and torsional stiffness to be close to zero for this condition, and simulates pinned connections for the structural model that is close to the real connection response exhibited in the actual structure, Ref 25. A full post-buckling response is included and was achieved by using low bending stiffness and the true axial stiffness thereby modelling the tower members as truss elements. In addition, the self-weight of the conductors and tower, together with the restraining forces exerted by the conductors are included in the analysis.

Fig 6

4. DETAILS OF THE MICROBURST MODEL It is known that wind speeds of up to 70 m/s have been recorded in microbursts and it is evident that maximum speeds may be even higher in unrecorded events. This being the case, a radial speed of 80 m/s was adopted for the present work, similar to that used by other researchers, Ref 16. In addition, a translational speed for the event of 20 m/s has also been chosen as being broadly representative of field data. In order to provide a representative shape for the radial velocity profile a characteristic radial length scale of 350 m was adopted. This being about half the core radius of 650m which is effectively the distance between the centre of the core region to where the barometric pressure is low within the ring vortex. A value that is similar in magnitude to that noted in real events, Ref 14. The transient nature of a microburst is such that its lifetime may be a short as 5 minutes or less once it has touched down, Ref 14. If one assumes that once this occurs then the radial velocity reduces to some 20% of its original maximum value, then a time scale for the above microburst model would be of the order of 200 secs. This assumption would allow a radial length scale and time scale for the model to be comparable to the correct radial velocity profile and decay rate as observed in nature. In addition to these spatial and temporal scales a position of 60 m above the ground was taken for the location of the maximum radial velocity, a value that has been determined by other workers, Ref 14, 15. The dynamic loading was determined for a microburst that touches down some 1.5 km upstream of the transmission

Fig 5 The numerical model adopted is an implicit, incremental, direct integration approach that allows modelling of the tower inertia forces due to microburst wind loads, thereby allowing an investigation of the possibility of a dynamic failure in the structure. This direct integration method provides numerical stability when non-linear responses are expected. For dynamically sensitive structures, such as the tower model (particularly under dynamic wind loading), a considerable amount of time is needed for each single analysis. Furthermore, a small amount of

CHAPTER 3 - Towers and masts

159

line, and which is convected along a path normal to the conductors with the microburst centre at a distance of 100 m from the nearest tower. This gives a configuration such that the windward face of the tower is normal to the microbursts path as shown in Fig 7. The time interval between successive calculations was set at 1 second time steps (giving step intervals of 20 m for the horizontal translation of the microburst).

80
U component V component

60
Height = 43.7m Core radius = 650m Max. rad. vel. = 80m/s Ambient vel. = 20m/s

Line Core rad. = 650m Max. radial vel. = 80m/s Tower detail X Y

Velocity

Tower

(m/s)

Microburst path normal to conductors Translational velocity = 20 m/s

40

20

100 m 1500 m (starting position)

-20

-40 0
V component of velocity Fy component of force

30 60 Time

90 120 150 180 (seconds)

Fig 8

U component of velocity Fx component of force


Fig 7

80
X component (Fx) Y component (Fy)

(kN) Total load on tower section 1

60

5. PREDICTED VALUES OF THE WIND FIELDS AND LOADS The numerical prediction for the U and V components of the velocity time history for the above microburst, occurring at the mid section of section 1 of the tower, is shown in Fig 8. The predicted results have been determined from the time of touch down of the microburst and indicate the effect of the decay of the flow field as it passes the tower. It may be observed that the event lasts for approximately 140 seconds with a normal component of velocity (U) reaching a peak of 70 m/s, some 40 seconds after the microburst has touched down. Since a microburst has little rotational speed, it is not unexpected that the lateral velocity component (V) is predicted as being rather weak. An analysis of the loading that is generated by these velocity components on this section 1 of the tower provides an insight into the possible loading time history as the event passes the tower. This highest section includes the top pair of cross-arms, and the two components of the forces acting on this section are shown in Fig 9. Here it may be seen that the normal force (Fx ) persists for about 60 seconds, reaching a peak of approximately 75 kN. Well in excess of the force due to ambient wind conditions. The lateral force acting on the tower section shows, as expected, little variation as the microburst passes the tower, which is consistent with the low value of the lateral velocity as shown in Fig 8.

