2 6 T H RC I I N T E R N A T I O N A L C O N V E N T I O N A N D T R A D E S H OW • A P R I L 7 – 1 2 , 2 0 1 1 K O O N T Z • 2 7 THE CORRELATION BETWEEN WIND RESISTANCE AND PHYSICAL PROPERTIES OF FIBERGLASS SHINGLES BY JIM D. KOONTZ, RRC, PE JIM D. KOONTZ & ASSOCIATES, INC. 3120 N. Grimes St., Hobbs, NM 88240 Phone: 575-392-7676 • Fax: 575-392-7602 • E-mail: jim@jdkoontz.com ABSTRACT Fiberglass shingles constitute the primary roofing material used for steep-sloped structures. During high wind events, shingle performance is critical to building protection. Shingle failure accounts for millions of dollars of insurance claims every year. The International Building Code (IBC) requires that the manufacture of fiberglass shingles comply with physical property standards of ASTM D3462. To test the wind resistances of shingles, ASTM D3161, the Fan-Induced Method, is used. This method subjects shingle test decks to varying wind speeds. This research reviews the relationship between fiberglass shingle physical properties and the relative wind resistance of the shingle. Test results will be compared to shingle manufacturers marketing claims and IBC requirements. SPEAKER BY JIM D. KOONTZ, RRC, PE — JIM D. KOONTZ & ASSOCIATES, INC. JIM D. KOONTZ, PE, RRC, has been involved in the roofing industry for 50+ years. Mr. Koontz obtained a bachelor of science degree in engineering and a master’s in business administration from Tulane University in New Orleans. He is a registered engineer in multiple states. Jim began his career as a roofer and roofing contractor in 1960, and in 1976, he started one of the first roof engineering firms in the United States. As a researcher, Jim has authored numerous technical articles on roofing, covering a variety of technical subjects. Mr. Koontz has been a guest speaker for organizations such as NRCA, ASTM, RCI, international symposia, and others. As a consulting engineer, he has worked in over 40 states, Canada, Mexico, and the Caribbean. Notable projects include the Kingdome in Seattle, the Denver International Airport, and the Church Street USPO at ground zero in New York in the aftermath of 9/11. 2 8 • K O O N T Z 2 6 T H RC I I N T E R N A T I O N A L C O N V E N T I O N A N D T R A D E S H OW • A P R I L 7 – 1 2 , 2 0 1 1 2 6 T H RC I I N T E R N A T I O N A L C O N V E N T I O N A N D T R A D E S H OW • A P R I L 7 – 1 2 , 2 0 1 1 K O O N T Z • 2 9 INTRODUCTION Fiberglass shingles constitute the primary roofing material used for steep-sloped roof structures within North America. Shingles are used in a wide variety of applications, including single-family homes, multitenant structures, commercial- and institutional-type buildings. There is an expectation by the designer, installer, and the user that fiberglass shingles will perform during wind events according to the claims of the marketing literature of shingle roofing manufacturers and the requirements of local and national building codes. Failures of shingles during wind events and subsequent damage to buildings on which they are installed account for millions of dollars of insurance industry claims on a yearly basis. Obviously, during high wind events such as hurricanes, shingle performance is critical to the protection of building structures and their contents. See Photo 1. This study reviews the relationship between fiberglass shingle physical properties and wind resistance. Several commonly used new fiberglass shingles from various manufacturers throughout North America were initially tested for physical properties following American Society for Testing and Materials (ASTM) procedure D3462. The shingles were installed over plywood test decks in accordance with the various shingle manufacturers’ installation instructions. A large oven was used to heat the test decks to a test-specified temperature in order to facilitate self-sealing adhesive strip activation. The test decks were then subjected to wind speeds up to 150 mph until failures occurred. A comparison was then made between the wind speed resistance and