Interference Excitation of Tall Buildings | Wind Engineering

In a typical city environment where many tall buildings and structures are grouped closely together, the proximity of one building to another can influence the flow field around and the wind forces acting on the buildings. Such interference effects are particularly marked for a group of identical structures such as a line of chimneys.

The interactions of neighbouring tall buildings, commonly referred to as interference or proximity effects, can cause a significant increase in wind loads and exacerbate the problem of occupant perception of motions due to enhanced dynamic vibrations.

The alongwind, crosswind and torsional responses of both upstream and downstream buildings in a group may be affected for wind from all directions. It is evident the owner of a new building may be held responsible for any adverse wind effect on existing buildings, adding a new dimension in litigation and professional liability.

The effects of interference from adjacent buildings, existing or proposed, are now routinely studied during wind tunnel model testing of tall buildings and structures. A wealth of useful information has been collected through these test but rarely is this information widely available because of the proprietary nature of those tests. Nevertheless, significant advances in the understanding of the various interference excitation mechanisms and their potentials in causing significant increases in the response of tall buildings have been achieved.

Alongwind and crosswind responses of tall buildings under interference excitations:

The effects of interference on the alongwind and crosswind responses of tall buildings have been investigated by Melbourne and Sharp (1976), Saunders and Melbourne (1980), Ruscheweyh (1980), Saunders (1980), Sykes (1983), Bailey and Kwok (1984, 1985), Blessmann (1985), Kareem (1987), Kwok (1989), Taniike (1991) and Yahyai et al (1992) using aeroelastic modelling techniques.

These tests were undertaken in turbulent boundary layer wind tunnels in simulated wind flows ranging from that over flat open country terrain to highly turbulent urban terrain. Most of the tall building models studied, conveniently referred to as the principal buildings, were of square or rectangular cross- sections with height to breadth ratio ranging from 2 to 9.

The tests covered a full range of reduced wind velocities from 2 to 14. In addition to having an identical interfering building upstream, the effects of a downstream interfering building were also investigated by Bailey and Kwok (1984, 1985), Kwok (1989) and Yahyai et al (1992).

Saunders and Melbourne (1980), Saunders (1980), and Kareem (1987) also tested configurations of a pair of upstream interfering buildings. Other variables studied include interfering building height, shape, and size.

The increased alongwind and crosswind responses of a square cross-section tall building with a height to breadth ratio of 9, the principal building, under interference excitation and proximity effects of a neighbouring building of the same height in open country, suburban and urban terrains were investigated in a series of wind tunnel model tests conducted by Bailey and Kwok (1984, 1985) and Kwok (1989). The interference excitation mechanisms were also studied in detail.

The interference effects are usually expressed in the form of contours of buffeting factor (BF) defined as follows:

The buffeting factor- contours shown in Fig. 1 for the principal building under interference excitation by an identical building in an open country terrain and at a reduced wind velocity of 6 show large increases in alongwind and crosswind responses of up to around 80% for an upstream interference situation and more than 400% in the alongwind response when the interfering building was in a critical downwind position.

A reduced velocity of 6 would be representative of a 5 to 10 year return period design wind speed for serviceability and occupant comfort consideration for typical tall buildings. The higher values of both the alongwind and crosswind buffeting factors span out diagonally from the principal building suggesting the excitation forces intensify near the edge of the wake of the upstream building. Similar distribution of the buffeting factors is evident in the results obtained by Yahyai (1992).

The interference excitation mechanisms can be studied by examining the force and wake spectra of the principal building. The alongwind and crosswind force spectra of the principal building at two critical interference positions near the edge of the wake of the upstream building presented in Fig. 2 show substantial increases in excitation forces which are consistent with the observed increases in responses. Excitation forces associated with vortex shedding from the upstream building are clearly evident in the alongwind force spectrum.

A downstream interfering building generally has very little effect on the response of the principal building upstream except for a small region shown in Fig. 1. Similar results were obtained by Yahyai et al (1992). When the spacing between the buildings is at a critical width which produces a channel wide enough for the wind to converge and stream between the two buildings, the wake characteristics of the principal building are significantly altered.

The spectrum of velocity fluctuations in the modified near wake presented in Fig. 3(a) reveals a dominant reduced frequency of 0.17 which has been shown to induce resonant type response for the principal building upstream at a reduced wind velocity of around 6.

