Internal pressures in a building with roof opening

The Eighth Asia-Pacific Conference on Wind Engineering, December 10 14, 2013, Chennai, India Internal pressures in a building with roof opening Shouying Li 1,Shouke Li 2, Zhengqing Chen 3. 1 Asociate Professor of Civil Engineering, Hunan University, Changsha,Hunan, China, shyli@hnu.edu.cn 2 Phd. Candidate of Civil Engineering, Hunan University, Changsha,Hunan, China,32090170@163.com 3 Professor of Civil Engineering, Hunan University, Changsha,Hunan, China,zqchen@hnu.edu.cn ABSTRACT Six TTU rigid models with dominant opening on roof are manufactured with a scale of 1:50 for carrying out wind tunnel tests. The fluctuating mechanisms and variations of internal pressures are specially studied, and the results from wind tunnel tests are compared with that of previous studies about wall opening buildings and that of several provisions. It is shown that there is a high frequency Helmholtz resonance for internal pressure, and the test results are out of the range of previous studies about wall opening buildings, as well as the provisions of both the Chinese wind loading code-the GB50009:2012 and the American wind loading code-the ASCE7-10,which are sometimes non-conservative in the evaluation of mean and peak internal pressure coefficients. Keywords: Wind tunnel test, Internal pressure, Roof opening, Helmholtz oscillation Introduction Building opening is usually divided into functional opening and destructive opening. Internal pressures form a large portion of cladding design pressures on a building with functional opening, such as door, window, sky-window or retractable roof. Moreover, severe external pressures could be transmitted into the internal space if windows are left open during a storm or envelope is subject to a failure present a destructive opening. Wind induced internal pressure inside a building with a single windward dominant opening is usually considered as the most critical design configuration by wind engineers. Evaluation of wind-induced internal pressures received a lot of attention starting in the 1970s. Holmes(1979) carried out BLWT experiments to investigate the effect of various opening configurations on internal pressures for a two story house. It is shown that the mean and root-mean-square (rms) of the internal pressure coefficients were both monotonically increasing functions of the ratio of the windward to leeward opening areas for this type of buildings, and inertia effects produced resonance amplification for internal pressures. A number of investigators (for example, Liu et al.,1986; Ginger et al.,1997,2010; Oh et al.,2007;woods et al.,1995;kopp et al.,2008;lu et al.,2006;) have increased our understanding of the characteristics of building internal pressure in the presence of dominant openings. Previous wind-induced internal pressures researches were mostly based on the wall opening low-rise buildings, and opening ratio is usually in the range of 3~10%. In fact, some Proc. of the 8th Asia-Pacific Conference on Wind Engineering Nagesh R. Iyer, Prem Krishna, S. Selvi Rajan and P. Harikrishna (eds) Copyright c 2013 APCWE-VIII. All rights reserved. Published by Research Publishing, Singapore. ISBN: 978-981-07-8011-1 doi:10.3850/978-981-07-8012-8 270 940

buildings often appear functional opening on the roof and their opening ratio is beyond 15%. For example, the DongSheng Stadium s retractable roof s opening ratio is 18.2% in China, and that of Amsterdam Stadium is 20.5%, as well as that of Millennium stadium is 25%.Moreover, damage opening usually first presents on the roof corner, and the opening ratio is not very large in early stage, but the opening could contribute a lot to internal pressures. This Paper is mainly focused on the wind induced internal pressures for these buildings with a single dominant opening on roof. The coherences, distributions and the effect factors on internal pressures are specially studied. The values of internal pressure coefficients from wind tunnel tests are compared with several provisions such as Chinese Code, AIJ, ASCE, NBCC. Wind Tunnel Tests Wind tunnel tests were carried out in the high wind speed section of boundary layer wind tunnel at Hunan University. Spires and roughness elements were used to simulate velocity and turbulence intensity profiles at 1:50 geometric scale for suburban terrain. Six roof opening buildings (9.14 13.72 3.96m) are manufactured with a scale of 1:50, which are 3 roof center opening models with opening ratio of 15%, 20%, 25%, and 3 roof corner opening models with opening ratio of 3%, 5%, 8% as shown in table 1. The photo of test model in BLWT is shown in Figure 1. The same number of pressure taps were installed in the roof s upside and downside surface (Fig. 1(b)) for measuring the external pressure and internal pressure. It is difficult to measure the external pressure at the opening without changing the flow pattern. Therefore, the external taps were placed around each opening and the averaged time history of these taps was used as the approximation of the external pressure at the opening. The internal pressure was approximated from the tap arranged away from the opening. Table 1 Summary of the tested case Case No. Opening ratio Case No. Opening ratio 1 Center Opening 15% 4 Corner Opening 3% 2 Center Opening 20% 5 Corner Opening 5% 3 Center Opening 25% 6 Corner Opening 8% 0 270 182 3% 5% 8% 15% 20% 25% 123.6 61.8 90 79.74 22306 a)photo of test model b)tap location of test model for case 1~6 Fig. 1 Plan of test model 275.84 180 941