the physical properties of the shingles. HISTORY Three-tab fiberglass shingles were first introduced into the roofing market in the late 1970s and early 1980s as a substitute for conventional three-tab organic shingles. The newer three-tab fiberglass shingles relied upon the use of fiberglass mats to be an equal-to or better-than product when compared to the use of conventional organic mats. In the early 1990s, splitting problems were reported with fiberglass shingles.1 As a result, numerous technical articles were published and seminars held that discussed these shingle failures. Splitting failures were attributed to a variety of factors that included workmanship, lack of proper below-deck ventilation, tab adhesion, thermal expansion/ contraction of the shingles and the deck, and insufficient tensile strength and tear resistance. Analysis of the relationship of these physical properties of shingles to wind resistance has been limited. Several factors that can affect the wind resistance of fiberglass shingles include installation errors, the design of a shingle, and the quality control of the shingle when manufactured. The International Building Code (IBC) requires that fiberglass shingles be manufactured to comply with various physical property standards set forth in ASTM D3462. Typical physical properties include the following: masses of the shingle, the matt, asphalt, and filler in a per unit area. Other minimum physical properties include: nail pull-through and tear resistance. Consumers, contractors, and roof designers take for granted that fiberglass shingles are manufactured in compliance with code requirements, and they rely upon manufacturers’ marketing representations and claims. The test wind speeds for shingles, once deemed adequate to predict performance, have changed dramatically since Hurricane Andrew in August of 1992. In the late 1980s and early 1990s, the maximum wind speed used in shingle testing was 60 mph. This was recognized as inadequate and unrealistic compared to real-world wind events. Today’s standards require testing at speeds ranging from 60 to 150 mph. SHINGLE INSTALLATION The number and placement of fasteners attaching the shingle to the substrate is critical to the performance of fiberglass shingles. Nails installed in the shingle too high or at improper locations can affect a shingle’s wind resistance. Obviously, nails placed too high, engaging only one layer of a shingle as opposed to two layers, are more THE CORRELATION BETWEEN WIND RESISTANCE AND PHYSICAL PROPERTIES OF FIBERGLASS SHINGLES Photo 1 – Shingle failure. likely to fail at lower wind speeds. The pullthrough resistance of a nail installed in only one layer of a shingle is obviously less than that of a properly placed nail. Thomas Smith studied the subject of acceptable fastener location variation in 1999.2 One of Smith’s recommendations was for shingle manufacturers to provide clear “nail lines” so that contractors could secure shingles at the proper locations. Shingle placement and exposure can also be important to the wind resistance of fiberglass shingles. Shingles installed with too much exposure and, therefore, a lack of headlap, are more vulnerable to failure during wind events. See Figure 1.3 Properly driven fasteners are critical for suitable performance of shingles during wind events. Improperly installed nails can be underdriven, overdriven, installed crookedly, or outside the nail zone. Tab adhesion by way of the self-sealing adhesive strip can be affected by fasteners either underdriven or installed crookedly. See Figure 2.4 Overdriven fasteners and fasteners installed crookedly can break the fiberglass matt at or near the nail head, resulting in a loss of nail pull-through resistance. Typically, a minimum of four nails, evenly spaced and just below the self-sealing adhesive strip, is required for conventional three-tab fiberglass shingles. Specific recommendations in high wind zones are somewhat vague. Several manufacturers require and show up to six fasteners per shingles in “high