Wind Loads on a Medium-Rise Building by Using Five Major International Wind Codes and Standards

Wind load calculations are done for a medium-rise building with 48 m height locates in an urban area of Sri Lanka. Five different international wind loading standards were selected to calculate wind loads according to prevailing conditions in Sri Lanka. Available 3 second gust wind speed is converted in to mean hourly wind speed and 10 minutes average wind speed to use selected wind loading standards. The difference of wind load calculations is evaluated by using member forces in columns, beams, and shear walls obtained from a 3–D finite element modelling. The comparison displays that both the Australian standards yields higher wind loads compared to other selected standards. The use of higher terrain-height multiplier for cyclone prone zone 1 contributes much larger member forces compared to other cases. None of the standards exceeds the serviceability criteria, which evaluate by means of drift index.


Introduction
Medium rise buildings are less sensitive to wind loads compared to high-rise buildings that are subjected to wind loads with same magnitude. Thus, when designing medium-rise buildings, design engineers are conservatively more concerned about vertical loads than lateral loads. As a result, medium-rise buildings employ less sophisticated designs and engineering solutions such as vibration control systems. This makes medium-rise buildings more popular building type in many countries including Sri Lanka. However, extreme wind conditions such as cyclones, tornados impose large wind loads on structures which require designers to design even medium-rise buildings more strongly and safely against wind loads. In such circumstances wind loading standards serve as a main design tool that can be used to design simple building geometries adequately in the absence of sophisticated testing methods such as wind tunnel tests, numerical simulations, etc. Medium-rise buildings are more common in urban areas of Sri Lanka. Most of these buildings have 2 to 5 storeys but sometimes can be as many storey as 15.Because Sri Lanka is yet to prepare its own wind loading standard, most of these buildings were designed according to various international wind loading standards. These standards range from previous British code of practice CP3 Chapter V-Part2:1972 [1] to newest wind loading standards such as EN 1991-1-4:2005 [2]. However, this kind of direct adoption of international wind loading standards is questionable because less harmonisation of wind load calculations exist among different wind loading standards. In addition to that, there is no clear agreement of the use of country specific factors which were originally derived for prevailing condition of native country of the standard. Though if one assumed effects of aforementioned factors are less, then the main uncertainty is associated with conversion of available basic wind speeds to different time averaged wind speeds to calculate wind loads as provided in different standards. Because the accuracy of wind load calculation mainly depends on the basic wind, the conversion process should be precise. For an example, the available 3 second gust wind speeds in Sri Lanka needed to be converted in to mean hourly or 10 minutes average wind speeds to use BS 6399-Part 2:1992 [3] and BS EN 1991-1-4:2005 [2] respectively. In this study, wind load calculation is done for a 48 m height building by using five international wind loading standards. For the calculation purpose existing 3 second gust wind speeds are converted in to 10 minutes and 1 hour wind speeds by using best available methods. The similarities and differences of derived wind loads from five different wind loading standards are evaluated by comparing several types of structural loads including axial force, shear force, bending moment obtained from Finite Element Model (FEM) of the building.
Following the introduction of this study, section 2 briefly describes the historical and current practices of wind engineering in Sri Lanka. The selected international wind loading standards will be introduced in section 3. Section 4 illustrates the details of 3-D finite element model. The conversion procedure of 3second gust wind speeds to wind speeds with different averaging time explains in section 5. Section 6 displays results of member forces obtained from finite element modelling. Finally conclusion illustrates the possible reasons for differences and similarities on wind load calculation done with five different wind loading standards.

