« PreviousContinue »
point, the condensation on it is liquid water, and air can escape. This forms a layer of clear, high-density ice.
Sooner or later, the particle encounters another strong updraft, starts back up and freezes, subcools and goes through its tumbling cycle over and over again. Thus, the hailstone is found to consist of alternate layers of milky (low density) ice and clear (high density) ice. When the hailstone encounters no updraft sufficient to lift it, it falls to earth, usually at a velocity approximating the free-fall terminal velocity .
The second type of storm occurs on the eastern slopes of the Rocky Mountains; thus, it is called an orographic storm. A front of warm, moist air hits the base of the mountains, expands upward until the nucleation, freezing and tumbling processes occur and then the hailstones drop out as in the frontal storm. This type of storm tends to drop its hailstones at about 6000 ft.
Figure 1 is a map of the central United States showing the distribution of storms during the years 1960–1966 in which at least $5000 worth of building damage was done by hail in each storm. The orographic storms form an imperfect line at the left of the figure; the frontal storms account for the rest of the points. Only infrequently do building-damaging storms occur outside of this area.
Hailstorms occur all over the world in open regions where rapidly moving air masses can develop. However, only meteorological reports on storms and studies on the physics of hail formation can be found in the literature. Oc
casionally reports appear in the trade literature [7, 8] on hail damage to buildings, but only one paper has appeared in which a serious effort has been made to evaluate the effects objectively. In this paper , J.A.P. Laurie reported that he used 21/2-in (6.4 cm) artificial hailstones, made by cutting cylindrical cores from blocks of ice, cutting them to heights equal to their diameter and molding them to roughly spherical shape. He fired these missiles at various velocities at building materials with a grenade launcher and determined the threshold energy of damage. The velocities were controlled by the size of the charge in the blank cartridges used in the launcher.
Because of the difficulties in controlling the velocities of the hailstones, an air-operated piston was developed and used as the launcher in the latter part of Laurie's study.
Laurie's paper, being the only one in its field, was the base from which this work was defactory, (2) the use of ice spheres was extremely desirable, if not absolutely necessary, (3) hail usually struck at its approximate freefall terminal velocity (corroborated by others), and (4) a criterion for failure was damage that would permit the penetration of liquid water to an appreciable extent. However, it was decided to use a less complicated launcher, use "hailstones" of various sizes, cast the "hailstones" to approximate spheres more closely and explore areas of different vulnerability on various roofing systems. The work was primarily directed at bituminous roofing materials, but a sampling of other roofings was made.
2. Apparatus 2.1. Test Apparatus
masking tape, also 2.0 ft (61 cm) apart. The The apparatus consisted of a compressed air
tapes are kept under tension such that any gun, for launching the hailstones, a timer, for
impact on them will close the microswitches and determining their velocity, and a target area,
actuate the triggering mechanism to start and for positioning the specimen to be tested. The
stop the counter (5). physical layout of the apparatus is shown in
The compressed air gun (4) is a commercially figure 2.
available device manufactured by Diamond The apparatus consisted of a specimen (tar
King, Inc.2 (El Segundo, Calif.). It is their get) area (1), timing range (2), gas gun (3),
Mark 14 model, with a 314-in (8.3 cm) inside gas cylinder (4), timer (5), hailstone carrier
diameter barrel and a maximum muzzle ve(6), hailstone molds (7), and a triggering
locity of 300 ft/s (9144 cm/s). The counter is mechanism (8). The roofing specimen to be
a Hewlett-Packard Model No. 523B microsecond tested is mounted on a roof deck, just as in
counter, with both starting and stopping gates service, and clamped in place against the back
and a direct readout. stop in position (1). The timing range consists
2.2. Hailstone Carriers of a metal frame of 34-in (1.9 cm) angle iron on which are mounted two microswitches 2.0 ft The hail carriers were made from 3-in (7.6 (61 cm) apart. The actuating levers on the cm) diameter foamed polyethylene cylinders microswitches contain metal hooks, which are (Ethafoam-Dow Chemical Co., Midland, used to hold one end of 1-in (2.5 cm) paper Mich.). This material was obtained as cylinders computer tapes, the other ends of which are 9 ft (274 cm) long, sliced into short cylinders fastened to the top members of the frames with 6-in (15.2 cm) long and split in half longitudi
The bail resistance of building materials was determined by shooting progressively larger hallotones at different parts of these materials until failure occurred.
6. Timer. 1. Test specimen.
6. Hallstone carriersone in gun and one in open to 2. Timing section.
show cavity. 8. Compressed gas gun.