Both the alongwind and crosswind responses were affected and the principal building underwent large elongated elliptical movements, as shown in Fig. 4, along the diagonal line between the two buildings. The effect on the crosswind force spectrum can be seen in Fig. 3(b) which shows a high concentration of excitation energy at the resonant reduced frequency of 0.17.

The interference effects on the principal building are dependent on the plan shape of the upstream interfering building. The effects of a circular cross-section building upstream of the square cross-section principal building were investigated by Bailey and Kwok (1984, 1985).

Due to the difference in Strouhal Number, 0.15 for the circular and 0.1 for the square cross-section buildings, resonant buffeting occurred at a critical reduced wind velocity of 6.8 at which the vortex shedding frequency of the upstream building coincided with the natural frequency of the principal building. The corresponding increase in excitation forces is clearly evident in the alongwind and crosswind force spectra shown in Fig. 5.

As a result, very large buffeting factors were recorded for both the alongwind and crosswind responses of the principal building at this critical reduced wind velocity, but at different critical interference positions, as shown in Fig. 6. Similar observations were made by Taniike (1991) using a small square cross-section upstream building which has a breadth 40% of that of the principal building.

The critical reduced wind velocity at which resonant buffeting occurs depends on the vortex shedding characteristics of the upstream interfering building and may occur at lower and more frequent wind velocities. The sensitivity of interference effects to reduced wind velocity is shown in Figs. 7 and 8 for the alongwind and crosswind responses respectively of the square cross-section principal building for a few critical interference positions.

Large values of the alongwind buffeting factor occur at a critical reduced velocity of around 11 for a square cross-section upstream building identical to the principal building. Similarly, large values of both the alongwind and crosswind buffeting factors occur at a critical reduced velocity of 6.8 for a circular cross-section upstream building.

However, the crosswind buffeting factor for a square cross- section upstream building decreases as reduced velocity increases due to the disruption of the vortex shedding mechanism of the principal building caused by the increase in turbulence created by the upstream building.

The effects of interference are reduced with increase in turbulence in the flow, according to Melbourne and Sharp (1976), Bailey and Kwok (1985), and Kwok (1989). Taniike (1991) suggested that interference effect caused by an upstream building is negligible when the turbulence intensity reaches 17% to 18% at two-third building height.

However, Melbourne and Sharp (1976) and Kwok (1989) reported increases in response of up to 25% and 40% for a square cross-section tall building with a height to breadth ratio of 6 and 9 respectively in an urban terrain of which the turbulence intensities are significantly higher than the value suggested by Taniike (1991).

The effects of a single upstream building have been reported by Melbourne and Sharp (1976) to be evident up to 16 building breadths downstream. However, interference effects have been shown by Melbourne and Sharp (1976) and Saunders and Melbourne (1980) to be significantly reduced when the height of the upstream building was less than two-third the height of the principal building.

For a pair of upstream buildings set 5 breadths apart and at equidistance upstream of the principal building, Saunders and Melbourne (1980) reported a further amplification of the interference effects compared with that of a single upstream building, and the interference effects were noticeable even at an equivalent prototype distance of 1 km or about 30 building breadths.

In addition to those aeroelastic model tests, pressure measurement and force balance techniques have also been employed to study the effects of interference on tall buildings. Blessmann and Riera (1985) studied interference effects using a pressure-tapped model of a square cross-section building with a height to breadth ratio of 6, and Pathak et al (1989) two buildings with ratios of 2 and 8.

Mean force coefficients, including torsion, were determined by integrating the mean pressure distributions. Thoroddsen et al (1985) used a high frequency force balance to determine the mean and dynamic wind loads on five tall building models of different shapes in the presence of an upstream building.

Taniike (1992) employed both force balance techniques and pressure-tapped models in the study of a square cross-section building with a height to breadth ratio of 4.5 under the influence of an identical or different size square cross-section buildings. The results obtained from these studies generally agree with those obtained from aeroelastic tests.

English (1985, 1990, and 1993) also studied the effects of interference using force balance technique and concentrated on the shielding effect instead of the amplification effect of square and rectangular cross-section buildings. The shield factor for the mean alongwind response was found to be sensitive to the building height to breadth ratio, wind flow turbulence characteristics, and in particular separation distance.