For maintaining the similarity of fluctuating internal pressures between the wind tunnel test and full scale test, a chamber was added to wind tunnel test model under floor as shown in Figure 2 by means of Holmes(1979).The total volume including the test model and the added chamber is 9 times as large as the test model, so the wind velocity scale is 1/3. The Scan pressure measurement system was used to acquire the data at a sampling frequency of 330 Hz. The sampling number was 6600. The reference dynamic pressure was obtained with the pitot tube 0.5m upstream of the model and 8m above the floor. The test wind velocity is 11.0m/s. All the data was amended by the tube transfer function in frequency domain. Test Model a) Cavity scaling method b) Added volume chamber Fig.2 Scaling method of internal volume of test building Fluctuating mechanism of internal pressures Fig.3 depicts coherence of internal pressures at different tap points which arranged away from the opening in the model for roof center opening case (Case 1) and roof corner opening case (Case 4). For Wind angle of 0 at Case No.1, the coherence amplitude of internal pressure coefficient is above 0.9 in the frequency range of 0~3Hz, indicating that the internal pressures are highly correlated; at Wind angle of 45, the coherence amplitude of internal pressure coefficient is greater than 0.9 within the band of 0~0.86Hz, and fluctuating significantly within the band of 0.86~3Hz, but the amplitude is greater than 0.8; at wind angle of 90, coherence amplitude of internal pressure coefficient is greater than 0.9 within the band of 0~1.9Hz, the inner pressure coefficient highly correlated. Among the three typical wind angle, wind angle of 0 had the best internal pressure coherence performance, and at 45 case, the conical vortex resulted in its internal pressure coherence dropped. Comparing the result of 0 and 90 wind angle, if the opening is wider, the coherence is worse. However, the internal pressure coherences amplitudes for these three wind angles are close to 1.0 in the main frequency ranges which show a good correlation for Case 1, so which indicated that the internal pressures for center opening case can be depicted by one uniform pressure time history curve. Figure 3b shows the coherence of internal pressures for roof corner opening case (Case 4) at wind angle of 0, 315 and 270. As can be seen from Figure 3b, the coherence function amplitude is greater than 0.98 in the frequency range of 0 ~ 3Hz at 0 wind angle, there is a good correlation that the internal pressure process can be accurately described by a uniform time history; there are the similar results for 270 and 315 wind angle. Comparing the result of case 1 with case 4, it is shown that there is a better coherence performance for 942

roof corner opening case 4. Coherence of Internal Pressures 1.0 0.8 0.6 0.4 0.2 0 Wind angle 45 Wi nd angl e 90 Wi nd angl e Coherence of Internal Pressures 1.0 0.8 0.6 0.4 0.2 0 Wind angle 315 Wi nd angl e 270 Wi nd angl e 0.1 1 10 f (Hz) 0.1 1 10 f (Hz) a) Case 1 b) Case 4 Fig.3 Coherence of internal pressures at different points in the model Fig.4 depicts auto-spectra of internal, external and typical external pressure for case 1 and case 4 for a wind angle of 0. Fig. 4 indicates that the low frequencies of the spectra contain the greatest proportion of the energy. The experimentally obtained spectra also indicate broadband peaks at 6.5 and 6.3 Hz for the three cases 1, and 4, respectively. Clearly, the resonance is the worst for the roof corner opening Case 4, consistent with the decrease in open ratio in the equation governing the internal pressures. It is because there is a significant vortex-driven Helmholtz resonance for case 4. Scp 1 0.1 1 1E-3 1E-4 1E-5 Helmohotz resonance f H =6.5Hz Cpi Cpe Cpe of typical tap 0.1 1 10 Scp 0.1 1 10 f (Hz) f (Hz) a) Case 1 b) Case 4 Fig.4 Power Spectra of internal, external and typical external pressure for case 1 and case 4 0.1 1 1E-3 1E-4 1E-5 Helmohotz resonance f H =6.3Hz Cpi Cpe Cpe of typical tap Variation of internal pressures with wind angle Fig.5 depicts mean, fluctuating, maximum, and minimum internal pressure coefficients for case 1~6. Fig. 4a indicates that the mean internal pressure coefficients are fluctuating significantly with wind angle, and all the mean values are negative in 0~360 wind angles for all cases, and the value of roof center opening case (Case 1,2,3) is larger than that of roof corner opening case (Case 4,5,6).There is no significant monotonous effect on internal pressure for opening ratio. Fig. 4b indicates that fluctuating internal pressure coefficients vary with wind angle significantly, and the smallest value is close to 0 for roof center opening case, the biggest value is close to 0.20 for roof corner opening case. The maximum value (0.36 for 275 ) of maximum pressure coefficient for roof corner opening case 6 (0.45 for 85 ) is greater than that of roof center case 1, and the minimum value (-1.80 for 290 ) of minimum pressure coefficient for roof corner opening case 5 is greater than that of roof center opening case 1 943

