Wind Loads - IS 875 - Where Does Our Code of Practice Stand

Wind Loads: Where does our code of practice stand?

Bhami Mohammed1*, C. Cini1, K. Suresh Kumar2

1Engineer/Scientist, 2Principal/Managing Director

RWDI Consulting Engineers (India) Pvt.Ltd, T-5, Thejaswini, Technopark Campus, Kariyavattom P.O., Trivandrum – 695581

ABSTRACT

This paper elaborates on how well our code of practice for wind loads (IS 875) predicts the wind-induced local loads as well as overall structural loads in conjunction with reality. For this investigation, the wind tunnel results from a standard CAARC building model have been utilized for comparing against the IS 875 predictions. Further predictions from other international codes of practice are also included in this paper for comparison purposes.

Key words: CAARC, Codes, IS 875, Wind loads, Wind tunnel.

INTRODUCTION

The IS code of practice (IS 875: Part 3 (1987)) is widely used for estimating wind loads by the structural engineers all over the country. RWDI India has worked as the wind engineering consultant for a dozen projects in India in the past few years and in several occasions, wind tunnel tested values are compared against analytically computed values using IS code and other international codes of practice. The IS code to our surprise stood apart from other codes and quite differently from the tunnel tested results.

This led us to embark on a test project to compare the difference in IS code to the tunnel tested results for a standard building. The famous CAARC building model (Melbourne (1980)) widely tested in many labs around the world has been chosen for this study. This building is rectangular in geometry and tested without any surroundings. Considering that the codes/standards are based on simple box-like structures without any immediate surroundings, it is anticipated that the wind tunnel results of CAARC buildings should be in reasonable comparison with values from codes of practice.

This paper presents both cladding and structural load comparisons between wind tunnel tests and IS 875. In order to add more credibility to this investigation we included a comparison with the American (ASCE7-05 (2005)) as well as Canadian (NBC 2005 (2005)) codes too. Note that in this study only vertical facades have been considered for the local wind loads considering the significant usage of glazing. *Email: [email protected], Tel: 0471-4060010, Mob: 09895559131

EXPERIMENTAL DETAILS

The data presented here were collected from wind tunnel studies of the standard CAARC tall building model using the HFPI technique. The HFPI method is based on the simultaneous measurement of pressures at several locations on a building. The CAARC is a simple rectangular box type building without any architectural features/geometric disturbances such as parapets, balconies, fins, mullions etc. These tests were conducted on a 1:400 scale model of the building without immediate surroundings, as shown in Figure 1, in RWDI’s 2.4m x 2.0m boundary-layer wind tunnel. A rural upwind terrain condition (zo = 0.1m, corresponding approximately to a power law exponent of 0.17) was simulated for all wind directions by means of floor roughness and upwind spires.

For the HFPI tests, the building was instrumented with 280 pressure taps, including the standard CAARC locations, plus additional taps near the building edges. Time series of the pressures at these locations were collected and stored for post-test analysis. The individual pressure time series were used to form time series of the base loads and generalized forces, from which statistics and spectra could be calculated. The peak base moments, shears and torsion could then be determined. Also, this data is used to derive the local pressure statistics.

Figure 1 provides the pressure tap locations for the HFPI study, along with the overall equivalent full-scale dimensions, axes system and wind flow angle. The wind tunnel tests were conducted for 36 wind directions at 10o intervals. For the analysis, the full scale natural frequency of the building was taken as 0.2 Hz in both sway directions and 0.3 Hz in the torsional direction. The structural damping was taken as 1% of critical and the mass distribution of the building was taken to be 160 kg/m3. These above values were the same as used in the previous CAARC studies (Dragoiescu et al. (2006); Melbourne (1980)).

Figure 1: CAARC building model in wind tunnel - HFPI pressure tap locations

The mean, root-mean-square, peak maximum and peak minimum values of the pressure signals were measured for all wind directions. Note that peak local pressure values corresponding to an averaging time of about 1-sec. All pressure measurements were done with a free stream wind velocity of approximately 15m/s.

