A. Baskaran, PhD, PEng S. Molleti, PhD H. Yes, MEng National Research Council Canada Ottawa, Ontario, Canada ABSTRACT Wind effect on roofs is a complex phenomenon. Poor wind design is one of the common factors in roofing failures. RICOWI’s wind investigation program on Hurricanes Charley and Katrina also confirmed this and challenged designers to develop proper roof wind design tools. To address this issue, the National Research Council of Canada developed a Web-based roof wind design calculator. It is named Wind-RCI (Wind-Roof Calculator on Internet). Wind-RCI can minimize possible mis¬ interpretations of code language. Calculation of cover wind uplift design loads is a function of various parameters, such as roof type, slope, wind speed, building height, roof area, building terrain, building type, and openings. As such, it involves several procedural steps. Wind-RCI integrates these procedural steps into a simple Web¬ based calculator. Demonstrating the viability and flexibility of the Wind-RCI is the objective of this paper SPEAKER Dr. “Bas” Baskaran is a group leader and senior research officer at the National Research Council of Canada, Institute for Research in Construction (NRC/IRC). He has been immersed for 25 years in researching the wind effects on building envelopes through wind tunnel experiments and computer modeling. He also acts as adjunct professor at the University of Ottawa. Mr. Baskaran is a member of RCI, ASCE, SPRI, RICOWI, ICBEST, and CIB technical committees. His work in the area of wind engi¬ neering and building envelopes has received national and international recognition. He has an outstanding research record with more than 150 publications in refereed journals and conference proceedings. Being a professional engineer, Baskaran received his master’s degree in engineering and PhD from Concordia University, Montreal, Canada. Both research topics focused on the wind effects on buildings and earned best dissertation award from the Canadian Society of Civil Engineers. Contact Information: Phone – 613-990-3616; E-mail – bas.baskaran@nrc.ca Baskaran, Molleti, and Yew – 26 Proceedings of the RCI 23rd International Convention
INTRODUCTION Natural wind hazard damages have historically been dramatic, incurring loss of life and property worldwide. Wind-induced roof failure is one of the major contrib¬ utors to insurance claims, and it is rising (AAWE, 1997). Recently, members of the Roofing Commit¬ tee on Weather Issues (RICOWI) completed two major wind investi¬ gations projects, documenting extensive roof damages (Figure 1), providing factual data, and chal¬ lenging designers to develop prop¬ er roof wind design tools. (RICOWI, 2006 and 2007). Simi¬ larly, the Federal Emergency Management Agency (FEMA) also published reports summarizing the observations, conclusions, and recommendations of the Mit¬ igation Assessment Team (MAT) in response to recent hurricanes [FEMA 488,489 (2005) and 549 (2006)]. Wind flow around buildings creates both negative and positive fluctuations over a roofing sys¬ tem. The wind effect on roofing and its response is dynamic. Wind pressure distribution varies spa¬ tially over a roof and can have high suction at the corner and perimeter, due to formulation of vortex and flow separations. Figure 2 illustrates the wind pres¬ sure variation on a building roof. This data represents wind tunnel measurements carried out by the National Research Council Canada (NRC) in the 30 x 30 ft (9 x 9) wind tunnel. These tests used full-scale roofing components [10 x 10 ft (3 x 3 m) in size] at differ¬ ent building heights for wind directions perpendicular to the building face (normal wind) and at 45 degrees (oblique wind) . A PVC roofing system was tested with pressure taps fitted in the PVC single-ply membrane to measure the unsteady pressure loads on the roof (Savage et al., 1996). To calcu¬ late wind up¬ lift loads on roof coverings, designers use wind standards Figure 1 – Roof damage during a high wind event. or building codes experimental data measured by such as the National