40

Height = 43.7m Core radius = 650m Max. rad. vel. = 80m/s Ambient vel. = 20m/s

20

-20 0 30 60 Time 90 120 150 180 (seconds)

Fig 9 6. STRUCTURAL ANALYSIS The finite element analysis of the response of the tower has been achieved for a horizontal wind load that was produced by a microburst. The results of this analysis are shown in Fig 10. Here it may be

observed that the displacement of the top of the tower in the x-direction reaches a maximum at the same time as the maximum wind load, and that the displacement remains proportional to the applied loading

PART I Structures

160

throughout the time history. This implies that no significant nonlinearity has occurred for this structure and that for this type of loading history, a quasi-static response appears dominant. However, this analysis has been based upon simple conditions for a particular tower geometry, and it is likely that failure of transmission towers is influenced by other factors, such as, smaller scale fluctuations within the microburst, or a broader fronted, higher magnitude loading pattern that encompasses significant loading contributions from the conductors, and the geometric properties of the tower. These aspects remain the subject of future investigations.

30

Load (x 10 kN) and Displacement (cm)

25

Total wind load (Fx) X direction displacement at top of tower

20

15

10

0 0 10 20 Time
Fig 10

30 40 50 (seconds)

60

7. CONCLUSIONS This preliminary investigation has provided us with a platform for studying the loading and response of HIWs on lightweight structures. In particular, we have looked at a specific structural form, namely that of a transmission tower. Although HIWs may be either of the tornadic form or that of the microburst phenomena, we have concentrated, at present, on the wind effects and structural analysis on such a tower subjected to a microburst. These predictions have shown that such a tower is subjected to large forces over a short time interval, and even though this analysis has been based upon many simplifications, it is evident that with more detailed parameters as inputs, a full dynamic structural analysis could be entertained. Although the results are encouraging, in as much that they provide a basis from which this work may evolve, a more detailed parametric study of the response of transmission towers to tornado and microbursts is needed. To this end, it is anticipated that an examination of the effects of scale and intensity of the HIW will be studied along with the path of the event and its proximity to a tower/line arrangement. Future interest will also be directed to studying these HIWs on different types of lightweight structures including other tower geometries, and in particular, the guyed mast type that are common in many countries where HIWs are a regular occurrence.

8. ACKNOWLEDGEMENTS The authors are indebted to the Engineering and Technology Division of The National Grid Company PIc, United Kingdom, for supporting participation in the International Task Force Committee on HIW on Transmission Lines.