LOCAL WIND LOADS

Local mean and rms pressure coefficients corresponding to the ring of taps from row 4 located approximately at 2/3rd height (see Figure 1) are shown in Figure 2 for a wind direction of 0°.Also in this figure, range of results from previous studies at 2/3rd height for a wind angle of 0° is also presented for comparison. Note that in our study, we have eight extra taps at the corners in each row and these taps have been avoided in this comparison in order to match with other studies. The results show a good degree of agreement and the observed deviations from other studies are mostly in the range of ±5%. These observed variations in pressure coefficients are potentially due to distortion of length scale, wind tunnel blockage effect, profile variations etc.

To get an insight into the wind tunnel results, peak factors and gust factors are derived and plotted in Figure 3. Peak factors (gp neg & gp pos) are defined as (peak-mean)/rms, while gust factors (G neg & G pos) are defined as peak/mean. As expected, the negative peak factors are much higher than positive peak factors. Considering the flow separation at edges and high negative pressures, the negative peak factors are getting as high as 12, while the mean of the positive peak factors are about 5 only. On the same token, the negative gust factors are also getting much higher than positive gust factors. Note that the positive gust factors are clustered around 3 which is closer to the 2.5 value used in NBC code.

Being satisfied with our wind tunnel CAARC simulation, the obvious next step is to compare the results with codes. To bring the wind tunnel results in similar format with the IS and ASCE codes, the wind tunnel results have been converted to pressure coefficients based on peak dynamic pressure at roof height. These results have been plotted in Figure 4 for selected elevations at corresponding tap locations for negative and positive pressures. The provided

Figure 2: Comparison of pressure coefficients – RWDI vs. Other studies

-1.0

-0.8

-0.5

-0.3

0.0

0.3

0.5

0.8

1.0

Mea

nC

p

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

RWDI Other studies

-1.0

-0.8

-0.5

-0.3

0.0

0.3

0.5

0.8

1.0

Mea

nC

p

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

RWDI Other studies

0.0

0.1

0.2

0.3

0.4

0.5

Rm

sCp

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

RWDI Other studies

0.0

0.1

0.2

0.3

0.4

0.5

Rm

sCp

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

RWDI Other studies

pressure coefficients in Figure 4 are minimum and maximum pressure coefficients at individual tap locations irrespective of the angle of attack.

Negative pressure coefficients are higher at the edges and moderate at the center as expected from typical flow regime. However, in case of positive pressure coefficients, they seem to increase along the height with the exception of extreme top region where they reduce as the flow tries to escape. Along with the wind tunnel results, the pressure coefficients from ASCE, NBC and IS are also provided at the top. Note that equivalent NBC pressure coefficients in the same format as ASCE and IS have been derived and presented in this plot.

As far as the negative pressure coefficients are concerned, IS code value is -1.2 and this code does not distinguish between edge and center zones. This coefficient is much lower than the edge zone pressure coefficients obtained from wind tunnel results. Also, ASCE and NBC codes provide higher values as well for edge zones. However, IS, ASCE and NBC values for center zone seem closer to the wind tunnel predictions. Note that certain variations in pressure coefficients can be attributed to the conversions from one format to the other. Overall, we can infer that the IS code significantly underestimates loads on claddings located especially on the edges. Also it appears from the IS code that some of local pressure coefficient data on the code could have been obtained under smooth flow conditions and this may be the reason for a lower value.

As far as the positive pressure coefficients are concerned, the wind tunnel results are in good match with ASCE and NBC values, while the IS code is not providing any value.

Figure 3: Peak factor and Gust factor plots

-14

-12

-10

-8

-6

-4

-2

0

2

4

6

8

0 1 2 3 4 5

Mag

nitu

degp neg gp pos G neg G pos

-14

-12

-10

-8

-6

-4

-2

0

2

4

6

8

0 1 2 3 4 5

Mag

nitu

degp neg gp pos G neg G pos

Finally, we worked out an example for predicting local cladding load on a CAARC type building in Mumbai using wind tunnel and various codes. The basic wind velocity for Mumbai of 44 m/s 3-sec gust has been used for this exercise. Figure 5 shows the comparison of peak negative pressures along the height of the building. Note that the wind tunnel predicted cladding loads are between ASCE and NBC predictions, while the IS code predictions are far below. When the average wind tunnel predicted pressure is -3.75 kPa, the maximum IS prediction is only -2.3 kPa. This result shows that the IS code under-predicts cladding loads significantly and the code should be used with caution.