7. Hailstone mold. 4. Gas cylinder.
8. Triggering mechanism. FIGURE 2. Hail resistance apparatus.
nally (Item 6 fig. 2). Each hemicylinder was truncated at one end at 45 deg to its long axis from the central cut to its outer wall and milled with one of a series of sizes of hemispheres centered 214 in (5.7 cm) from its other end. Thus, when the two hemicylinders were reassembled, they formed carriers for the several sizes of hailstones and permitted one size barrel to be used for all of the hailstones. Carriers with recesses for 112 (3.8 cm), 2 (5.1 cm), 212 (6.4 cm) and 234 in (7.0 cm) hailstones were made. The 114-in (3.2 cm) hailstones were carried in the 11-in (3.8 cm) carrier and the 134-in (4.5 cm) hailstones, in the 2-in (5.1 cm) carrier.
tainer and sliced through with a razor blade at a great circle of the float. The float and rod were removed and the cut interface covered with a thin layer of silicone grease.
The hailstones were cast in these molds in two stages, in order to permit expansion of water during freezing to occur without shattering the hailstones. Water was poured into the mold through the opening (called the gate) left by the removal of the suspending rod until the cavity (left by the float) was about one-half full and frozen in the freezing compartment of a conventional refrigerator. Four hours later water was added to fill the mold just to the bottom of the gate and the mold was returned to the freezer. Only by this twostage process was it possible to freeze ice
2.4. Specimen Construction The shingle specimens were applied with four staples (per strip) to 1 ft 6 in x 3 ft 0 in (46 x spheres without shattering. While the structure of these synthetic hailstones is different from that of naturally formed hailstones, it was felt that the differences in structure did not affect their performance.
References to specific articles in the description of apparatus used in these experiments are for the purpose of definition of the experimental details, and should not be construed as preferential endorsements of these articles.
2.3. Hailstone Molds The hailstones were cast in molds made from a silicone casting resin (RTV-60—General Electric Co.). The models for the hailstones were plastic fishing floats, which are produced in increments of 14-in (0.6 cm) diameter from 1-in (2.5 cm) to 3-in (7.6 cm). Each float was suspended on the end of a rod, which fit the indentation in the float, in the center of a cylindrical polyethylene container of suitable size. The casting resin was deaerated, poured into the mold and cured. The following day the casting was removed from the polyethylene con
1 The ice spheres were stored in a chest-type freezer at about 10°F (-12°C) until ready for use. 91 cm) decks, representative of those used in construction (3/8 in (1 cm) and 1/2 in (1.3 cm) plywood, 1 in x 6 in (nominal) T & G boards). The decks were supported on 2–2 in x 4 in (nominal) “rafters,” to which they were fastened 6 in (15.2 cm) from each of the short sides by 8d common nails. Thus, each deck represented a 1 ft 6 in x 3 ft 0 in (46 x 91 cm) section out of a conventional roof.
Wood, slate, asbestos cement, tile, and sheet metal roofing were applied as directed by their suppliers to decks supported on 2 in x 4 in rafters, 2 ft (61 cm) on centers.
The built-up roofing specimens, 1 ft (30.5 cm) square, were solidly mopped to 1/2 in (1.3 cm) plywood or 1 in (2.5 cm) asbestos cement board (to simulate a concrete deck) or to various types of insulation mopped solidly to these decks. Also, where metal decking was used, the insulation was mopped solidly to the decking.
3. Procedures 3.1. Shooting Hailstones at Roofing
The indentation on the specimen was meas
ured and the condition of the specimen noted The specimen on its deck was held against
after each firing. A minimum of two hits in the backstop in figure 2 with large C clamps.
each area of vulnerability was observed and The 1-in (2.5 cm) paper computer tapes were
average values of damage used. Granule losses, hooked to the microswitches and fastened to
coating and felt fractures and deck damage the top of the timing frame with masking tape.
were recorded. The average velocities and enerThey were held under tension, just insufficient
gies of the hailstones in the 2 ft (61 cm) of to close the switches. A hailstone of the desired
travel immediately in front of the test specimen size was taken from the freezer, cleaned of any
also were calculated and recorded. burrs or projecting pieces of ice (from the gate in the mold), weighed, and placed in its carrier,
3.2. Evaluating Failure which was slid into the barrel of the gun as far as possible. Air, or nitrogen, was permitted to enter the gun until the desired pressure was
Damage to roofing by hail falls into two genreached. The valves between the gun and the
eral categories: (1) Severe damage, which leads tank were closed (to protect the pressure regu
to penetration of the structure by the elements lator), the pressure gage was removed from and (2) Superficial damage, which affects apthe gun, and the gun was fired by opening the
pearance but does not materially interfere solenoid valve, which relieved the pressure
with the performance of the roofing. While the behind the floating cylinder in the gun and
latter is distracting and leads to insurance permitted the remainder of the pressurized
claims, the former is the type of damage that gas to escape into the barrel and expel the
should be of most concern, because the possible hailstone carrier.
loss can exceed the replacement cost of the The carrier was propelled out of the gun,
roofing many fold. Thus, while the dents will where the air resistance opened the two halves
be reported, only the fractures of the coating, and permitted the hailstone to travel alone felt or other shingle material will be called toward the target. As the hailstone hit the first failure in this report. For each material and tape it started the counter, and as it hit the roofing system, the thresholds of failure, or the second tape, it stopped the counter. Then it hit smallest hail size producing these failures, are the specimen.