9. REFERENCES [1] Engineering Sciences Data Unit (ESDU). Lattice structures Part 1: mean fluid forces on single and multiple plane frames Wind Engineering SubSeries, Data Item 81027 , London, 1988. [2] Engineering Sciences Data Unit (ESDU). Lattice structures Part 2: mean fluid forces on tower-like space frames. Wind Engineering Sub-Series, Data Item 81028, London 1988. [3] International Electrotechnica1 Commission (IEC), Loading tests on overhead line towers. Report 60652, I st ed., Technica1 committee 11' Overhead Lines, 1979. [4] International E1ectrotechnical Commission (IEC). Loading and strength of overhead transmission lines. Report 60826 TRO, 2nd ed., Technical Committee II. Overhead Lines, 1991. [5] American Society of Civil Engineers (ASCE). Guidelines for electrical transmission line structural loading. ASCE Manuals and Reports on Engineering Practice, No.74, New York, 1991, [6] Loredo-Souza AM, Davenport AG. Wind tunnel modelling of transmission lines, In: Proc. 3rd Int, ColI. on Bluff Body Aerodynamics and Applications. Virginia, USA Blacksburg, July 1996, Paper No B XI 9-B XI 12. [7] Dempsey D, White HB, Winds wreak havoc on lines, Transmission and Distribution World, June 1996:32-42. [8] Milford RV, Goliger AM. Tornado risk model for transmission line design. J Wind Engng and Ind Aero 1997;72:469-78. [9] Lezaola J, de Schwarzkopf MLA, Rosso LC, Carstairs D. The influence of severe meteorological events on EHV conductor selection. Paper at University of Western Ontario Workshop on Control of Damage to Transmission Lines due to HIW, London, Ontario, Canada, October 1997. [10] Krishnasamy SG, Assessment of weather induced transmission line ]oads on a probabilistic basis. IEEE Trans Power App and Syst 1985;104:25]0-6. [ 11] McMahon B. Reliability and maintenance practices for Australian and New Zealand HV transmission lines. In' Proc. 2nd Int. Conf. on Reliability of Transmission and Distribution Equipment, Coventry, UK, March] 995: 198203. [12] Oliver SE, Moriarty WW, Holmes JD. A severe thunderstorm risk model for transmission line design: Queensland and New South Wales. Australian Bureau of Meteorology Special Services Unit, ESAA Report No.4, July 1996. [13] Holmes JD. Modelling of extreme thunderstorm winds for wind loading of structures and risk assessment. In: Larson A, Larose GL, Livesey FM, editors. Wind engineering into the 21 st century. Rotterdam. Balkema, 1999.1409-15. [14] Fujita TT. Downbursts: meteorological features and wind field characteristics. J Wind Engng and Ind Aero 1990;36.75-86. [15] Vicroy DD. Assessment of microbur.5t models for downdraft estimation. J Aircraft 1992;29.1043-8. [16] Holmes JD, Oliver SE. A model of downburst winds near the ground for transmission line loading. CSIRO Div. ('f Building, Construction and Engineering, Australia. Report DBCE Doc 96/3 (M), J anuary 1996. [17] Holmes JD, Oliver SE. An empirical model of a downburst. Engng Struct 2000;22(9).1167-72. [18] Zhu S, Etkin B. Model of the wind field in a downburst. J Aircraft 1985;22.595-601. [19] Ivan M. A ring-vortex downburst model for flight simulations. J Aircraft 1986;23.232-6. [20] Letchford CW, IIlidge G. Turbulence and topographic effects in simulated thunderstorm downdrafts by wind tunnel jet. In. Larson A, Larose GL, Livesey FM, editors. Wind engineering into the 21st century. Rotterdam: Balkema, 1999:1907-12. [21] Wood GS, Kwok KCS, Motteram NA, Fletcher DF. Physical and numerical mode1ling of thunderstorm downdraft. In. Larson A, Larose GL, Livesey FM, editors. Wind engineering into the 21st century. Rotterdam: Balkema, 1999.1919-24. [22] Wen Y -K, Ang AHS. Tornado risk and wind effects on structures. In. Proc. 4th Int. Conf. on Wind Effects on Buildings and Structures, London, 1975:63-74. [23] Wen Y -K. Dynamic tornadic wind loads on tall buildings. J Struct Div ASCE STl 1975;101.169-85. [24] Lomas C. Transmission tower development in the UK. Engng Struct 1993;15(4)"277-88. [25] ECCS Recommendations for angles in lattice transmission tow- ers, European Convention For Constructional Steelwork. Technical Committee 8 Structural Stability, Technical Working Group 8.1 ~ Components, Report No.39, 1985. [26] British Standards. Lattice towers and masts. Codes of Practice for Loading, BS 8100: Part 1, London, 1986. [27] Scruton C. An introduction to wind effects on structures, Engineering Design Guide No.40. UK. Oxford University Press, 1981.

CHAPTER 3 - Towers and masts

161

LIGHTWEIGHT STRUCTURES IN CIVIL ENGINEERING


PROCEEDINGS OF THE INTERNATIONAL SYMPOSIUM Warsaw, Poland, 24-28 June , 2002

AN EVALUATION OF STRUCTURAL STEEL DESIGN SYSTEMS FOR TRANSMISSION AND TELECOMMUNICATION TOWERS
J. G. S. da Silva1, P. C. G. da S. Vellasco2, S. A. L. de Andrade3, M. I. R. de Oliveira4
3

