Got Hail? Proving the performance of EPDM roof assemblies

by Katie Daniel | March 8, 2016 3:56 pm

By Jim D. Koontz, PE, RRC, and Thomas W. Hutchinson, CSI, AIA, FRCI, RRC, RRP
Damage to roof assemblies from around the world from hail or other forms of physical damage (e.g. foot traffic and workmanship) results in millions of dollars of economic loss annually. Owners of properties that are largely self-insured are beginning to realize the importance of installing hail-resistant roofing systems.

Based on empirical evidence, ethylene propylene diene monomer (EPDM) roofing assemblies seem to fare very well in hailstorm events. The EPDM Roofing Association (ERA), however, desired scientific validation; it decided to embark on a hail testing program. The technical committee decided that in addition to new EPDM roof membrane material, the real question in the design, insurance, and contractor communities was over how well aged, in-situ EPDM roof covers performed.

It was determined the test sample pool would include new EPDM membrane material, new membranes that would be heat-aged, and existing in-situ material procured from existing roofs with five to 20 years of actual exposure. The major EPDM manufacturers each provided:

• 1.2 x 1.2-m (4 x 4-ft) new 1.52-mm (60-mil) material EPDM samples;
• new, but heat-aged, 1.52-mm material; and
• 1.52-mm samples from roof covers that had been exposed between five and 20 years, procured from the field.

Prior to sending the EPDM samples for testing, the material was fully adhered to various 1.2 x 1.2-m substrates: mechanically fastened polyisocyanurate (polyiso) insulation, mechanically fastened wood fibre board, and 13-mm (½-in.) plywood. Between 20 and 35 samples of each roof cover category were sent for testing.

Field experience from the examination of thousands of roofs, and through Roofing Industry Committee on Weather Issues (RICOWI) hail impact investigations (For example, the “2011 Dallas−Fort Worth Hail Storm Report” can be downloaded from www.ricowi.com[1]), has clearly shown hail damage to a roofing system can be the result of several factors:

• hailstone diameter;
• roofing system type;
• roof cover age;
• substrate beneath the primary roof assembly; and
• surface temperature at the point of impact.

To evaluate a roofing system’s resistance to hail, these reference points have to be considered as part of a research project.

NBS impact research
In the early 1960s, the National Bureau of Standards (NBS) in Washington, D.C., conducted research by impacting roof systems with ice spheres. The group’s Sydney H. Greenfield performed this initial research and generated technical article NBS 23, Hail Resistance of Roofing Products. Referring to previous research (This came from “Hail and Its Effects on Buildings,” by J.A.P. Laurie, Council for Scientific and Industrial Research (CSIR) Research Report No. 176, Bull. 21, 1-12, published by the South African National Building Research Institute in 1960), he initially determined the freefall or terminal velocity of hail.

The technical data shown in Figure 1 indicates the free-fall velocity of the hail increases with hailstones of larger diameters. A key factor is the amount of ‘impact energy’ imparted to a target or roof surface. Simply stated:

Impact Energy = Kinetic Energy = ½ Mass * Velocity²

The mass of a hailstone obviously depends on the volume of the ice sphere and the density of the ice. The density of hailstones is typically valued at .91.

Volume of a sphere = 1.33 * Π * Radius³

A substantial difference in impact energy occurs with only slight changes in diameter. As shown in Figure 2, increasing hailstone size from 25 to 51 mm (1 to 2 in.) only represents a 100 per cent change in diameter, but the impact energy increases by 1580 per cent.

Industry impact research
Historically, the hail resistance of roofing products has been tested by dropping steel balls or darts onto the roofing product. The procedures used to impact roofing products have varied between Canada, the United States, and European organizations.

In Canada, groups have utilized the now-withdrawn impact procedure of Canadian General Standards Board (CGSB) 37-GP-52M, Roofing and Waterproofing Membrane, Sheet Applied, Elastomeric. As is the case for UL and FM Global south of the border, this procedure employed steel darts to impact targets, typically at room temperature.

Other organizations have developed impact tests that use steel darts, such as ASTM D3746, Standard Test Method for Impact Resistance of Bituminous Roofing Systems. Within the last few years, greater consideration has been given to impacting targets with ice spheres. Prior research by this article’s co-author has also reviewed the issue of ice spheres versus steel darts. (For more, see Koontz’s “Simulated Hail Damage and Impact Resistance Test Procedures for Roof Coverings and Membranes,” published by RICOWI in 2000). (The use of ice spheres, obviously, comes closer to replicating what occurs during a real hailstorm.)

Hail launcher
A key factor in performing the test is to have reproducible impact energies with each shot of ‘hailstone.’ The hail gun or launcher propels ice spheres by utilizing the quick release of compressed air from a tank to a barrel (Figure 3). To achieve reproducibility, several factors have to be taken into consideration. Consistent ‘air pressure’ is required for each shot. This necessitates controlling the air pressure to 70 Pa (0.01 psi).