Previous practice in Sri Lanka
Much less attention was paid in designing ordinary buildings against high wind conditions until a massive cyclone hit Eastern and North-Eastern coastal areas of Sri Lanka in 1978. The damage that occurred in 1978 cyclone was huge not only due to underestimation of wind load at design stage but also due to poor workmanship and low quality of building materials [4].The huge losses in both human lives and properties led the production of the first Sri Lankan mandatory document about wind loading" Design of buildings for high winds -Sri Lanka" [5] (Here after referred as design manual). The design manual was based on the previous British code of practice CP 3 Chapter V -Part 2: 1972 [1]. This document consists two main sections; the first section is wind speeds and wind zoning map of Sri Lanka. The second section is about proper design and construction methods to enhance wind resilient ability of a low rise building. These simple techniques are addressing strengthening roof structure, proper fixing of roof covering materials, increase the structural integrity of a low rise building, etc. One of the important aspects of the design manual is defining basic wind speeds and wind zones in Sri Lanka. However, these basic wind speeds were not derived either by using a properly measured data set or using sound statistical derivation methods. According to the Clarke et al [4] these wind speeds are derived based on information such as damage survey done along the cyclone path, available historical data, considering distance from the coast line and particular location, and using Australian practices. Therefore, some researchers ( [6], [7]) raised doubts about magnitude of these wind speeds not only due to the lack of theoretical basis on derivation but also low wind speeds compared to other countries with similar geographical and meteorological conditions. Not only the basic wind speeds but also wind zoning is needed to be revised because some parts of the island may not be free from cyclones anymore as assumed by the design manual. For an example, according to the historical cyclone data, there were two cyclones that entered the country through West coast in 1925 and 1967 [8], which is basically located in wind zone 3 with low probability of hitting by a cyclone.
Addition to the design manual, the most common practice was use of previous British code of practice; CP 3 Chapter V-Part 2: 1972 [1]. The quasi-static approach used in CP 3 Chapter V-Part 2: 1972 [1], is quite accurate in estimating wind loads on medium-rise buildings but less accurate in predicting behaviour of tall slender structures with dynamic behaviour like modern high-rise buildings [9]. On the other hand, many factors given in previous code of practice such as pressure coefficients, statistical factor, concept of return period has changed with the advance of wind engineering researches done over last few decades. The new standards, which supressed CP 3 Chapter V -Part 2 such as BS 6399-Part 2:1997 [3], and EN 1991-1-4:2005 [2] incorporate those new data sets and advance methods for wind load calculation which may lead to have better building design. However, even after several decades general practise in Sri Lanka has not changed much yet. Thus evaluating wind loading standards in Sri Lankan conditions is a timely requirement before employing them. Comparison of different wind loading standards is a common objective completed before adopting any international wind loading standard by a country. A series of technical papers published on Asia-Pacific Conference of Wind Engineering (APCWE) ( [10], [11], [12]) focused to harmonize different wind loading standards used in Asia-Pacific region. However, this objective is yet to be accomplished as wind loading standards display significant discrepancies when use them in conditions outside the country of origin. Nevertheless, these Comparisons are done in systematic manner for low-rise, medium-rise, and high-rise buildings separately due to their inherent structural responses ( [11], [12]). Tamura el al [10] compared fifteen different wind loading standards in Asia-Pacific region by using 48 m height building and found that no significant correlation in basic reactions or dynamic response. However, better correlations were observed for peak cladding pressure and force coefficient. The comparison done by Kola [13] by using 4 international standards including American, Canadian, Australian and British displayed that design pressures on component and cladding can be differ as large as 200%.

Selected codes and standards for the study
Many Sri Lankan engineers are using different international wind loading standards to design multi-storey buildings. These standards are ranging from old code of practices such as CP 3 Chapter V -Part 2 to newest code including Euro standard. By considering most popular practices found in Sri Lanka, following codes and standards were selected for the comparison purpose of this study; CP 3 Chapter V -  [15]but it also allows adjusting methods and factors according to the prevailing conditions in adopted country by means of a national annex.

Case study
The majority of medium-rise buildings in Sri Lanka are 2 to 15 storey buildings. Therefore, as a case study, a rectangular shaped office building with 48 m height and plan dimension of 60 m x 30 m was analysed by using SAP 2000 software [16] (Figure 1(a)).The same building was used by previous researchers [10]for comparison of wind load calculation by using different international wind loading standards. This building was assumed to be located in anurban area of Sri Lanka. It is also assumed that building has 12 storey with 4 m floor-tofloor height. All structural components are designed in detail to simulate realistic structural behaviour of the building. Building is designed as a concrete wall-frame structure. Columns and beams act as a frame structure while shear walls are served as walls of the service core.600mm x 600mm columns are in a 6m x 6m grid from ground floor to 12th floor. All columns are connected by using 600 mm x 400 mm (depth x width) beams at the slab level. Slab thickness is selected as 175 mm to satisfy deflection criterion. Shear walls are designed as 300 mm thick reinforced concrete walls around service core of the building. The soft zoning lift system, ducts, and toilets are located within the service core. All columns and beams are modelled in SAP 2000 [16] FEM package as frame members with 6 degree of freedom while slab and shear walls are simulated as fournodes shell elements with membrane and platebending behaviour. The diaphragm constraint is used for slabs to preserve membrane action of the slab. Other than the dead load of structural members, super imposed and live loads are applied in model according to the BS6399: Part 1: 1996 [3]. Wind loads on building are applied in two orthogonal directions as joint loads at the column -beam junctions on the wind ward and leeward faces separately as shown in Figure 1 (b) .

Basic wind speeds with different averaging time
Wind loading standards use different time averaged wind speeds to calculate wind loads on structures. However in Sri Lankan context,  available wind speed data are only 3 second gust wind speeds in three wind zones. Therefore, it is necessary to convert 3 second gust wind speeds to one hour or 10 minutes average wind speeds by using some conversion factors. Firstly, 3 second gust wind speeds are converted in to mean hourly wind speed as method proposed by Cook [17]. According to Cook's method, the mean hourly wind speed (Vmean) at can be calculated from gust wind speed (VG) as shown Equation (1).
Where,S b is the terrain and building factor at 10 m height in an open terrain as given in Table 4 of BS 6399.2:1997 [3].
The value of S b is only changing with distance from sea toa site of interesting to calculate wind speed.
In this study, it is assumed that building sites in all three zones are within 2 km from the sea, thus having 1.78 for value of S b .
Calculated mean hourly wind speed is converted in to 10 minutes average wind speed by using factor 1.06 as proposed by the Institute of Civil Engineers in United Kingdom (ICEUK) [18]. The wind speeds in all three zones with different average times are shown in Table 1.