Associate Professor, Mechanical Engineering Department, State University of Rio de Janeiro, BRAZIL 2 Associate Professor, Structural Engineering Department, State University of Rio de Janeiro, BRAZIL Associate Professor, Civil Engineering Department, Pontifical Catholic University of Rio de Janeiro, and Structural Engineering Department, State University of Rio de Janeiro, BRAZIL 4 Undergraduate Student, Structural Engineering Department, State University of Rio de Janeiro, BRAZIL

ABSTRACT: The traditional methods of structural analysis involved in the design of steel telecommunication and transmission towers tends to assume simple truss behaviour. In this structural solution all the steel element connections are considered as simple or hinged. Despite this fact, the most commonly used tower geometries possess structural mechanisms that could compromise the assumed structural behaviour. A possible explanation for the structure stability is related to the actual behaviour of semi-rigid connections instead of the simple connections assumed in the structural design. Previous investigations on a 40m high steel tower showed that maximum stresses and displacements for the structural modelling based on the two investigated methodologies (simple truss element and combined beam and truss element modelling), lead to similar results. This paper describes further comparisons of the two above mentioned design methods for an existing 75m high steel telecommunication tower. The authors intention is to further investigate if the previous differences in results still apply for slender telecommunication towers. Keywords: Telecommunication and transmission towers, steel tower design, static and dynamical behaviour. 1. INTRODUCTION The steel transmission and telecommunication tower design is not a straightforward process, but an interactive compromise between many factors, which must ultimately satisfy basic strength requirements. The traditional methods of structural analysis involved in the design of steel telecommunication and transmission towers tends to assume simple truss behaviour. In this structural solution all the steel element connections are considered as hinged. Despite this fact, the most commonly used tower geometries presents structural mechanisms that can compromise the assumed structural behaviour. A usual solution to this problem is the use of fictitious bars in the structure to prevent the occurrence of the unwanted degrees of freedom. A possible explanation for the structures stability is related to the behaviour of semi-rigid connections instead of the simple connections assumed in the structural design. Previous investigations, Refs 1,2,3, on a 40m high steel tower showed that maximum stresses and displacements for the structural modelling based on the two investigated methodologies (simple truss element and combined beam and truss element modelling), lead to similar results. However, the simple truss method substantially increases the amount of work to model the structure and generates a potential error source if the rigidities and/or number of spurious bars were not properly considered. This fact conducted to the use of a methodology in which the main bars were modelled as spatial beam finite elements while the bracing system are spatial truss finite elements. This paper describes further comparisons of the two above mentioned design methods for an existing 75m high steel telecommunication tower. The authors intention is to further investigate if the previous differences in results still apply for slender telecommunication towers. The comparison is based on a parametric study of the tower geometry in order to obtain the structures response in terms of displacements and stresses associated to the tower ultimate and serviceability limit states. Both structural models, in first order elastic analysis and also second order analysis, are investigated in terms of their qualitative and quantitative behaviour focusing on the actual structural safety provisions. The design of steel transmission and telecommunication towers in this slenderness range is very competitive market aiming on lower global costs and higher quality issues. Fig.1 Tower Geometry Layout. 2. THE STRUCTURAL MODELLING The steel telecommunication tower investigated in the present paper uses a truss type geometry configuration. It should be pointed out that the present work is based on a real structure, designed, fabricated and erected by a practising company, Ref 4. Actual member properties were used in this analysis. The structure possessed a square cross section divided into two segments: a pyramidal lower part and a prismatic upper part. Rolled angle sections connected by bolts composed the main structure. The imposed load consists of a standard, together with associated waveguides, cabling, etc. The staircase is arranged vertically in the tower centre being intermediately supported by the main platforms and by additional bracings running approximately across the tower diagonal. The working platforms, stairs and carriers where positioned inside the tower structure. The geometry configuration of the simply supported tower structural system, depicted in Fig 1, presented a total height of 75m.