wind areas. The decision of what is considered a high wind area is left up to the contractor, consumer, and the design professional. The building codes refer back to recommendations of the shingle manufacturer for proper installation. The timing of installation of shingles can also be critical. The self-sealing adhesive strip may not activate until spring for shingles installed in the late fall or winter because of cold weather. In some areas of the country, windstorms can deposit dust beneath shingle tabs before sealing occurs, resulting in imperfect bonding strength.5 BUILDING CODES Designers, contractors, and consumers assume that fiberglass shingles will provide a reasonable degree of performance during expected wind events as required by local building code. Maps found in the IBC depict expected basic wind speed for a 50-year mean recurrence interval for all areas of the United States and territories. Refer to Figure 3, IBC R301.2(4). Once a wind speed is determined from the IBC R 301.2(4) map, reference is then Figure 1 — Proper and improper nailing. Figure 2 — Improperly driven fasteners. Figure 3 — IBC map. Figure 4: Table 905.2.4.1(1) Maximum Basic Classification ASTM Type, Class Wind Speed Requirement Wind Speed (mph) Figure 301.2(4) (mph) 85 D, G, or H D = 90, G = 120, H = 150 90 D, G, or H D = 90, G = 120, H = 150 100 G or H G = 120, H = 150 110 G or H G = 120, H = 150 120 G or H G = 120, H = 150 130 H H = 150 140 H H = 150 150 H H = 150 3 0 • K O O N T Z 2 6 T H RC I I N T E R N A T I O N A L C O N V E N T I O N A N D T R A D E S H OW • A P R I L 7 – 1 2 , 2 0 1 1 CLASSIFICATION OF ASPHALT SHINGLES PER ASTM D7158 made to determine the proper shingle classification based upon the basic wind speed. Refer to IBC Tables 905.2.4.1 (1) and (2), Figures 4 and 5. Depending upon the test procedure followed, a letter classification is then designated for a shingle: D, G, or H for ASTM D7158; or A, D, or F for ASTM D3161. The test wind speed is then listed within the ASTM standards. Depending upon performance characteristics, shingles are consequently marketed by manufacturers with specific wind-resistance classifications, e.g., A = 60 mph, D = 90 mph, or F = 110 mph. WIND TESTING, ASTM D7158 AND D3161 The wind resistances of shingles are tested in accordance with ASTM D7158, which involves measuring pressure differences above and below shingles and performing mechanical uplift resistance using a tensile-testing apparatus described in ASTM D6381. Calculations are then required to determine the relative uplift resistance of a shingle. The test procedure described was developed by the Asphalt Roofing Manufacturers Association (ARMA) in 1991 at a time when the prior shingle testing that was performed at wind speeds up to 60 mph was believed to be inadequate.6 See Photo 2. The Fan- Induced Method, ASTM D3161, involves building shingle test roof decks and then subjecting the shingles to various wind speeds. The test decks are conditioned at a temperature of 57° to 60°C (135° to 140°F) for a continuous period of 16 hours at a slope of 2 in per ft or 17%. The temperature- conditioned test decks are then placed in front of the fan outlet on an incline of 2 in per horizontal ft (17% slope). The fan delivers a horizontal stream of air through a rectangular opening 3-ft wide by 1-ft high, placed 7 in from the leading edge of the third course of shingles. See Photo 3. The test decks are then subjected to various wind speeds. Failure is defined in ASTM D3161 as the inability to restrain any portion of a shingle lifting that would prevent the shingle from standing upright or bending back on itself. Depending upon performance characteristics, shingles are consequently marketed by manufacturers with specific wind resistance classifications: A = 60 mph, D = 90 mph, or F = 110 mph. Shingle manufacturers will typically represent, in writing, that their shingles meet certain ASTM and IBC standards; however, shingle manufacturers’ warranties may actually be written for a lower wind speed than the wind speed