Moulds for ice spheres are fabricated using precise-diameter steel spheres (Figure 4). Each ice sphere of a given diameter is then weighed to 0.01 g prior to each shot. Laboratory-grade barrels or tubes with precise internal diameters are also necessary to develop consistent impact energies. Basically, the charge (i.e. air pressure), the quick-release valve, and the bullets (i.e. ice spheres) require precise fabrication in order to achieve reproducible impact energies.

The ice spheres are initially weighed and then placed in the barrel, similar to a lead shot for a muzzle loader. As the ice sphere is pneumatically launched toward the target, the velocity is measured with a ballistics timer. The kinetic (or ‘impact’) energy is then calculated for each hail shot. The minimum kinetic energies listed by NBS are maintained within a tolerance of minus zero plus 10 per cent.

The specified slider does not exist.

The specified slider does not exist.

EPDM targets and impact procedures
Manufacturers provided a total of 81 test targets constructed with 1.52-mm (60-mil) non-reinforced EPDM for impact testing. The new, heat-aged, and field-aged targets are listed in Figure 5. The field-aged and exposed EPDM samples were collected from six states and ranged in age from five to 20 years. It is important to note some of these field-aged samples—listed in Figure 6—represent more heat exposure than would typically be found in the Canadian climates, translating to an added safety factor for buildings in this country. The U.S. samples taken from the Great Plains—a region experiencing climatic weather conditions similar to those of the plains of Alberta, Saskatchewan, and Manitoba—provide a good barometer for the performance of roofs in these provinces.

The artificially heat aged samples were prepared at Cascade Technical Services of Hillsboro, Oregon. The samples were heat-aged for 1440 hours at a temperature of 116 C (240 F). The 1.2 x 1.2-m (4 x 4-ft) EPDM ‘targets’ were installed over a variety of substrates that included polyiso and wood fibre insulation, plywood, and oriented strandboard (OSB). Fully adhered EPDM was utilized in the target construction. Figure 7 indicates the material age, substrate, and number of samples of each prepared.

Each target with substrate was vertically mounted. Hailstones measuring 38, 51, 63, and 76 mm (1 ½, 2, 2 ½, and 3 in.) impacted the targets at a 90-degree angle at velocities listed by NBS. To replicate severe weather conditions, such as cold rain during a hailstorm, the test targets were sprayed with water at 4 C (39 F). Prior research and experience has shown roof assemblies exhibit different levels of impact resistance depending upon surface temperature.

The various targets were impacted both in the ‘field area’ and also directly over fasteners and plates utilized to secure the substrate below the EPDM. Failure was defined as a visible split or cut in the surface of
the EPDM.

Impact results
Of the 25 ‘new’ EPDM targets tested, 24 were not damaged by the 76-mm (3-in.) hail balls. None of the 20 ‘heat-aged’ targets failed when impacted with 76-mm hail balls.

The ‘field-aged’ EPDM target samples included 18 over a 51-mm (2-in.) thick polyiso insulation substrate, and 18 over a 13-mm (½-in.) thick OSB substrate, supported by 38-mm (1 ½-in.) thick polyiso roof insulation. Fourteen of the EPDM targets adhered directly over the polyiso did not fail when impacted with 76-mm hail balls. One sample failed with a 76-mm hail ball, a second sample failed with a 63-mm (2 ½-in.) hail ball, and the two other samples failed with a 51-mm diameter hail ball. None of the 18 EPDM ‘field-aged’ targets over OSB were damaged by 76-mm diameter hail balls (Figure 8).

Hail and roofs
Hail forms in the core of a thunderstorm. Water vapour in warm, rapidly-rising air masses (convection currents) condenses into water at higher, cooler altitudes to produce heavy rain showers. If it is cold enough, ice crystals can form around minute particles, such as dust whipped up from the ground, and increase in size as more water freezes onto their surfaces. When the ice pellets are too heavy for the ascending air currents to lift, they fall as hail. They may become larger, heavier and more damaging if they collect more water on the way down.

Hailstones have a minimum diameter of 5 mm (0.2 in.)—smaller and they are defined as snow or ice pellets. Hail can grow larger than 100 mm (4 in.), reaching the size of grapefruit. It can hit the ground at 130 km/h (80 mph), causing severe damage to crops, houses, and vehicles, along with injuring people and animals.

As shown in Figure 9, hail occurs right across Canada, but more frequently in Alberta, the southern Prairies and in southern Ontario. Roof assemblies should be capable of resisting impact from reasonably expected hail storms for a given area. Just as roofs are required to perform in various meteorological events (e.g. wind, snow, and rain), a roof should be able to withstand some degree of hail impact over its expected service life.