PART I Structures

162

This investigation considered as acting vertical loads: structure selfweight, stairs, internal platforms, vertical carriers, antennas, cables and etc. The wind effects over the steel tower were the main horizontal loads considered in the structural analysis. Several investigations have been conducted in order to access the best strategy to model the steel tower, Refs 1,2,3,5,6,7,8,9. The methodology proposed in the present investigation uses truss and beam threedimensional finite elements. The main purpose of the adopted system was the prevention of spurious structural mechanisms that could lead to an uneconomic or even an unsafe structure. Truss Element Traditional Modelling. The steel tower modelling presented problems, due to the loss of structural continuity in some parts of the structure, related to the presence of the hinges associated with the used spatial truss finite element. The diagonal angles were connected to the leg angles by a three-bolt connection (single row) leading to a semi-rigid behaviour. Using truss elements to model the angles, a loss of moment continuity occurs, since the truss connections are flexible. On the other hand, by using only one bolt in the horizontal vs. leg angle connections, another discontinuity appears. Although the in-plane behaviour can be considered flexible the out-of-plane behaviour disregards the torsional and flexional continuity present in the structure. To overcome this problem, dummy bars, with associated low axial stiffness values, are introduced into the structure. Every new bar represents the suppression of an internal degree of freedom. Diamond shape horizontal sets of bars are introduced along the tower height, Fig 2. The diamond bar set restricts the displacements of the mid-side (highlighted) nodes. These nodes shown in the vertical diamond shape set of bars are present in every structures face, Fig 3, and if not restrained can collapse by means of a simple structural mechanism. Another reason for using these bars is to improve the structures torsional rigidity. Finally an extra dummy bar, positioned inside the horizontal diamond shape set of bars, Fig 2, is used to create a typical truss configuration system preventing another possible in-plane horizontal mechanism.

Combined Beam and Truss Element Modelling. Based on an extensive parametric investigation, Refs 1,2,3, a modelling strategy combining three-dimensional beam and truss finite elements was proposed. In this methodology the main structure uses beam elements while the bracing system utilises truss elements. This method models the structure as a static determined system discarding the need of dummy bars present or in the traditional analysis. The used beam finite elements presented seven degrees of freedom per node, six associated with translation and rotation displacements in space and the last representing the warping degree of freedom. 3. RESULT ANALYSIS AND DISCUSSION Table 1 and 2 present the results of a linear static analysis of the investigated steel telecommunication tower according to the two abovementioned modelling strategies. Maximum values of stresses and horizontal displacements are presented and compared. The acting load considered in the analysis where self-weight and two wind load cases (perpendicular and diagonal to the tower face). Tower Structure Maximum Stress. Self-Weight and Wind Loads. Simply Supported Conditions. Modelling Strategies: I - Truss Element II - Combined Beam and Truss Element Perpendicular Wind Diagonal Wind Direction Modelling Vmax. (MPa) I II Difference 100. (Compression) 190. (Compression) 47.4% Vmax. (MPa) 180. (Compression) 290. (Compression) 37.9%

Table 1. Structure Maximum Stress. Tower Structure Maximum Horizontal Displacements. Self-Weight and Wind Loads. Simply Supported Conditions. Modelling Strategies: I - Truss Element II - Combined Beam and Truss Element Perpendicular Wind Diagonal Wind Direction Modelling umax. (mm) I II Difference 465. 455. 2.2% umax. (mm) 600. 585. 2.7%

Section AA Fig.2 Dummy Bars Used in the Structures Transversal Plan.

Table 2. Structure Maximum Displacements. The difference between the maximum stress values obtained from the simple truss traditional method and combined beam and truss element modelling is about 47.4% of the value obtained from the second modelling. When quantitative analysis of the data is performed it is possible to confirm that the maximum stress values, Table 1, were significantly modified. However, the lateral displacements values, Table 2, were not significantly changed when the combined beam and truss model or the simple truss model were considered. The maximum stress value of 290MPa results from these three contributions: 2MPa from gravitational loading, 145MPa from wind load axial stress and 143MPa from wind load bending stress. These results can be factored for resistance checks, but undoubtedly the structure index criterion is modified and lower safety is expected. Table 3 presents the horizontal displacements of a series of nodes situated in the studied steel tower edge. An expected increase in lateral nodal displacement is clearly noticed in this table as the tower top nodes are considered.

Fig.3 Dummy Bars Used in the Structure Vertical Plan.