Building Code of Canada (NBC-2005) and the American Society of Civil Engineers (ASCE 7-05). Code pro¬ visions are a collection of facts testing models in wind tunnels or full-scale measurement data, which are obtained from instru¬ mented structures, field observa¬ tions and consensus of expert and knowledge based on the opinions. Figure 2 – Spatial wind-induced pressure distribution over a roof system. Proceedings of the RCI 23rd International Convention Baskaran, Molleti, and Yew – If sure and internal pressure across the specific component. This can be mathematically expressed as shown in Equation 1. P = lwq(CeCgCp – CeCgiCpl) (1) Where: p = design pressure, l» = importance factor q = reference velocity pressure, Ce = exposure factor Cg = gust factor, Cp = external pressure coefficient Cgi = internal gust factor, Cp< = internal pressure coefficient Equation 1 Wind load calculation for roof coverings involves several proce¬ dural steps, since the design pres¬ sure is a function of various para¬ meters such as roof type, slope, wind speed, building height, roof area, building terrain, building type, and openings. Interpreting wind standards to identify the various parameters is indeed a time-consuming process and it can often lead to misinterpreta¬ tion and misinterpolation of code language. Considering this complexity, the RCI Foundation (RCIF) has offered a research grant to NRC to develop a roof wind design tool. This tool has been named Wind- RCI (Wind – Roof Calculator on In¬ ternet). Wind-RCI was developed based on the National Building Code of Canada (NBC – 2005), and therefore, it is applicable for all Canadian provinces and territo¬ ries. Nevertheless, this paper also presents and discusses the Na¬ tional Roofing Contractors Asso¬ ciation’s (NRCA) recently devel¬ oped wind load calculator based on ASCE 7-05. We will first de¬ scribe the roof covering design pressure calculation, followed by the functionality of the Wind – RCI. Through case studies, we shall demonstrate how this tool integrates several procedural steps involved in the wind load calculation into a simple Web load calcula¬ tor. WIND-RCI NBC (2005) is a model code that sets out techni¬ cal provisions for the design of buildings in Canada. Wind-RCI com¬ putes the wind design using NBC 2005’s Users Guide for Structural Commentaiy. NBC 2005 specifies that for structur¬ al components and cladding, the design wind pressure is the algebraic sum of the external pres- To obtain the design pressure as per Equation 1, a six-step pro¬ cedure was developed by Baskaran and Smith (2005) as follows: 1. Define corner zone. 2. Calculate dynamic pres¬ sure. 3. Calculate external pres¬ sure component. 4. Calculate internal pres¬ sure component. 5. Calculate design pressure (Equation 1). 6. Load diagram. Wind-RCI integrates these procedural steps into a simple Web-based calculator. Table 1 gives a quick snapshot of this Web tool for a hypothetical building located in Toronto, Ontario. The Wind-RCI is capable of performing calculations for various parame¬ ters as listed in Table 2. This includes building height, ranging from low-rise to high-rise; roofs with various slope configurations, from low-sloped to steep-sloped; and building categories both in terms of occupancies and open¬ ings. Thus, the Wind-RCI is capa¬ ble of providing a complete tool for determining the wind design pres¬ sures of a roof. Building Roof Type Height Exposure Openings Importance Type Slope Low rise to medium rise buildings H <=20 m (66 ft) OPEN ROUGH CATEGORY 1 CATEGORY 2 CATEGORY 3 LOW NORMAL HIGH Low Slope a<= 7° Stepped flat a = 0“ Gabled single-ridge a<=7u a>7° Gabled, multiple-ridge a<= 10° a> 10° Monosloped a <= 3U a >3° Sawtoothed a <= 10° a >10° High Rise H > 20 m (66 ft) High-rise Table 2 – Capability of the Wind-RCI for parameter investigation. Baskaran, Molleti, and Yew – 28 Proceedings of the RCI 23rd International Convention Screen # 1 Screen # 2 Roof Calculator on Internet Building Location Province: Ontario City: ■ Toronto (Metropolitan), Toronto If the city you are entering is not available, please select the nearest city {[Next]] Screen # 3 Type of Roof Reference height, h Roof Calculator on Internet Screen # 4 Roof Calculator on Internet Building Dimensions Reference Height, h: ;40 ft Width, w ,60 jft Length. I: ;9Q_ jft | Next | (Conversion Unit: 1 ft = 3.281 m) < Previous oaae R.oof Calculator on Internet Building Exposure Exposure Type: open y] Open: Level terrain with only scattered buildings, trees, or other obstructions. Rough Suburban, urban or wooded terrain Volume 1 – Division 8. Clause 4.1. 