listed in the IBC or ASTM classification. Prior research7 utilizing pressure-sensing taps has studied the wind flow mechanism that produces uplift on shingles. This research has shown that as wind flows up and over the surface of a shingled roof, forces exerted by the wind vary depending upon location. Refer to Figure 6.8 Negative or uplift pressure develops just above and behind the leading edge of the tab of the shingle. In front of the butt edge of the shin- Photo 2 — Wind testing. Photo 3 — Fan. 2 6 T H RC I I N T E R N A T I O N A L C O N V E N T I O N A N D T R A D E S H OW • A P R I L 7 – 1 2 , 2 0 1 1 K O O N T Z • 3 1 CLASSIFICATION OF ASPHALT SHINGLES PER ASTM D3161 Figure 5: Table 905.2.4.1(2) Maximum Basic Classification ASTM Type, Class Wind Speed Requirement Wind Speed (mph) Figure 301.2(4) (mph) 85 A, D, or F A = 60, D = 90, F =110 90 A, D, or F A = 60, D = 90, F =110 100 A, D, or F A = 60, D = 90, F =110 110 F F = 110 120 F F = 110 130 F F = 110 140 F F = 110 150 F F = 110 gle, a stagnation region develops that creates downward pressure. Prior research was instrumental in the development of ASTM D7158, which calculates wind resistance of asphalt shingles. SHINGLES’ MARKETING LITERATURE Each shingle of every manufacturer used in this research represented compliance with a letter classification listed in ASTM D3161, ASTM D7158, or both. Seven out of eight manufacturers indicated compliance with ASTM D3462. The manufacturers also listed a wind speed warranty in miles per hour (see Figure 7). LABORATORY TESTING, PHYSICAL PROPERTIES – ASTM D3462 Jim D. Koontz & Associates, Inc. acquired new, three-tab fiberglass shingles from five different commercially available retail or supplier sources. Prior to the installation of the shingles on the decks for wind testing, the physical properties of the shingles were determined. • Mat weight • Tensile strength • Tear strength • Fastener p u l l – through: s i n g l e l a y e r , double layer Small samples from the fiberglass shingles were placed in a bath of hot, circulating, organic solvent. The asphalt and granules were extracted from each sample until a clean mat was obtained. The samples of the desaturated fiberglass mat were dried and weighed. The desaturated dry fiberglass mat weights ranged from a low of 1.71 lbs/100 ft2 to a high of 2.07 lbs/100 ft2. All of the shingle samples tested contained fiberglass mats that exceeded the ASTM D3462 minimum desaturated mat weight requirement of 1.35 lbs/100 ft2. The shingles tested had a gross weight ranging from a low of 74.7 lbs/100 ft2 to a high of 101.9 lbs/100 ft2. All of the shingles exceeded the ASTM D3462 total minimum weight requirement of 70.0 lbs/100 ft2. See Photo 4. The shingles were also tested for tensile strength. Tensile strength is not listed as a requirement under ASTM D3462. The tensile strength test was performed for purposes of information and comparison. The tensile strengths ranged from a low of 84.0 lbf to a high of 126.0 lbf. See Photo 5. Fastener pull-through testing was performed with a single layer and two layers of shingles. The purpose of employing two methods of testing was to simulate improper and proper fastener placement. Obviously, if a shingle were nailed above the recommended nail zone, only one layer would be penetrated. The single-layer pullthrough resistance ranged from a low of 20.0 lbf to a high of 33.0 lbf. The ASTM D3462 pull-through requirement for a single shingle layer is a minimum of 20.0 lbf. The double shingle layers’ pull-through resistance ranged from 40.0 lbf to 70.0 lbf. Figure 6 — Shingle separation. Figure 7 Shingle ASTM ASTM D3161 ASTM D7158 Manufacturer Manufacturer D3462 Classification Classification Warranty (mph) 1 Yes A, F H 60 2 Yes F H 60 3 — A H 54 4* Yes F — 100 5 Yes F H 60 6 Yes F H 60 7 Yes F H 60 8 Yes F H 60 3 2 • K O O N T Z 2 6 T H RC I I N T E R N A T I O N A L C O N V E N T I O N A N D T R A D E S H OW • A P R I L 7 – 1 2 , 2 0 1 1 Photo 4 — Mat weight. Photo 5 — Tensile strength. *Modified Asphalt Shingle A = bundle; F = literature ASTM D3161: A = 60 MPH, F = 110 MPH; ASTM D7158: H = 150 MPH See Photos 6A and 6B. The Elmendorf Tear Test, ASTM D1922, was performed on each of the eight new shingle samples. ASTM D1922 is one of the ASTM Standards referenced in D3462. The tear strength ranged from a low of 1,290.0 g to a high of 2,400.0 g. Under ASTM D3462, the minimum tear strength requirement is 1,700.0 g. Seven of the eight shingle samples tested did not meet the minimum ASTM requirement. See Photo 7. Seven of the eight shingles that fell below the 1,700 g tear resistance requirement were packaged in wrappers that stated the shingles were manufactured in accordance with ASTM D3462 standards. Most importantly, the IBC requires that shingles meet the minimum standards of ASTM D3462, and that includes D1922. A summary of the physical property test is included within Figure 8. LABORATORY TESTING, WIND RESISTANCE – ASTM D3161 Eight test roofs/decks were carefully constructed consisting of ½-in plywood covered with a Type 15 organic felt underlayment and overlaid with different, new fiberglass shingles. Each shingle was attached to the test deck with four roofing nails at the manufacturer’s specified usual locations just below the self-sealing adhesive strip. Each test roof was then conditioned following ASTM D3161 in a large oven for 16 hours at a temperature of 57° to 60°C (135° to 140°F) in order to activate the self-sealing adhesive strip. Initial testing showed that the conditioning time and temperature did not result in an adequate activation of the self-sealing adhesive strip. This resulted in a poor bond of the shingle layers, typically found in field inspection of real-world projects. The shingles conditioned for 16 hours at a temperature of 57° to 60°C (135° to 140°F) failed at 60 mph. The test samples were then reheated for 16 hours at 160°F, and tested after reheating. Prior research by William Cullen9 also found that conditioning the shingles at higher temperatures over a longer period of Figure 8 Shingle Tear Tensile Net Desaturated Pull Pull Sample Strength Strength Weight Mat Weight Through Through Single Double Layer Layer (g) (lbf) (lbs/sq) (lbs/sq) (lbf) (lbf) 1 1,670 99 74.7 1.77 25 48 2 1,450 91 84.6 2.04 25 55 3 1,290 87 81.8 1.96 23 45 4* 2,400 87 101.9 2.00 26 54 5 1,960 126 80.5 1.89 33 70 6 1,430 84 93.1 2.07 22 40 7 1,870 87 83.6 1.83 23 49 8 1,430 84 82.3 1.71 20 49 Average 1,688 9893 85.3 1.91 25 51 ASTM 1,700 — 70.0 1.35 20 — Minimum 2 6 T H RC I I N T E R N A T I O N A L C O N V E N T I O N A N D T R A D E S H OW • A P R I L 7 – 1 2 , 2 0 1 1 K O O N T Z • 3 3 Photos 6A and 6B — Fastener pull-through. Photo 7 — Elmendorf tear test. *Modified Asphalt Shingle time enhances self-sealing adhesive strip bond strength. The ASTM D3161 basic test procedure requires that a test deck be exposed to one wind speed for two hours; however, it is not prohibited to use different test velocities and different time intervals. In this research, each test deck was placed in front of a blower, and it was then tested at wind speeds that began at 60 mph and increased in 30 mph increments every thirty minutes, up to a maximum of 150 mph. The mode of failure of the different shingles on each test deck was recorded (Figure 9). According to ASTM D3161, failure is defined as follows: The end point for failure shall be taken as the time at which the sealing feature fails to restrain one or more full shingle tabs. In addition, no free portion of a shingle shall lift so as to stand upright or bend back on itself during the test. For purposes of this research, failures of the shingles were described by two different modes: Mode 1 — Tab loose, unbroken, meets the ASTM D3161 minimum Mode 2 — Full shingle detachment, exposure of felt underlayment, wood deck, or both Seven of eight shingles tested complied with the wind speed rating as required under ASTM D3161. Two of eight shingles rated under ASTM D7158 performed well at a wind speed of 150 mph; six shingles, however, failed. The shingles tested performed at