Wind induced forces
Wind loading standards only facilitate to calculate wind pressures at different heights of the building. By multiplying these values with corresponding contributory areas will give the wind load at a particular height. However, it is not the actual force experienced by structural members due to their load sharing mechanisms. These actual member forces are important in designing structural members. Structural member forces in columns, beams, and shear walls can be obtained by using 3-D finite element model by applying external loads, in this study it is wind loads. The member forces used for comparison in this study are maximum values of axial force, shear force and bending moment in columns, shear forces and bending moments in beams, base moment and base shear at the support level and maximum compressive stress in shear walls.
Wind induced forces such in columns, beams, supports and shells on 48 m high building in all three zones are shown in Figure 2   According to the FEM analysis, the governing load case for designing structrural members is load combination 1.2G+1.2Q+1.2W, which gives the maximum structural load. Except results obtained for both Australian standards in zone 01, other standards generate approximately equal member forces. The maximum difference in column shear force and bending moment is about 10% in zone 01 when wind flows perpendicular to 60 m long wall for both Australian standards. That differnce is about 6-7% when wind flows perpendicular to 30 m wall. According to Figure 2(b) and 6(b) difference of forces in beams for Australian standards can be high as 15-18% and 5-6%in zone 01 when wind flows perpendicular to 60 m side and 30m side respectively. This might be due to use of higher terrain-height multiplier in zone 1 to accountit as a cyclone prone area, which is subsequently resulteded higher wind loads. However, the difference of calculated wind loads from different standards is clearly visible from results of load combination 1.0G+1.4W as the effect of dead load is minimal. The use of higher terrain-height multiplier cause to have 50% higher shear force and 30% higher bending moment in columns in zone 1 when wind flow perpendicular to the 60m side. For rest of cases, both Australians standards lead to have slightly higher member forces compared to other wind loading standards. This means that use of higher terrain-height multiplier is not indeed necessary unless a designer willing to design highly cyclone resilient building such as cyclone shelters. However when the contribution of dead load has maximum such as in load combination 1.4G+1.4W column loads are reduced but beams loads are increasing with respective to the previous load combination 1.0G + 1.4W. According to the Figure 9, in most of cases resulted column loads are equals for CP 3 Chapter V -  [3] other three wind loading standards resulted slightly larger member forces for each load combination. This means that calculated wind loads by using Australian and Euro code are higher than the wind load calculated by using British standards.

6.2.
Base reactions The summation of wind loads calculated from different wind loads and their overall effect reflect from base reaction as summarized in Table 2.The base shear and base moments valuesvary according to the derived pressure values and their vertical distribution along the height of the building.Therefore, in zone 1, maximum base moment and base shear can be observed for Australian standards, because of their higher wind speeds resulting from higher terrain-height multiplier used in zone1 reulted larger wind loads. These values are almost twice as the values derived from CP 3 Chapter V-Part2:1972 (Figure 10(a)). In zone 2 and zone 3, Euro codes yielded higher base moment as well as base shears values compared to CP 3 Chapter V-Part2:1972 (   Table 2, it can be concluded that this might be resulted from combine effect of higher calculated wind loads and the use of divisionby-parts method for pressure distribution in EN 1991-1-4:2005.

.3
Maximum shell stress in shear walls Similar to member forces, maximum compressive stress in shear walls is observed for both Australian standards in zone 1. From  Figures 11(a) and (b), the maximum normalized compressive stress can be as high as 2.1 and 2 for wind flow normal to 60 m side and 30 m side respectively. Another striking feature of Figures 11(a) and (b) is all standards contribute larger compressive stress in shear walls compared to CP 3 Chapter V-Part2:1972 . That difference is more prominent when wind flow normal to 30 m side than wind flow perpendicular to 60 m side.   2. Both Australian standards give higher forces in zone 1 due to use of higher terrain multiplier for cyclonic areas. Unless designing an important post disaster structure, this factor can be omitted to have a more economical design.
3. Same trend can be seen in maximum stresses in concrete shell of shear wall in both models. Both Australian standards gave higher stresses because of their higher wind pressure values in zone 1. However, these differences were reduced in zone 2 and 3. In these zones 2 and 3, Euro codes yielded higher stresses due to higher wind loads derived from the code. 4. Base shear and base moments are two major measurements on determining the effects of wind loads, which are directly combine with the wind load values and their distribution along the height of the building. Higher base moment and base shear obtained from British and Euro codes because of they use division by parts rule for distributes wind loads along the height of the building.
5. Division -by -parts rule trends alter the vertical necessary to balance the derived wind loads and its linrear vertical variation carefully.
6. For the medium -rise builing used in this study does not exceed the drift limit irespective to building location in any wind zone.