7.1. Volume 1 – Division B: Clause 4.1.7.1. ($}(&) (5)(u) * Previous Page Table 1 – Snapshot of Wind-RCI. Proceedings of the RCI 23rd International Convention Baskaran, Molleti, and Yew – 29 Screen # 5 WAIVSIlRIBldQ”- 1^ Roof ■■Mf Calculator on g^ Qga H Internet Building Openings Openings Category: Category 1 y Category 1: Buildings without any large or significant openings, but having small uniformly distributed openings amounting to less than 0.1% of total surface area (User’s Guide, Commentary I, Sentence 31) Category 2: Buildings, in which significant openings, if there are any, can be relied on to be closed during storms. Ex. Low rise buildings (User’s Guide, Commentary I, Sentence 31} Category 3: Building with large openings through which gusts are transmitted to the interior. Ex. Sheds, industrial buildings (User’s Guide, Commentary!, Sentence 31} | Next | < Previous page Screen # 6 Importance Category Importance Category: Law y’ Low: Low human occupancy farm buildings having an occupant load of 1 person or less per 40 m2 of floor area. Minor storage buildings that represent a low direct or indirect hazard to human life in the event of structural failure. (Volume 2 – Division B Appendix A: A-Table 4.1. 2.1) Normal: Buildings equipped with secondary containment of toxic, explosive or other hazardous substances. (Volume 2 – Division B Appendix A: A-Table 4.1. 2.1) High: Buildings containing sufficient quantities of toxic, explosive or other hazardous substances. Example: petrochemical facilities, fuel storage facilities and manufacturing or storage facilities for dangerous goods. (Volume 2 ■ Division B Appendix A: A-Table 4.1. 2.1) | Calculate | < Previous page Screen # 7 Building Location Low sloped roof 0’ « a <= 7′ Building Dimensions L W * Height Width: Length: 40 ft 60 ft 90 ft Roof Type Type of Roof: Province: City. Ontario Toronto (Metropolitan), Toronto Building Openings Openings Category. Category 1 Roof Calculator on Internet Corner, © Edge. ®: Field,© (Conversion Unit: 1 ft = 3.281 m, 1 psf = 478 Pa) < Previous page |’■ ■ ar .uteti on Roof Cladding Wind Design Loads End Zone Width, z: 6 ft -49 psf -23 psf Building Exposure Exposure Type: Open Building Importance Category Category Type: Low Table 1 – Snapshot of Wind-RCI. Baskaran, Molleti, and Yew – 30 Proceedings of the RCI 23rd International Convention RoofWindDesigner (WWW.ROOFWINDDESIGNER.COM) Similar to Wind-RCI, the RoofWindDesigner is also a Web¬ based roof wind load calculator. RoofWindDesigner applies the American Society of Civil Engineers’ (ASCE) Standard No.- 05 “Minimum Design Loads for Buildings and Other Structures.” It was developed by the NRCA in cooperation with the Midwest Roofing Contractors Association (MRCA) and the Northeast Roofing Contractors Association (NERCA). Apart from computing the roof covering design pressures, it also recommends the roof covering uplift resistance by multiplying the calculated design pressures with a minimum safety factor of 2 (ASTM D6630). It is worth noting that the RoofWindDesigner is lim¬ ited to calculating wind uplift on low-rise buildings [the mean roof height, h, must be less than or equal to 60 feet (h < 60 ft)]. More¬ over, one can use the RoofWind¬ Designer only for low-slope config¬ urations (roof incline less than 7.5 degrees, which is about 1-1/2:12 slope). sure; and 2. One that is pertinent to the building – effects of building height, openings and roof slope. Using the Wind-RCI, all the parametric investigations are car¬ ried out through