Similarly, the positive pressure distributions have been compared in Figure 6. Note that the wind tunnel predictions are between the ASCE and NBC codes. However, the IS code does not provide any positive pressure value.

Figure 4: Pressure coefficients on CAARC facades

OVERALL WINDS LOADS

The mean base overturning moments normalized with 12

2 2ρUH H Dy are shown in Figure 7,

where UH = reference velocity at roof height, n = natural frequency (0.2 Hz) and Dy = wide dimension of building cross section (45.72 m). The plots also show the corresponding range of results from other studies (Melbourne (1980)). Equivalent mean base overturning moment

Figure 5: Comparison of peak negative pressures

0

20

40

60

80

100120

140

160

180

200

-6.0 -5.0 -4.0 -3.0 -2.0 -1.0 0.0

Peak Negative Pressure (kPa)

Hei

ght(

m)

RWDI

ASCE 7-05

NBC 2005

IS 875

0

20

40

60

80

100

120

140

160

180

200

0 0.5 1 1.5 2 2.5 3

Peak Positive Pressure (kPa)

Hei

ght(

m)

RWDI

ASCE 7-05

NBC 2005

0

20

40

60

80

100

120

140

160

180

200

0 0.5 1 1.5 2 2.5 3

Peak Positive Pressure (kPa)

Hei

ght(

m)

RWDI

ASCE 7-05

NBC 2005

Figure 6: Comparison of peak positive pressuresFigure 4: Pressure Coefficients on CAARC

coefficients for orthogonal directions were also derived and plotted in Figure 7 corresponding to IS and ASCE codes. The results from RWDI and the codes on the standard CAARC model compare reasonably well with results predicted by others in the past. Part of the variation would be the result of differences in the flow simulations. The previous studies were conducted in rougher terrain (suburban or rougher) simulations compared to the rural type terrain condition used for this study.

CMM

UH H Dyx

x= 12

2 2ρCM

M

UH H Dyy

y= 12

2 2ρ

For the analysis, full-scale 3-sec gust wind speed of 44 m/s at 10 m height in open terrain is used. This is the IS code recommended wind speed for Mumbai. The raw overall base loads, moments and torsion obtained from the tunnel, for each wind direction is plotted in Figure 8. From this figure it can be noticed that there is cross-wind loading at angles 90° and 270° in the X direction, while the cross-wind loading is dominating in the Y direction for wind angles 0° and 180°. This cross-wind effects can be seen in moment as well as shear plots. Considering the aspect ratio, as expected the cross-wind effects are significant along the Y-dir when the wind is blowing perpendicular to the wider face or along X-dir. Winds blowing along Y-dir also induce cross-wind loading in the X-dir but only as high as the along-wind loading.

The overall base loads calculated analytically using the IS code is compared against ASCE and the wind tunnel values and they are presented in Table 1. The results show that the values predicted in the X direction using IS is generally in good agreement with the tunnel values, while the ASCE values are lower. In the Y-direction, both IS and ASCE are predicting lower values than the tunnel values. This is because the tunnel predicted values are based on cross-wind loading when both codes are predicting along-wind loads. This shows that the codes will under predict the loads in cases where cross-wind phenomenon is dominating.

-0.7

-0.5

-0.3

-0.1

0.1

0.3

0.5

0.7

CM

y

0 60 120 180 240 300 360 Wind Direction (degrees)

RWDI Other studies ASCE IS

-0.7

-0.5

-0.3

-0.1

0.1

0.3

0.5

0.7

CM

y



-0.7

-0.5

-0.3

-0.1

0.1

0.3

0.5

0.7 C

Mx



Figure 7: Wind-induced mean response of CAARC building along with predictions from other studies and codes

-3.0E+09

-2.0E+09

-1.0E+09

0.0E+00

1.0E+09

2.0E+09

3.0E+09

Bas

eO

vertu

rnin

gM

omen

t(N

-m)

10 60 110 160 210 260 310 360 Wind Direction (degrees)

My

-3.0E+09

-2.0E+09

-1.0E+09

0.0E+00

1.0E+09

2.0E+09

3.0E+09

Bas

eO

vertu

rnin

gM

omen

t(N

-m)