wind speeds that exceeded the velocities listed in the warranties. The test data show very little correlation between the wind resistance and the desaturated mat-weight of the shingle. There is also very little correlation between the gross shingle weight and the wind resistance of the shingle. The only exception was Shingle Sample No. 4, the modified asphalt shingle, which had the highest total gross shingle weight. Once the self-sealing adhesive strip has failed by separating, it is clear that shingles with higher fastener pull-through resistance and higher tear strength have a higher destructive wind-speed resistance (refer to Figures 10 and 11). Shingle Mode 1 Mode 2 Class Class Warranty Tab Loose Underlayment D3161 D7158 mph mph or Deck mph mph Exposure mph 1 110 130 60/110** 150 60 2 130 130 110 150 60 3 90 110 60 150 54 4* 150+ 150+ 110 — 100 5 150 150 110 150 60 6 90 110 110 150 60 7 130 130 110 150 60 8 130 130 110 150 60 Figure 9 Figures 10 and 11 — Tear strength/wind speed and pull-through/wind speed. 3 4 • K O O N T Z 2 6 T H RC I I N T E R N A T I O N A L C O N V E N T I O N A N D T R A D E S H OW • A P R I L 7 – 1 2 , 2 0 1 1 *Modified Asphalt Shingle **Data on Wrapper / Marketing Literature ADDITIONAL TESTING, WORKMANSHIP ERRORS Test decks were also constructed with often-seen workmanship errors: high nailing (above the required nailing zone), installation with less than four nails, and excessive tab exposure. Samples built with excessive exposure and nails improperly placed failed at wind speeds of 60 mph. ADDITIONAL TESTING, SHINGLE TAB RESEALED Shingles were subjected to various wind speeds until one tab failed but did not crack or bend back from the test deck, according to ASTM D3161. The shingle tabs that became loose at 60 mph were resealed with asphalt roof cement. The repaired roof deck was then retested. The roof with repaired shingles performed up to 110 mph. SHINGLES INSTALLED USING SIX NAILS VERSUS FOUR NAILS Before the advent of self-sealing shingles, the installation of organic shingles using six nails per shingle was a common practice and was required by the Corp of Engineers on military installation projects. The shingle tabs were hand sealed with asphalt roof cement. Manufacturers currently refer to the use of six nails in areas of high wind zones. Shingle samples were tested on test decks using both six and four nails per shingle. The initial tab failures were found to be similar on both sets of decks. Catastrophic failures (seen as fully bent back or broken tabs), however, were minimized when six nails were utilized for attachment. Moreover, the catastrophic failure mode was delayed, requiring a longer exposure to wind before failure. Prior research conducted by Carl Cash indicated that at wind speeds of up to 60 mph, the number of nails used to secure shingles to a deck was not found to be particularly critical. Thomas Smith stated, however, that shingles installed with six nails provided additional wind resistance over shingles installed with four nails.10 DISCUSSIONS Under the current 2009 Edition of the International Building Code, all of the United States falls within a minimum wind speed zone of 90 mph. The IBC also requires that all shingles comply with ASTM D3462. It would appear that Class-A shingles, only tested to 60 mph under ASTM D3161, cannot be code-compliant. The 2012 Edition of the IBC will adopt 2010 American Society of Civil Engineers (ASCE) 7. Significant increases in the basic wind speeds that shingles need to resist will occur, with the minimum standard throughout the United States becoming 115 mph.11 Classifications of shingles at 60 mph, 90 mph, and 110 mph may no longer be adequate for code compliance. There is an assumption that the marketing literature representing a shingle’s physical characteristics and performance attributes is accurate. Prior research has shown that representations of shingle manufacturers claiming shingle compliance with ASTM D3462 are not necessarily accurate. 