case studies. For discussion purposes, the paper presents design pressures only for the roof-corner zones and not for the edge and field zones. Based on this approach, it is easier to understand the effect of the influ¬ encing parameters on the roof¬ design pressures. Designers can develop a similar inference for the edge and field zones with only variation in the magnitude of the design pressure. This is due to the fact that the intensity of pressure moderates as one moves away from the corner towards the field, as discussed in Figure 2. Effect of Wind Speed on Wind Uplift Pressures NBC 2005 uses mean hourly wind speed to determine the refer¬ ence dynamic pressures, q. The dynamic pressures are based on the probability of being exceeded per year of 1 in 50. This is tabu¬ lated for major Canadian cities in Appendix C of the NBC (2005). It is worth mentioning that there is no wind speed map in NBC (2005) that is similar to the ASCE 7 (2005). Nevertheless, one can esti¬ mate the wind speed for a given dynamic pressure using the fol¬ lowing equation. q = CV2 (2) Where: q = reference wind pressure C = factor which depends on atmospheric pressure and air temperature V = wind speed Equation 2 To investigate the effect of wind speed, a 50-ft (15-m) build¬ ing is selected. The building has an importance factor of 1, open¬ exposure condition, and category 1 as an internal opening provi¬ sion. Beta version of the Wind-RCI can perform calculations for the provinces of British Columbia, Parametric Investigations Using Wind-RCI To demon¬ strate the viability and flexibility of the Wind-RCI, var¬ ious parameters that can influence design pressures are studied. These parameters can be grouped into two segments: 1. One that is pertinent to the up¬ coming wind – ef¬ fects of wind speed and ter¬ rain expo- 1 mph = 1.6 kmph Figure 3 – Effect of wind speed on wind uplift pressures. Proceedings of the RCI 23rd International Convention Baskaran, Molleti, and Yew – 3 1 Ontario, and Quebec. For each province, using the Appendix C of NBC (2005), two cities are select¬ ed such that they represent the minimum and maximum i.e., q™n, and qmax dynamic pressure levels. Equation 2 was used to calculate the respective wind speeds. With these assumptions, Figure 3 illus¬ trates the effect of wind speed on the design pressures. As shown in Figure 3, Harrington-Harbor in Quebec has the highest mean hourly wind speed of 89 mph (142 kmph), and therefore, its roof cor¬ ner zone has its highest design pressures of 113 psf (5.4 kPa), while Armstrong in Ontario expe¬ riences the lowest mean hourly wind speed of 48 mph (77 kmph) and has a corner design pressure of 33 psf (1.6 kPa). Therefore, the data clearly indicate that the roof cladding design pressures are directly proportionate to the upcoming wind speed. For the benefit of U.S. readers, equivalent 3-second gust wind speeds are also calculated from the Canadian Figure 4 – Exposure types as per NBC (2005). mean hourly wind speed and included in Figure 3. Effect of Exposure on Wind Uplift Pressures “Exposure” refers to the ground roughness surrounding the building. The exposure factor characterizes the turbulence gen¬ erated by the ground roughness; therefore, it can directly reflect changes in wind speed. NBC 2005 defines exposure in terms of open and rough (urban) as shown in Figure 4. Designers often face challenges in identifying the ap¬ propriate exposure condition dur¬ ing the wind load calculations. Building exposure can change from the design stage to construc¬ tion stage when the surrounding neighborhood develops and Agassiz Newcastle (Bowmanville) Harrington-Harbour British Columbia Ontario Quebec Figure 5 – Effect of exposure on wind uplift pressures. Baskaran, Molleti, and Yew – 32 Proceedings of the RCI 23rd International Convention 180 Figure 6 – Effect of building height on wind uplift pressures. changes the design building land¬ scape. In those situations, it may be useful for the designer to vali¬ date the original wind load esti¬ mation. Wind-RCI makes this process much