My

-2.0E+09

-1.0E+09

0.0E+00

1.0E+09

2.0E+09

Base

Ove

rturn

ing

Mom

ent(

N-m

)


Mx

-2.0E+09

-1.0E+09

0.0E+00

1.0E+09

2.0E+09

Base

Ove

rturn

ing

Mom

ent(

N-m

)


Mx

-3.0E+07

-2.0E+07

-1.0E+07

0.0E+00

1.0E+07

2.0E+07

3.0E+07

Base

Shea

rFor

ce(N

)


Fx

-3.0E+07

-2.0E+07

-1.0E+07

0.0E+00

1.0E+07

2.0E+07

3.0E+07

Base

Shea

rFor

ce(N

)


Fx

-2.0E+07

-1.5E+07

-1.0E+07

-5.0E+06

0.0E+00

5.0E+06

1.0E+07

1.5E+07

2.0E+07 Ba

seSh

earF

orce

(N)


Fy

-2.0E+08

-1.5E+08

-1.0E+08

-5.0E+07

0.0E+00

5.0E+07

1.0E+08

1.5E+08

2.0E+08

Bas

eTo

rsio

nalM

omen

t(N

-m)


Mz

Maximum

Mean

Minimum

Legend

Figure 8: Raw overall base moments, shears and torsion on CAARC building

Table 1: Overall base loads on CAARC building

Fx (kN) Fy (kN) My (kN-m) Mx (kN-m) Mz (kN-m) IS 2.39E+04 1.54E+04 2.42E+06 1.56E+06 -

ASCE 2.27E+04 1.41E+04 2.18E+06 1.36E+06 1.10E+05 RWDI 2.39E+04 1.64E+04 2.66E+06 1.93E+06 1.55E+05

Estimated shear force distribution is provided in Figure 9 along with predictions from codes. It is evident that the wind tunnel distribution has a much steeper slope than the code distributions. This brings the shear forces on upper floors higher than code predictions while the reverse is happening towards the bottom. Overall, this difference in distribution does not affect the overall base shear force which is evident from Table 1, but this difference in distribution does affect the overall base moments. Considering the steep slope of the distribution in the tunnel generated values, it is expected that the corresponding moments to be higher considering the increasing influence of higher loads towards the top.

In the wind tunnel analysis, firstly the inertial loading is distributed based on mode shape and mass at each floor level. Later, the quasi-static loading is distributed based on an iterative approach of finding a suitable pressure distribution which can satisfy the overall base shear and moment. Initially, these pressure distributions are assumed based on the measured quasi-static base shear and moment. However, the code philosophy itself is very different which is mostly based on a gust effect factor which is constant. Also, the code distribution mostly represents along-wind loading with mild inertial component. When the inertial component becomes significant even in along-wind loading, the distribution is expected to be steeper. Note that the Y-dir distribution is steeper than X-dir distribution and this is due to the cross-wind influence in the Y-dir.

To get an insight into the load distribution resulted from tunnel tests, we have divided the total distribution into inertial and quasi-static components and they are plotted in Figure 10. Note that

020406080

100120140160180200

0 100 200 300 400 500 600Shear Fy (kN)

Hei

ght(

m)

RWDI ASCE IS875

020406080

100120140160180200

0 100 200 300 400 500 600Shear Fy (kN)

Hei

ght(

m)

RWDI ASCE IS875

020406080

100120140160180200

0 200 400 600 800

Shear Fx (kN)

Hei

ght(

m)

RWDI ASCE IS875

020406080

100120140160180200

0 200 400 600 800

Shear Fx (kN)

Hei

ght(

m)

RWDI ASCE IS875

Figure 9: Shear force distribution on CAARC building

in X-dir, the quasi-static component is noticeable compared to Y-dir. The inertial distributions in both X and Y directions are as expected. But in Y-dir, the inertial distribution is dominant compared to the quasi-static loading and this is once again due to the influence of cross-wind effect. These results once again reinforce the comparatively steeper distribution in tunnel compared to codes.