This article’s co-author has examined hundreds of EPDM roofs impacted by hail. Two noteworthy investigations include a telephone building in Fort Worth, Texas, that was impacted by softball-sized hail in 1995. The non-reinforced EPDM over polyiso did not fail. A second investigation was at the University of Nebraska in Kearney. All campus buildings covered with non-reinforced EPDM survived softball-sized (i.e. 96-mm [3.8-in.]) hail. The manufacturer of the roof was notified of the performance of the aged EPDM assembly. The gravel built-up roofing (BUR), metal, slate, clay tile, and modified bitumen (mod-bit) roof covers on the 65 other university buildings all failed.

During the examination of hundreds of roofs, direct impacts over fasteners and plates used to secure underlayment have been extremely rare. Damage observed of that kind has not constituted a failure of the entire roof, and has been repairable. The increasing use of adhesives to fasten insulation and coverboards is eliminating the already unlikely chance of damage caused by hail-ball impact with mechanical fastener plates.

Conclusion
The new, heat-aged, and aged non-reinforced EPDM tested within this study provided excellent resistance to large hail. Of the 81 targets installed over polyiso, wood fibre, plywood, and OSB, 76 did not fail when impacted with hail ice balls up to 76 mm (3 in.) in diameter.

The overall results of this research clearly indicate non-reinforced EPDM roof assemblies offer a high degree of hail resistance over a variety of substrates, validating empirical experiences. The impact resistance of both the field-aged and heat-aged membrane also clearly demonstrates EPDM retains the bulk of its impact resistance as it ages.

Owners of critical facilities (e.g. hospitals, schools, data centres, airports, and sensitive government buildings), who demand durability and long-term service lives of their roofs, have come to realize the importance of installing a hail-resistant assembly. The use of non-reinforced EPDM can provide an additional level of long-term protection.(Koontz’s “A Comparative Study of Dynamic Impact and Static Loading of One-Ply Roofing Assemblies” is a reprint from ASTM Special Technical Publication (STP) 959 1988 (1988): 24).

In heating-dominated northern climates like Canada, roof system designers should consider using black EPDM in fully adhered assemblies where the top layer of cover board is set in urethane adhesive—beads or full-coverage sprayed polyurethane foam (SPF). From a design standpoint, this assembly has many built-in safeguards, such as the durability of non-reinforced EPDM without any hard insulation fastening plates or fastener heads directly under the membrane. A ballasted EPDM roofing assembly utilizing black EPDM can be a prudent choice for hail resistance.

Further, roof cover durability in hail events is linked to membrane thickness. Designers should consider specifying greater roof membrane thickness.

From empirical observation by the authors and by testing, several key characteristics of roof systems designed for hail resistance can be summarized:

1. Utilize non-reinforced roof covers with greater thickness—2.28 mm (90 mil) over 1.52 mm (60 mil).
2. Provide high-density support for the roof cover (i.e. do not allow the membrane to be depressed under hail ball impact).
3. When used, place screw fasteners and stress plates below the cover board, and do not place below roof covers.
4. Ballasted roof systems provide the greatest protection.
5. Specify roof covers with a history (i.e. at least 20 years) of general in-situ performance to ensure performance as the membrane ages and is exposed to ultraviolet (UV) radiation and climatic conditions.

It is important to remember this hail testing was completed in the United States, but the data can be useful to design/construction professionals north of the border as hail is not aware of territorial lines. As damage from these storms can be quite dramatic, it will be important for Canadian testing agencies to consider this in the future.

Jim D. Koontz, PE, RRC, has been involved in the roofing industry since 1960 and began testing roofing materials in 1976. President of Jim D. Koontz & Associates, he has experience as a roofer, estimator, consultant, lecturer, researcher, and expert witness. Koontz has authored numerous articles relating to roofing material/product research, including research on single-ply products and hail/wind impacts. He can be reached at jim@jdkoontz.com[2].

Tom Hutchinson, AIA, FRCI, RRC, is a principal in Hutchinson Design Group. A licensed architect and registered roof consultant, he specializes in roof design, contract document preparation, specifications, inspections, and the determination of moisture penetration and failure of existing roof system. Hutchinson is a former president of RCI Inc. He is also a Certified Energy Professional in Chicago and secretary of the Conseil International du Bâtiment/Réunion Internationale des Laboratoires d’Essais et de recherches sur les Matériaux et les constructions (CIB/RILEM) International Joint Committee on Roof Materials and Systems. Hutchinson can be contacted via e-mail at hutch@hutchinsondesigngroup.com[3].

Source URL: https://www.constructioncanada.net/got-hail-proving-the-performance-of-epdm-roof-assemblies/

---