easier. To demonstrate the exposure effect, in addition to the open country exposure, calcula¬ tions are performed using the Wind-RCI for urban exposure keeping the building configura¬ tion and other assumptions the same as that previously men¬ tioned. Figure 5 shows such cal¬ culated wind uplift pressures. Comparison of the data in Figure 5 clearly indicates that the building located in open-exposure conditions has higher wind¬ design pressure when compared to the building in urban exposure. With all the parameters kept con¬ stant, the data also indicate that the roof constructed in open¬ exposure condition has 45 per¬ cent more roof-corner design pressure in comparison to the building in urban-exposure con¬ dition. Even though the data are presented for the roof-corner zone, the trend will be similar for the other two zones – namely, edge and field. Effect of Building Height on Wind Uplift Pressures NBC (2005) provides for two sets of building heights: h <= 66 ft (20 m) and h >66 ft (20 m). To demonstrate the capabilities of the Wind-RCI, two additional building heights (30 and 70 ft) are selected. This can provide an indi¬ cation of how much the wind¬ uplift load changes when there is change of 20 ft (6 m) from the base configuration building, which is 50 ft (15 m) as presented above. Figure 6 shows the com¬ puted design pressures for these three buildings. As per the NBC 2005, 30-ft (10-m) and 50-ft (15- m) buildings are low-rise build¬ ings, while the 70-ft (20-m) build¬ ing is a high-rise building. The graphical plot in Figure 6 shows an increase in the design pressure as the building height increases. This is due to the fact that in¬ crease in height will increase the wind speed. As presented in Equa¬ tion 2, the dynamic pressure will increase by the square of the wind speed. Changes in building height also change the corner zone (©) shape as depicted in Figure 7. Effect of Building Openings on Wind Uplift Pressures As discussed under Equation 1, CeCgiCpi refers to the internal pressure component. The wind design pressure is calculated based on the algebraic sum of the external- and internal-pressure components. The magnitude of this internal-pressure component, depends on the amount and dis¬ tribution of the openings in a building. It is rather difficult to quantify the size and distribution of openings for a given building due to variation in construction process and components used. To account for these uncertainties, NBC 2005 provides ranges for the internal pressure coefficient in three categories as follows: Proceedings of the RCI 23rd International Convention Baskaran, Molleti, and Yew – 33 w h=50 ft z = 10% of the least horizontal dimension and 40% of height, but not less than 3.3 ft (1m) Agassiz = 92 psf New Castle = 74 psf Harrington =113 psf Harbour h=30 ft Agassiz = 83 psf New Castle = 66 psf Harrington =102 psf Harbour h=70 ft z =10% of width Agassiz =105 psf New Castle = 84 psf Harrington = 129 psf Harbour Figure 7 – Comparison of corner zone – low-rise vs. high-rise building. • Category 1: without large or significant openings. Cpi: -0.15 to 0.0 • Category 2: significant openings that can be closed during storms. CPi: -0.45 to 0.3 • Category 3: large open ings through which gusts are transmitted. Cpi: -0.7 to 0.7 With Wind-RCI, one can easily modify from one category to other during wind uplift calculations. Previous examples above used Category 1 for internal pressure coefficient selection. Wind-RCI repeated the computations by selecting the other two categories. Figure 8 shows the calculated wind uplift pressures. The building openings can influence the air intrusion (Dregger 1991, Baskaran et al. 2003, Molleti 2005) into a roofing assembly. Buildings with large openings and roof assemblies with air permeable components can significantly increase the design pressures. This relation¬ ship is clearly shown in Figure 8, with all the parameters kept con¬ stant, the corner design pressure increased by 10% and 26% when a Category 1 building changed to Category 2 and 3 respectively. A building’s