(-1.23 for 190 ). Overall, the suction and pressure is bigger for roof corner opening case than that of roof corner opening case. Mean internal pressure coefficients -0.2-0.4-0.6 Case 1 Case 2-0.8 Case 3 Case 4 Case 5 Case 6-1.0 0 45 90 135 180 225 270 315 360 Wind angle a)mean internal pressure coefficients Fluctuating internal pressure coefficients 0.20 0.15 0.10 5 Case 1 Case 2 Case 3 Case 4 Case 5 Case 6 0 0 45 90 135 180 225 270 315 360 Wind angle b)fluctuating internal pressure coefficients Maxima internal pressure coefficients 0.6 0.3-0.3-0.6 Cas e 1 Cas e 2 Cas e 3 Cas e 4 Cas e 5 Cas e 6 0 45 90 135 180 225 270 315 360 Wind angle Minima internal pressure coefficients -0.3-0.6-0.9-1.2-1.5-1.8 Cas e 1 Cas e 2 Cas e 3 Cas e 4 Cas e 5 Cas e 6 0 45 90 135 180 225 270 315 360 Wind angle c) Maxima internal pressure coefficients d) Minima internal pressure coefficients Fig.5 Variation of internal pressure coefficients with wind angle Comparison with previous tests The building with a dominant opening on the wall was the major subject in previous experimental research on internal pressure. The WOODS model (Woods et al. 1995) and the SHARMA model (Sharma et al. 2003) are the same models to currents test, however, which did not have the added volume scale for fluctuating internal pressure. The opening ratios are 1,4,9,16,25% for WOODS model on the wall, and 2.7,5.3% for SHARMA model on the wall, as well as 3,5,8,15,20,25% for current test on the roof. The other previous test models are close to current test model, and the opening ratios are close to current roof corner opening. The summary of previous experimental research models on internal pressure is shown in Tab.2. Fig.6a depicts the comparison of mean internal pressure between previous wall opening tests and current roof opening tests. For the small opening ratios (3,5,8%), when the opening is located in windward wall, the mean internal pressures of previous tests are positive, which have a larger value than the current model. However, when the opening is located in leeward wall, the mean internal pressures of previous tests are negative, which have a smaller value than the current model. For the large opening ratios (15,20,25%), the mean internal pressures of previous tests have a larger value than the current model. Fig.6b depicts the comparison of fluctuating internal pressure between previous wall 944