4. Results Although hailstones vary in size, shape, den- and reported graphically by Laurie ; they sity, and velocity, those that do damage tots are tabulated below as taken from these graphs buildings tend to fall within the narrow limits for the hailstone sizes used: of ice spheres falling at about their free-fall terminal velocity .
TABLE 1. Terminal velocities and energies of hailstonesa The density of large hailstones has been
impact shown to approximate that of solid ice  and seems to range between 0.89 and 0.91
ft lbs (2.5)
(<1.36) g/cm3. Hailstones, while rarely smooth spheres, 144 (3.2)
(5.42) can be treated aerodynamically as smooth
(10.85) 12 (4.5)
(18.96 ) spheres and conclusions reached are close to
(71.9) observed results . The terminal velocities 24 (7.0)
(162.7) and energies of ice spheres have been calculated
• Read from graphs in reference (9).
50 56 61 66 72
73 82 90 97 105 117 124 130
4 8 14 22 53 81 120
All of the results reported are based on hailstones of a given size traveling at velocities within + 10 percent of the terminal velocities reported in table 1 for hailstones of that size. The results are reported under the types of roofing studied.
4.1. Asphalt Shingles When applied according to the recommendations of their manufacturers, Type 235 squaretab shingles have three regions of different vulnerability: (1) The tab edges, (2) The surface over the unsupported areas between the top of one strip and the “line” where the strip above it contacts the deck or underlayment, and (3) The triple coverage area solidly supported from the deck up .
The resistances of these areas to hail damage are different; therefore, results are reported for each area. The results for the Type 235 square-tab shingles are shown in table 2.
These specimens were also exposed to 11/4-in dentations were made in the shingles by the (3.2 cm) hailstones. Only small, superficial in114-in (3.2 cm) hailstones. The larger size hailstones produced progressively larger dents. In general, the smaller hailstones produced circular indentations approximating one half their diameter and the larger hailstones, those above the felt-damage threshold, produced dents greater in diameter than one-half the hailstone diameter. Hailstones 234 in (7.0 cm) in diameter produced damage to the decks on which the shingles were mounted.
TABLE 2. Hail resistance of Type 235 square tab shingles exposed 5 in (12.7 cm)
Shingles on 3/8-in (1 cm) and 1/2-in (1.3 cm) plywood performed equally well; those on 1 x 6-in T & G roof boards were more resistant to hail damage than those on plywood.
The shingles without an underlayment consistently had a higher threshold of hail damage than did those with the conventional 15 lb saturated felt underlayment on all three decks. Apparently, the soft layer of felt makes the shingle slightly more vulnerable. The improved performance usually involved only 1/4-in (0.6 cm) larger hailstones, but this represented resistance to 6.3 or 9.5 more foot pounds (8.5 or 12.8 joules) of kinetic energy. From these results, it would seem that the solidly supported roofings performed better than those with some soft underlying layer in their construction. This observation is consistent with the fact that shingle materials are stronger in compression than in tension and the best performance can be expected when the impact forces can be kept as pure compression forces. Any soft layer within the system permits the back of the layer above it to be in tension and fail more easily.
As shingles age during exposure they tend to undergo a number of physical changes, which may affect their resistance to hail. A number of shingles that had been exposed on 1/2-in (1.3 cm) plywood to the weather in Washington, D. C. for 912 years became available and were
tested. These shingles had been exposed at a 4-in pitch (10 cm in 30 cm) facing due south. Three different Type 210 shingles showed failures (felt cracking) on all three areas of different vulnerability with 114-in (3.2 cm) hailstones. One Type 255 and one Type 290 shingle experienced spalling of the coating with 114-in (3.2 cm) hailstones, but felt damage did not occur until 112-in (3.8 cm) hailstones were used. Two Type 250 shingles showed felt damage in all three areas of vulnerability with 11/4in (3.2 cm) hailstones; however, one Type 250 and one Type 275 shingle showed no damage below 134-in (4.5 cm) hailstones on the tab centers, but both developed felt damage in the other two areas with 11/4-in (3.2 cm) hailstones. No direct comparison can be made between these aged shingles and unexposed ones because of changes in design and production. However, the aged shingles tended to be less re sistant to hail damage than the new ones.
A number of heavy weight and premium shingles were also investigated. Some of these resisted hail no better than the regular Type 235 square-tab shingles. However, a few performed significantly better, as discussed below.
A Class B shingle based on a glass fiber mat, instead of the conventional organic felt, did not show felt-type failure on its tab edges and unsupported areas with hailstones smaller than 2 in (5.1 cm). It failed with 21/2-in (6.4 cm)