internal pressure characteristics can also change during its service life. This can be a small variation when the HVAC requirement changes or moderate when building envelopes are refurbished. Also the internal pressure can change drastically and suddenly during catastrophe event such as window/ door fail¬ ure during high wind events (hur¬ ricanes). Baskaran et al. (2007) investigated several such failures as part of the RlCOWI’s wind investigation program. In those scenarios, a roof assembly de¬ signed as Category 1 can experi¬ ence significantly higher uplift due to sudden increase in the internal pressure component. Therefore, it is quite critical for the designer to understand this transition and design the roof assembly accordingly. Based on the field investigation data, Bas¬ karan et al. (2007) recommends that building with high probabili¬ ty of envelope (such as garage doors) failures during high wind events can be designed with the Category 3 internal pressure coef¬ ficient value. This conservative approach is a good design prac¬ tice for wind mitigation measures provided that glazing and doors are not designed for windborne debris. Baskaran, Molleti, and Yew – 34 Proceedings of the RC1 23rd International Convention British Columbia Ontario Quebec Figure 8 – Effect of building openings in wind uplift design pressures. Effect of Roof Slope on Wind Uplift Pressures Building aerodynamics changes when there is change in the roof slope. When the upcoming streamlines of wind hit the roof, the pres¬ sure distribu¬ tion on the roof surface depends upon the extent to which the roof slope modi¬ fies the airflow. This can alter the wind pres¬ sure distribu¬ tion and magni¬ tude. It is worth mentioning that the distribution presented in Figure 2 corresponds to a low slope-roof configuration (± < = 100), as do all the above calcula¬ tions. Wind-RCI can perform cal¬ culations for various roof slope configurations as listed in Table 2. Note that this is one of the major differences between the Wind-RCI and RoofWindDesigner, where the later is limited to only low-slope roof configuration. To demon- Britlsh Columbia Quebec Figure 9 – Effect of roof slope on wind uplift pressures. Proceedings of the RCI 23rd International Convention Baskaran, Molleti, and Yew – 35 strate the effect of roof slope, two additional roof slopes [medium (100 < ± < = 300) and steep (270 < ± < = 450)] are selected and calcu¬ lations are repeated with the Wind-RCI. All other parameters are kept similar to that of those listed above. Figure 9 shows the effect of roof slope on the computed design pressures. It is clear from Figure 9 that the roof slope modifies the design pressure and it is not directly proportional to the increase in the roof slope. This observation is different from the observation of previous parame¬ ters – namely the exposure, height, and building openings had on the design pressure. One can¬ not assume a linear relationship between roof slope and design pressure; i.e., increasing or de¬ creasing the roof slope cannot necessarily increase or decrease the design pressures. Alternative¬ ly, whenever there is change in the roof slope, new calculations are warranted. As presented, the Wind-RCI makes that process easier. CONCLUSION Wind-RCI is developed based on the National Building Code of Canada 2005 (NBC). NBC (2005) is a model code that sets out tech¬ nical provisions for the design of buildings in Canada. In the Wind- RCI, wind design loads are com¬ puted using NBC 2005’s Users Guide for Structural Commentary. A professional engineer certified the released beta version. Some of the benefits of this Web-based tool are as follows: • Obtaining design loads at the palm of a designer’s hand. • Minimizing the complexity of the referenced code and the involved calculations. The design pressures calculat¬ ed by the Wind-RCI should be multiplied by an appropriate safe¬ ty factor to obtain the required system resistance. Wind-RCI also has certain limitations. It does not provide calculations for the topo¬ graphic/ terrain