opening tests and current roof opening tests. The fluctuating internal pressures are smaller than 0.2 for all roof opening test cases. Overall, the fluctuating internal pressures of previous tests have a larger value than the current model. Table 2 Summary of previous experimental research on internal pressure Ref. Dimension of Test Scale Opening Ratio(%) Wind angle Field Volume Scale Building(mm) Holmes(1978) Domestic 1/50 0.35~21.6 Normal Open No Liu (1986) 140*140*290 N/A 2~1.4 Normal N/A No Woods(1995) 182*274*80 1/50 1,4,9,16,25 0~180 Open No Ginger(1997) 9.1*13.7*4 1/1 1.1,2.2,5.5 Normal Open No Sharma(2003) 182*274*80 1/50 2.7,5.3 0~360 Open No Lu(2006) 800*500*200 N/A 6.25,1.5,0.25 Normal Suburban Yes Oh(2007) 381*244*122 1/100 3.3,0.3 0~360 Open/Suburban Yes Kopp(2008) 183*207*156 1/50 3,5,6 0~360 Open Yes Ginger(2010) 200*400*100 1/200 1,3.1,6.25,10 Normal Open/Suburban Yes Fan(2011) 900*450*360 1/100 5,10 0~360 Open/Suburban No Mean internal pressure coefficients Fluctuating internal pressure coefficients 0.8 0.4-0.4-0.8 0.6 0.4 0.2 0 5 0.10 0.15 0.20 0.25 Opening ratio 0 5 0.10 0.15 0.20 0.25 Opening ratio (a) Mean internal pressure coefficients Biggest test value in this paper(roof Open Case) Smallest test value in this paper(roof Open Case) Biggest test value from Liu(Wall Open Case) Smallest test value from Liu(Wall Open Case) Biggest test value from Ginger's Field test(wall Open Case) Smallest test value from Ginger's Field test(wall Open Case) Test value from Ginger(Wall Open Case) Test value from Oh(Wall Open Case) Biggest test value from Kopp(Wall Open Case) Smallest test value Kopp(Wall Open Case) Test value from Holmes(Wall Open Case) Biggest test value from Woods(Wall Open Case) Smallest test value Woods(Wall Open Case) Test value from LuDan(Wall Open Case) Biggest test value from Sharma(Wall Open Case) Smallest test value Sharma(Wall Open Case) Biggest test value from Fan(Wall Open Case) Smallest test value Fan(Wall Open Case) Biggest test value in this paper(roof Open Case) Smallest test value in this paper(roof Open Case) Biggest test value from Liu(Wall Open Case) Smallest test value from Liu(Wall Open Case) Biggest test value from Ginger's Field test(wall Open Case) Smallest test value from Ginger's Field test(wall Open Case) Test value from Ginger(Wall Open Case) Test value from Oh(Wall Open Case) Biggest test value from Kopp(Wall Open Case) Smallest test value Kopp(Wall Open Case) Test value from Holmes(Wall Open Case) Test value from LuDan(Wall Open Case) Biggest test value from Sharma(Wall Open Case) Smallest test value Sharma(Wall Open Case) Biggest test value from Fan(Wall Open Case) Smallest test value Fan(Wall Open Case) (b) Fluctuating internal pressure coefficients Fig.6 Comparison of internal pressure between roof opening and wall opening building 945

Comparison with several provisions Several values from main provisions are listed in Tab.3. ASCE 7-10 groups buildings into three categories, namely Enclosed Building, Partially Enclosed Building and Open Building by opening ratio. The Significant Opening Building is assigned the highest mean Cpi as ±0.70 for NBCC-2005.The Enclosed Building is assigned the highest mean Cpi as 0, and the lowest mean Cpi as -0.40 for AIJ-2004. The Enclosed Building is assigned the highest mean Cpi as ±0.20 for GB5009-2012 in China. The minimum and maximum mean Cpi is -0.94 and -7 from current roof corner opening test models. The minimum and maximum mean Cpi is -0.54 and -0.10 from current roof center opening test models. So NBCC overestimated the internal pressure for roof center opening case, and AIJ-2004, GB50009-2012 underestimated the internal pressure for roof center opening case because they did not take account the opening effect into provisions. Table 3 Comparison of Mean internal pressure coefficients between Test and Provisions Comparison Test for roof corner open Test for roof center open NBCC 2005 AIJ -2004 GB50009-2012 ASCE 7-2010 Minimum -0.94-0.54-0.70-0.40-0.20 GCpi=-0.55 Maximum -7-0.10 0.70 0 0.20 GCpi=+0.55 Pressure coefficients from wind tunnel data should be transformed to correspond to ASCE 7-10 (St. Pierre et al. 2005). The equation used to determine the equivalent GCpi from the wind tunnel data is 1 2 ρu H ( GC ) 2 ˆ ˆ Pi eq = CP = F i WTCPi 1 2 ρu10 m,3s gustkztkhkd I 2 (4) Where U 10 m,3 s gust =3 s gust wind speed obtained at a height of 10 m in an open country terrain; Kzt=topographic factor of 1.0; Kd=wind directionality factor of 1.0; Kh=velocity pressure exposure factor; I=importance factor of 1.0; and F WT =resulting pressure conversion factor between wind tunnel data and ASCE (which is 0.39 for the current experiments).the Partially Enclosed Building is assigned the highest GCpi as ±0.55 (G is the gust factor). The Enclosed Building is assigned the highest GCpi as ±0.18. Note that these correspond to peak internal pressure coefficients in the present experiments of ±1.41 and ±0.46 respectively, by applying Eq. 4. For negative peak internal pressure, at the roof corner opening case, the current model results in negative peak internal pressure coefficient for Case 5 at 290 which is -1.80 for roof corner opening, is higher than the equivalence of -0.55 specified in ASCE 7-10 for Partially Enclosed Building and the equivalence of -0.18 specified in ASCE 7-10 for Enclosed Building; at the roof corner opening case, the current model result negative peak internal pressure coefficient for Case 1 at 190 is -1.23 for roof center opening, is smaller than the equivalence of -0.55 specified in ASCE 7-10 for Partially Enclosed Building and higher than the equivalence of -0.18 specified in ASCE 7-10 for Enclosed Building. For positive peak internal pressure, the current model result positive peak internal pressure coefficient for Case 946