12 The letter classification, particularly under ASTM D7158 where a wind speed resistance is calculated, may also provide inaccurate information. As Thomas Smith notes, “Research studies13 by others have concluded that shingles [sic] performance during wind events is related to the variability in the physical properties of shingles.” CONCLUSIONS • Successful performance of fiberglass three-tab shingles is clearly dependent upon three interdependent properties: proper installation, adequate activation of the self-sealing adhesive strip, and shingles manufactured actually to meet or exceed the industry-adopted standards of ASTM D3462. • The ratings determined under ASTM D7158, in some cases, list a wind speed of 150 mph. The representation that some shingles can perform at this wind speed may not always be accurate. • Testing shingles in strict accordance with ASTM D3161 procedures may not reflect real-world conditions. Conditioning shingles for 16 hours at a temperature of 57° to 60°C (135° to 140°F), in most cases, did not properly activate the self-sealing adhesive strip. Heating the test decks at higher temperature was required to properly activate the self-sealing adhesive strip. • Shingles with a higher nail pullthrough resistance and a tear resistance of 1,700.0 grams or greater, as required by ASTM D3462, exhibit a higher degree of wind resistance. Shingles that do not meet ASTM D3462 are not code-compliant. • Depending upon the type of damage and the extent of such damage, some repair can be performed on three-tab shingles with an expectation of shingle performance in compliance with code requirements. • Based upon the testing performed, it is clear the first line of defense from high wind speeds is the performance of the self-sealing adhesive strip. Once the self-sealing adhesive strip fails, the physical properties of the shingles come into play. Shingles with enhanced tear strength and fastener pull-through strength tend to have a greater resistance to catastrophic failure. END NOTES 1. Jim D. Koontz, “Fiberglass Shingles,” Western Roofing Magazine, 1990. 2. Thomas L. Smith, Matt Millen, “Influence of Nail Locations on Wind Resistance of Unsealed Asphalt Shingles,” North American Conference on Roofing Technology, Ontario, Canada, Sept. 16, 1999, 98-111. 3. Redrawn from The National Roofing Contractors Roofing and Waterproofing Manual – Fourth Edition, 1,049. 4. Ibid., 1,048. 5. J.A. Peterka, J.A. et al., “Wind Uplift Model for Asphalt Shingles,” Journal of Architectural Engineering, December 1997, 147-155. 5. Donald E. Shaw, “ARMA’S New Approach for Evaluation of Asphalt Shingle Wind Resistance,” International Symposium on Roofing Technology, 1991, 216-218. 7. J. Peterka, G.D. Lamb, J.E. Cermak, “Wind Resistance of Asphaltic Roofing Shingles,” Proceedings of the Seventh International Conference on Wind Engineering, Volume 3, 1987, 21-30. 8. Redrawn from an illustration from Carl G. Cash and Frank W. Kan, “Finite Element Analysis of Racked vs. Traditionally Applied Three-Tab, Seal-Tab, Strip Shingles, and Blister Tests for Asphalt-Glass Felt Shingles,” ASTM STP 1349, Roofing Research and Standards Development, 4th Vol., May 1999, 123–131. 2 6 T H RC I I N T E R N A T I O N A L C O N V E N T I O N A N D T R A D E S H OW • A P R I L 7 – 1 2 , 2 0 1 1 K O O N T Z • 3 5 9. William C. Cullen, “Research and Performance Experience of Asphalt Shingles,” 10th Conference on Roofing Technology, National Institute of Science and Technology, National Roofing Contractors Association, April 1993, 6-12. 10. Thomas L. Smith, “Improving Wind Performance of Asphalt Shingles: Lessons From Hurricane Andrew,” 11th Conference on Roofing Technology, Gaithersburg, Maryland, September 21 – 22, 1995. 11. Thomas L. Smith, “Mapping the 2010 Wind Changes,” Professional Roofing, NRCA, August 2010, 35-39. 12. Jim D. Koontz, “Performance Attributes of Fiberglass Shingles,” Interface, July 2007. 13. Thomas L. Smith and Matt Millen. 3 6 • K O O N T Z 2 6 T H RC I I N T E R N A T I O N A L C O N V E N T I O N A N D T R A D E S H OW • A P R I L 7 – 1 2 , 2 0 1 1