influences on a building, such as a structure sit¬ uated on hills and escarpment. Also, it cannot calculate design pressures for hipped roofs, for post disaster building configura¬ tions, and for roofs with over¬ hangs. ACKNOWLEDGEMENTS The presented research is being carried out for a consor¬ tium, the Special Interest Group for Dynamic Evaluation of Roofing Systems (SIGDERS), formed from a group of partners interested in roofing design. These partners include: Manufacturers: Atlas Roofing Corporation, Canadian General Tower Ltd., Carlisle SynTec, GAF Materials Corporation, Firestone Building Products Co., IKO Indus¬ tries Ltd., Johns Manville, Sika- Sarnafil, Soprema Canada, Ste¬ vens Roofing, Tremco, and Trufast. Building Owners: Canada Post Corporation, Department of Na¬ tional Defence, Public Works and Government Services Canada. Industry Associations: Cana¬ dian Roofing Contractors’ Asso¬ ciation, Canadian Sheet Steel Building Institute, National Roof¬ ing Contractors Association, RCI, Inc., and Roofing Contractors Association of British Columbia. REFERENCES American Association for Wind Engineering (AAWE), report on the “Workshop on Large-scale Testing Needs in Wind Engineering,” Washington, DC, 1997. American Society of Civil Engineers, Standard ASCE 7-2005, “Minimum Design Loads For Buildings And Other Structures.” Baskaran, A., Molleti, S., and Sexton, “Wind Uplift Per¬ formance of Mechanically Attached Roofing Systems with Vapor Barrier,” 9th Conference on Building Science and Technology, Vancouver, British Colum¬ bia, Canada, February 27- 28, 2003. Baskaran, A. and T.L. Smith, “A Guide for the Wind Design of Mechanically Attached Flexible Mem¬ brane Roofs.” National Re¬ search Council of Canada, Ottawa, Ontario, Canada, KIA 0R6, 2005. Baskaran, A., Molleti, S., and Booth, R.J., “Understand¬ ing Air Barriers in Mechan¬ ically Attached Low-Slope Roofing Assemblies for Wind Uplift,” Pp. 443-450 in Proceedings of the 3rd International Building Phy¬ sics/ Science Conference, Montreal, August 2006. Baskaran, A., Molleti, S., Roodvoets, D., “Under¬ standing Low-slope Roofs under Hurricane Charley from Field to Practice,” 6th Symposium on Roofing Research and Standards Development, Tampa, Flor¬ ida, December 01, 2007, pp. 1-25. Dregger, Phil D., “Role of Air Retarders Deserves Closer Scrutiny,” Professional Roofing, p 46-49, October 1991. Federal Emergency Manage¬ ment Agency (FEMA)-488, “Mitigation Assessment Team Report: Hurricane Charley in Florida,” 2005. • Offering designers a faster, reliable tool for wind cal¬ culations. Baskaran, Molleti, and Yew – 36 Proceedings of the RCI 23rd International Convention Federal Emergency Manage¬ ment Agency (FEMA)-489, “Hurricane Ivan in Ala¬ bama and Florida: Obser¬ vations, Recommendations and Technical Guidance,” 2005 Federal Emergency Manage¬ ment Agency (FEMAJ-549, “Hurricane Katrina in the Gulf Coast: Mitigation As¬ sessment Team Report, Building Performance Ob¬ servations, Recommenda¬ tions, and Technical Guid¬ ance,” 2006. Molleti, S., “Performance Evaluation of Mechanically Attached Roofing Systems.” PhD thesis, University of Ottawa, Ottawa, Canada, 2006. National Roofing Contractors Association, “Hurricane Charley: A Preliminary Re¬ port,” Professional Roofing, NRCA, October 2004. NRC (National Research Council). 2005. National Building Code of Canada, Part 5. Ottawa: National Research Council of Cana¬ da, Ottawa, Ontario, Cana¬ da, KIA 0R6. Roofing Industry Committee on Weather Issues, Inc., Hurricanes Charley and Ivan Wind Investigation Report, Powder Springs, GA, 2006. Roofing Industry Committee on Weather Issues, Inc., Hurricane Katrina Wind Investigation Report, Pow¬ der Springs, GA, pp. 183, August 01, 2007. Savage, M.G, Baskaran, A., Cooper, K.R., and Lei, W., “Pressure Distribution Da¬ ta Measured During the November 1994 Wind Tun¬ nel Tests on a Mechanically Attached, PVC Single-Ply Roofing System,” IAR Re¬ port LTR-A-003, National Research Council, Canada, 1996. Proceedings oj the RCI 23rd International Convention Baskaran, Molleti, and Yew – 37