1 at 275 is +0.36 for roof center opening, is smaller than the equivalence of +0.55 specified in ASCE 7-10 for Partially Enclosed Building and the equivalence of +0.18 specified in ASCE 7-10 for Enclosed Building. In a word, ASCE overestimated the positive and negative peak internal pressure for roof center opening case, and underestimated the negative peak internal pressure and overestimated positive for roof corner opening case. Conclusion The internal wind effects of six different configurations of roof opening buildings have been investigated. The major conclusions are as follows. The internal pressures in roof opening buildings are highly correlated everywhere, so it is appropriate to use one uniform time-history to represent the internal pressure. The low frequencies of the internal pressure spectra contain the greatest proportion of the energy, and there is a significant Helmholtz resonance at high frequency point for internal pressure. For the small opening ratios (3,5,8%), when the opening is located in windward wall, the mean internal pressures of previous researches have a larger value than the current model. When the opening is located in leeward wall, the mean internal pressures of previous tests are smaller than the current model. For the large opening ratios (15,20,25%), the mean internal pressures of previous tests have a larger value than the current model. The fluctuating internal pressures of previous tests have a larger value than the all cases of current model. NBCC-2005 overestimated the internal pressure for roof center opening case, and AIJ-2004, GB50009-2012 underestimated the internal pressure for roof center opening case because they did not take account the opening effect into provisions. ASCE 7-10 overestimated the positive and negative peak internal pressure for roof center opening case, and underestimated the negative peak internal pressure and overestimated positive for roof corner opening case. Reference J. D. Holmes(1979), Mean and fluctuating internal pressures induced by wind, Proceeding 5th International Conference on Wind Engineering, Fort Collins, Colorado, USA: Pergamon Press,.435-450. Liu H, Rhee K H(1986). Helmholtz oscillation in building models. Journal of Wind Engineering and Industrial Aerodynamics,24(2),95-115 Ginger J D, Mehta K C, Yeatts B B(1997). Internal pressures in a low-rise full-scale building. Journal of wind engineering and industrial aerodynamics,72, 163-174 Ginger J D, Holmes J D, Kim P Y(2010). Variation of internal pressure with varying sizes of dominant openings and volumes. Journal of Structural Engineering, 136,1319-1326 Oh J H, Kopp G A, Inculet D R(2007). The UWO contribution to the NIST aerodynamic database for wind loads on low buildings: Part 3. Internal pressures. Journal of wind engineering and industrial aerodynamics,95(8),755-779 Kopp G A, Oh J H, Inculet D R(2008). Wind-induced internal pressures in houses. Journal of Structural Engineering.134,1129-1138 Woods A R, Blackmore P A(1995). The effect of dominant openings and porosity on internal pressures. Journal of wind engineering and industrial aerodynamics, 57(2-3),167-177 947

Lu Dan(2006). Study on Wind-induced Internal Pressure and Its Effects on Structures. Zhejiang University, Hangzhou, China. Sharma R N, Richards P J(2003). The influence of Helmholtz resonance on internal pressures in a low-rise building. Journal of Wind Engineering and Industrial Aerodynamics, 91(6),807-828 Fan youchuan(2011). Wind Loads and Wind-induced Vibration of Typical Industrial Buildings. Tongji University, Shanghai, China China Engineering Construction Standardization Association(2012). Load code for the design of building structures 5009-2012. Beijing, China. National Research Council of Canada(2005),Canadian Commission for Building and Fire Codes, NRC User s guide NBC, 2005. Structural Commentaries (Part 4), Ottawa, Canada. Architectural Institute of Japan(2004), AIJ Recommendation for loads on buildings. Tokyo, Japan American Structural engineering institute(2010).minimum design loads for buildings and other structures 2010 (ASCE 7-2010). 948