CHAPTER 3 - Towers and masts

163

Steel Tower Horizontal Displacements. Self-Weight and Wind Loads. Simply Supported Conditions. Modelling Strategies: I - Truss Element II - Combined Beam and Truss Element Perpendicular Wind Height Nodes (m) I II Diagonal Wind I II

umax. (mm) umax. (mm) umax. (mm) umax. (mm) 14 136 186 214 254 300 344 368 392 416 438 460 464 468 10 19 23.5 29 34 39 44.8 50.4 55.6 60 63 67 71 75 12. 30. 42. 58. 76. 100. 135. 177. 223. 269. 304. 355. 410. 465. 12. 30. 41. 58. 75. 98. 132. 172. 217. 261. 295. 345. 398. 455. 17. 40. 63. 76. 99. 129. 174. 228. 288. 347. 392. 459. 530. 600. 17. 40. 61. 76. 98. 127. 171. 222. 280. 337. 380. 445. 515. 585. Steel Tower Moments. Diagonal Wind Direction. Self-Weight and Wind Loads. Simply Supported Conditions. Bars Vertical 7 Nodes Height (m) My (kNm) Mz (kNm) Fig.5 Steel Tower Nodal Stresses (N/m2). Diagonal Wind Direction. Table 4 presents the moments found in the node connections of a series of bars situated in studied steel telecommunication tower, considering a modelling strategy combining three-dimensional beam and truss finite elements.

Table 3. Steel Tower Horizontal Displacement Profile. The maximum stress present in both analysis ranged approximately around 290MPa (in compression) for a bar situated close to the tower base associated with a diagonal wind load case. The maximum lateral displacement reached 600mm for a node situated in the towers top, Tables 1 and 2. Figure 4 and 5 illustrate a deformed simply supported tower configuration and stress contours for a combined beam and truss element strategy associated to a diagonal wind load case.

8 -0.20 0.06 0.0 9 -16.00 0.01 14 11.00 0.02 Vertical 329 10.0 133 -10.90 0.00 136 1.38 -0.00 Vertical 476 19.0 184 -1.05 0.00 185 -11.50 0.00 Vertical 478 22.0 186 -0.00 19.85 214 6.77 -0.00 Vertical 676 29.0 252 -8.84 0.00 254 8.71 0.00 Vertical 794 34.0 298 -8.85 0.00 299 -1.12 0.00 Vertical 796 37.33 300 3.76 -0.00 332 4.22 -0.00 Vertical 931 41.9 343 -4.58 0.00 344 2.05 0.00 Vertical 963 44.8 355 -2.92 -0.00 368 2.00 -0.00 Vertical 1027 50.4 379 -2.80 -0.00 392 1.30 -0.00 Vertical 1091 55.6 403 -1.48 0.00 416 0.81 -0.00 Vertical 1155 60.0 427 -1.15 0.00 461 0.09 -0.13 Vertical 1222 68.0 462 0.03 0.13 467 0.05 -0.14 Vertical 1228 74.0 468 -0.00 0.10 Table 4. Moments on the Steel Tower Connections. The maximum moment node connection present in the analysis ranged approximately around 20kNm for a bar situated near the tower base associated with a diagonal wind load case. Figure 6 illustrates the steel tower moment contours for a combined beam and truss element strategy associated to a diagonal wind load case.

Fig.4 Steel Tower Horizontal Displacements (m). Diagonal Wind Direction.

PART I Structures

164

Most designers neglect these moments when using the traditional methods of structural analysis involved in the design of steel telecommunication and transmission towers as simple truss behaviour is assumed.