Torsional distributions derived from wind tunnel results and ASCE code for the CAARC building is presented in Figure 11. As evident from figure, the ASCE torsion values are much lower compared to the tunnel results. Also the total base torsion provided in Table 1 is significantly higher than ASCE prediction for the same. Note also that the obvious aerodynamic

Figure 11: Torsional distribution on CAARC

020406080

100120140160180200

0.E+00 1.E+03 2.E+03 3.E+03 4.E+03 5.E+03

Torsion Mz (kN-m)

Hei

ght(

m)

RWDI ASCE

020406080

100120140160180200

0.E+00 1.E+03 2.E+03 3.E+03 4.E+03 5.E+03

Torsion Mz (kN-m)

Hei

ght(

m)

RWDI ASCE

Figure 10: Quasi-static and inertial load distribution on CAARC building

020406080

100120140160180200

0 100 200 300 400 500

Force Fx (kN)

Hei

ght(

m)

Quasi-static-Fx Inertial -Fx

020406080

100120140160180200

0 100 200 300 400 500

Force Fx (kN)

Hei

ght(

m)

Quasi-static-Fx Inertial -Fx

020406080

100120140160180200

0 100 200 300 400 500

Force Fy (kN)

Hei

ght(

m)

Quasi-static-Fy Inertial -Fy

020406080

100120140160180200

0 100 200 300 400 500

Force Fy (kN)

Hei

ght(

m)

Quasi-static-Fy Inertial -Fy

coupling between torsion and X-dir response from Figure 8. This is the sole reason for wind tunnel results to be much higher than the ASCE predictions. Typically, the codes do underestimate torsion on a building and many researchers are on the work to bring reasonable values into the codes. It is recommended that IS code also look into this matter and provide a provision for calculating torsional loads.

CONCLUDING REMARKS

In this study, cladding and structural wind load predictions for CAARC building using IS 875 have been compared against the wind tunnel test results as well as predictions using other international codes of practice. Our results indicate that concerning the cladding loads predicted by IS 875 code, the pressure values on edges stand much lower than the tunnel tested and ASCE figures. In summary, it appears that IS code predictions for cladding design should be used with caution and it is strongly recommended to look at other international codes as well at the preliminary design stage. Also, this study shows that it is time to revamp the tabular values for external pressure coefficients provided for local cladding design.

As an initial step, we recommend the pressure coefficient to be increased to -1.8 from the current value of -1.2 at least for an edge zone of 20% building width. For the central region, use the same pressure coefficient of -1.2 currently in the code. We recommend to use the load calculation procedure in IS 875 with minimum risk (r=0.63, life of 50 years) by using the above recommended pressure coefficients.

Also in this study, code predicted structural loads were compared and the comparisons indicate that IS 875 is indeed good enough for preliminary structural load predictions in the absence of across-wind response domination.

ACKNOWLEDGMENT

The authors are thankful to Prof. G.N.V. Rao (Retired Professor, Department of Aerospace Engineering, Indian Institute of Science, Bangalore, India) for his valuable comments and suggestions towards the improvement of this paper. Prof. Rao is a well-known figure in the field of Wind Engineering in India and the third author of this paper was blessed to start his career in Wind Engineering under his guidance.

REFERENCES

ASCE7-05 (2005). “Minimum Design Loads for Buildings and Other Structures”, Published by the Structural Engineering Institute of ASCE, Reston, VA.

Dragoiescu C., Garber J., Suresh Kumar K. (2006). “A Comparison of Force Balance and Pressure Integration Techniques for Predicting Wind-Induced Responses of Tall Buildings”, ASCE Structures Congress 2006 – Extreme Event Loading - Extreme Wind Engineering.

IS:875 Part 3 (1987), “Code of Practice for Design Loads (Other than Earthquake) For Buildings and Structures, Part 3 wind Loads”, Indian Standard.

Melbourne W.H. (1980). “Comparison of measurements on the CAARC standard tall building model in simulated model wind flows”, Journal of Wind Engineering and Industrial Aerodynamics, 6, 73-88.

NBC2005 (2005). “User’s Guide – NBC 2005, Structural Commentaries (Part 4)”, Issued by the Canadian Commission on Building and Fire Codes, National Research Council of Canada.

Wind Loads - IS 875 - Where Does Our Code of Practice Stand

Documents

Transcript of Wind Loads - IS 875 - Where Does Our Code of Practice Stand