By incorporating available data from tests performed in well-simulated atmospheric turbulent boundary layer flows, the following equation was proposed for the mean alongwind response shielding factor:

SF = -0.05 + 0.65x + 0.29x² – 0.24x³ in which x = log(RSD)

RSD = reduced separation distance = d(h+w)/hw

d = separation between buildings

h = height of building

w = width of building

An approach similar to that adopted by English (1993) may be employed to derive empirical equations for the amplification effects of interference of which there is much useful data on buffeting factors.

Torsion Response of Tall Buildings under Interference Excitations:

Results of mean torsion coefficient obtained by Blessmann and Riera (1985) and Pathalc et al (1989), and mean and dynamic torsion coefficients by Thoroddsen et al (1985) for buildings under interference suggest the amplification factor can be as high as 3.

Zhang et al (1994) employed aeroelastic technique to study the torsional response of a square cross-section building with a height to breadth ratio of 5 under interference by four types of building, a square and a circular cross-section building of the same breadth (diameter) and of 60% the breadth (diameter) of the principal building, in an open country terrain.

Critical upstream and downstream interference positions and the corresponding buffeting factors were identified for a range of reduced wind velocities. The buffeting factors for the dynamic torsional response of the principal building at a reduced velocity of 6 are shown in Fig. 9 for an identical interfering building, and in Fig. 10 for the smaller square cross-section interfering building.

Fig. 9 Buffering factors caused by an identical interfering building.

Fig. 10 Buffering factors caused by a smaller interfering building.

The torsional interference excitation mechanisms are similar to those for the alongwind and crosswind responses and are dominated by resonant buffeting which occurs when the frequency of vortex shedding from the upstream building coincides with the natural frequency of the principal building.

The dependence of torsional interference excitation on reduced wind velocity can be clearly seen in Fig. 11 which highlights the critical reduced velocities of 7 for the circular cross-section interfering building with the same diameter as the principal building, and 6 and 4.2 for the smaller square and smaller circular cross-section interfering buildings respectively. Torsional buffeting factors are as high as 2.2 for the circular cross-section interfering building with the same diameter as the principal building.

Zhang et al (1992) also studied the torsional interference effects on a building with a 10% eccentricity between the mass and elastic center of the building. The torsional interference excitation mechanisms and the resultant torsional response characteristics of the eccentric principal building were found to be essentially the same as those for the normal centric building but the buffeting factors due to resonant buffeting were substantially increased, as shown in Fig. 12, to a value as high as 4.6.

Design Considerations:

The significance of interference effects in causing large increases in alongwind, crosswind and torsional responses of tall building has been clearly demonstrated for either an upstream or a downstream interfering building. Upstream interference can be felt at a considerable distance away and has been found to be most prominent when the principal building is located near the edge of the wake shed by the upstream interfering building.

The vortex shedding characteristics of the plan shape of an upstream interfering building can cause resonant buffeting, at which the vortex shedding frequency of the upstream building coincides with the natural frequency of the principal building, at a lower and more frequent wind velocity. Downstream interference is confined to a relatively small area in close proximity which produces a critical separation between the two adjacent buildings.

The magnitude of increase in wind load and response due to interference is strongly dependent on the turbulence created by the approach terrain. Tall buildings situated at the edge of the water front or at the edge of a city center are the most vulnerable. For tall buildings located well within a city center with its highly turbulent wind flow characteristics, the effects of interference are significantly reduced.

However, if there are a few tall buildings or other tall structures which protrude well above the surrounding building canopy, the potential for interference effects would not be diminished and should be assessed in accordance with the wind flow condition above the canopy.

The range of reduced wind velocities at which interference effects are significant encompasses design wind speed which is typical for serviceability consideration, particularly in relation to occupant comfort, to that well in excess of ultimate strength design. In terms of design, a potential increase in wind load of between 10% to 30% may be expected to be within the normal safety factor in design.

However, an increase of between 30% to 50% would be of considerable concern. While an increase of such magnitude may not be catastrophic, the serviceability requirement of the building, in particular the occupant comfort criteria, will almost certainly be tested. An increase of greater than 50%, let alone the more than doubling of wind load which has been observed in some of those wind tunnel experiments, will be well outside the safety margin used in a normal design.

As a result, a comprehensive program of wind tunnel model test is now routinely conducted to aid the design of modern tall buildings and structures to investigate the effects of interference from adjacent buildings, both existing and proposed.