of a number of dummy bars to prevent the occurrence of structural mechanisms. This fact increases the amount of work to model the structure and generates a potential error source if the rigidities and/or number of spurious bars were not properly considered. A quantitative analysis enabled the determination of the maximum stresses found in the analysis, for a 75m high steel telecommunication tower, reached around 290MPa in compression for a bar located near the tower base. This value corresponded to 82% of the steel yield stress used in the structures bars. The maximum moment found in the node connections was 19.85 kNm for a node near the structures foundation. These were the main reasons for the recommendation of the methodology in which the main structure uses spatial beam finite elements while the bracing system utilises spatial truss finite elements. Further studies will contemplate the dynamical and stability non-linear effects, due to nature of the imposed loads, on such tower structures. 5. REFERENCES 1. M.N. Policani: Comportamento estrutural de torres metlicas, MSc Dissertation, Engineering science laboratory, UENF, in portuguese, Brazil, 2000. 2. M.N. Policani, J.G.S. da Silva, L.F. Estrella Jnior, P.C.G. da S. Vellasco, S.A.L. de Andrade: Structural assessment of steel telecommunication towers, International conference on steel structures of the 2000's, pp. 251-256, Istanbul, 2000. 3. J.G.S. da Silva, P.C.G. da S. Vellasco, S. A. L. de Andrade, L.F. Estrella Jnior, M.N. Policani: Comportamento estrutural de torres metlicas de telecomunicaes, 21th Iberian latin american congress of engineering computational methods, pp.1-10, CD ROM, in Portuguese, Rio de Janeiro, 2000. 4. Communications Tower, Case Studies: Volume Four, Imperial College of Science, Technology and Medicine, Civil and Environmental Engineering Department, Internal report, UK, 1985. 5. Y. Hui, L. Yanjun, Z. Diansheng: Geometric non-linear analysis of transmission tower with continuous legs, Advances in steel structures, Proceedings of international conference on advances in steel structures, pp 339-344, Hong Kong, 1996. 6. R. Saxena, N. Popplewell, P.G.S. Trainor, A.H. Shah: Vibrations of complex guyed towers, Proceedings of the 12th biennial conference on mechanical vibration and noise control, Montreal, 1989. 7. Y.M.F. Wahba, M.K.S. Madugula, G.R.Monforton: Evaluation of non-linear analysis of guyed antenna towers, Computer & Structures, Vol. 68, pp 207-212, 1998. 8. K. Natarajan, A.R. Santhakumar: Reliability-based optimization of transmission line towers, Computer & Structures, Vol. 55, N0 3, pp 387-403, 1995. 9. G. Visweswara Rao: Optimum designs for transmission line towers, Computer & Structures, Vol. 57, N0 1, pp 85-92, 1995.

Fig.6 Moments on the Steel Tower Connections (Nm). Diagonal Wind Direction. 4. FINAL REMARKS This paper proposes a new structural analysis modelling strategy, based on qualitative and quantitative comparisons, for transmission and telecommunications steel towers. The proposed methodology, less conservative than traditional methods of analysis, uses a combined solution of three-dimensional beam and truss finite element to model the structural behaviour of spatial tower structures under several loading and support conditions. These promising results enable the following conclusions: The structural analysis methodology, substantiated by the finite element method, has shown coherent results when the global structural behaviour was considered. Despite these results, it become clear to the authors that the continuation of this research has to deal with case studies and comparisons with experiments and/or field measurements of steel towers found in practice. Generally in all the studied cases the maximum stresses values for the structural tower modelling based on the two investigated methodologies were significantly modified. The lateral displacements values were not significantly changed when the combined beam and truss model or the simple truss model were considered. On the other hand, based on the difficulties found in the analysed steel towers, present in current engineering design practice, and corroborated by the nature of the spatial truss finite element, an analysis based only on this element cannot be advisable. This method also implies in the use

) J.G.S. da Silva, Mechanical Engineering Department, State University of Rio de Janeiro, UERJ, BRAZIL ) P.C.G. da S. Vellasco, Structural Engineering Department, State University of Rio de Janeiro, UERJ, BRAZIL 3 ) S.A.L. de Andrade, Civil Engineering Department, Pontifical Catholic University of Rio de Janeiro, PUC-Rio and Structural Engineering Department, State University of Rio de Janeiro, UERJ, BRAZIL 4 ) M.I.R. de Oliveira, Undergraduate Student, Structural Engineering Department, State University of Rio de Janeiro, UERJ, BRAZIL
2

CHAPTER 3 - Towers and masts

165

PART I Structures

166

Vous aimerez peut-être aussi