13

Wind Damage to Buildings and Structures

Billy R. Manning, Southern Building Code Congress, Birmingham, Alabama

INTRODUCTION

This report is based on ground surveys along the South Carolina coast from Edisto Island north to Myrtle Beach, and as far inland as Charlotte, North Carolina An aerial survey was also made over the area contained by Sumter, Walterboro, and Seabrook Island in South Carolina to Yaupon Beach, North Carolina

The ground and air surveys and meteorological reports indicate that Hugo did not generate any tornadoes in the areas. Tornado effects were reported in the area but this was the result of the numerous wind cells (downbursts) in the northern quadrant of the hurricane. The surveys also indicated that the strongest winds and highest storm surge occurred in the Bull Bay area, approximately 25 miles northeast of Charleston.

Wind damage occurred as far south as Edisto Beach and as far north as North Myrtle Beach. The wind damage in the Edisto Beach and North Myrtle Beach areas was minor (primarily cladding failures). However, damage was extensive in the Charleston, Mt. Pleasant, and Bull Bay areas. Most of the damage inland was caused by falling trees; however, cladding failures were found as far inland as Charlotte, North Carolina, which is approximately 180 mi from the Charleston coast.

Buildings and structures were grouped into three categories for the purpose of this report, based on the level of engineering effort that may have been involved in the design.

  1. Nonengineered. These are buildings and structures that receive no specific engineering attention. Examples are most single and duplex residences, small commercial buildings, and small and medium size signs.

  2. Marginally Engineered. These are buildings and structures that receive minimal engineering attention. Examples are one- to three-story motels, apartments, offices, light industrial buildings, and large signs.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 247
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA 13 Wind Damage to Buildings and Structures Billy R. Manning, Southern Building Code Congress, Birmingham, Alabama INTRODUCTION This report is based on ground surveys along the South Carolina coast from Edisto Island north to Myrtle Beach, and as far inland as Charlotte, North Carolina An aerial survey was also made over the area contained by Sumter, Walterboro, and Seabrook Island in South Carolina to Yaupon Beach, North Carolina The ground and air surveys and meteorological reports indicate that Hugo did not generate any tornadoes in the areas. Tornado effects were reported in the area but this was the result of the numerous wind cells (downbursts) in the northern quadrant of the hurricane. The surveys also indicated that the strongest winds and highest storm surge occurred in the Bull Bay area, approximately 25 miles northeast of Charleston. Wind damage occurred as far south as Edisto Beach and as far north as North Myrtle Beach. The wind damage in the Edisto Beach and North Myrtle Beach areas was minor (primarily cladding failures). However, damage was extensive in the Charleston, Mt. Pleasant, and Bull Bay areas. Most of the damage inland was caused by falling trees; however, cladding failures were found as far inland as Charlotte, North Carolina, which is approximately 180 mi from the Charleston coast. Buildings and structures were grouped into three categories for the purpose of this report, based on the level of engineering effort that may have been involved in the design. Nonengineered. These are buildings and structures that receive no specific engineering attention. Examples are most single and duplex residences, small commercial buildings, and small and medium size signs. Marginally Engineered. These are buildings and structures that receive minimal engineering attention. Examples are one- to three-story motels, apartments, offices, light industrial buildings, and large signs.

OCR for page 247
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA Fully Engineered. These are buildings and structures that are individually designed by professional engineers and architects. Examples are highrise buildings, hospitals, and public buildings. In general, nonengineered buildings and structures, particularly homes and signage, received the most extensive damage; the fully engineered buildings and structures received the least. The marginally engineered buildings suffered moderate to severe damage. The height of the storm surge from Mt. Pleasant north made it difficult to determine if wind or water or both caused the initial damage. Not knowing which occurred first could cause incorrect assumptions. An example was a fish processing plant near the Ben Sawyer Memorial Bridge in Mt. Pleasant. The roof of this building was sheared off by a wind blowing toward the southwest. The storm surge carried the roof debris in the opposite direction and deposited it approximately 300 yards from the processing plant site. One could have assumed the wind had been blowing from the opposite direction, and that the plant had been hit by a tornado. In the Bull Bay area, the storm surge caused the initial damage to homes with low elevations in a residential area. Since some of these homes had already received considerable water damage by the time maximum wind velocities were realized, the wind compounded the damage, and could have led an investigator to overestimate the wind velocity. Several houses in the area that had been properly elevated received only minor wind damage, allowing the investigator to more correctly evaluate the actual wind velocity. One of the most significant aspects of the storm was the extent of tree damage. There were areas where wind caused significant damage to trees, while damage to buildings was nonexistent. The primary reason for this is that the pine is the predominant tree in the area, and vegetation for mature pine trees is located 30-40 ft above ground, where the wind speeds are higher than those at or near ground level. The falling trees caused extensive damage to residences as far inland as Charlotte, North Carolina Near the coast, the trees served as a windshield, preventing the wind from damaging many of the residences, but in many of these areas, residences were damaged by broken trees or branches. NONENGINEERED BUILDINGS AND STRUCTURES In residential areas where the most intense winds occurred and the extensive damage was observed, one could find houses that received little or no damage standing beside the remains of a house totally destroyed by the storm. Some have concluded that this type of damage was caused by tornadoes. However, when a review was made of all houses, it became apparent that the houses that were standing had been constructed in accordance with the building code and federal government flood-plain requirements, while the remains of those destroyed indicated a lack of compliance with codes and flood-plain requirements.

OCR for page 247
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA Failure of roof coverings was the most widespread damage observed. Metal roof coverings that were not adequately attached, as well as corner and eave regions of asphalt shingle roofs, were frequently damaged. Loss of weather protection caused a great deal of water damage to building interiors and contents. On residential buildings damaged roofs were the most common structural failure observed to be caused by high winds. Roofs were frequently blown off because of the lack of a proper connection between the roof and the exterior walls. In some cases, the rafters were attached only by toenails to the top plate, while hurricane clips were used in other buildings only to attach the rafter to the top plate. When the roof was blown off, the walls would lose the support provided by the roof system, and later, lesser winds would collapse the exterior walls. Another frequent failure was the displacement by wind or water of a residence from its foundation. Adequate connections between the superstructure and the foundation were seldom found. As in previous hurricane surveys, inadequate pier foundations or a complete lack of pier reinforcement was a common problem. A significant number of mechanical components—air-cooled condensing units, outdoor heat pump units, and underfloor air conditioning ducts—received major damage or were totally destroyed by the storm. Damage to these elements occurred because of a lack of equipment platforms above the base flood elevation or because the ductwork was installed below the base flood elevation. The survey team identified one unique structural failure to a home in the Bull Bay area. It resulted from a horizontal wind force causing the residence to rotate the floor joists on the supporting beams, dropping the superstructure dropping approximately 11 1/2 inches until it came to rest on flat floor joists. Current building code requirements even for nonhurricane areas require solid blocking or bridging at the ends of all joists. If the builder had complied with this code section, much of the damage to this residence would not have occurred. The houses most frequently destroyed or having major damage were older homes; however, in a number of cases homes that had been constructed in the past 5 years were also destroyed or had major damage. Signage and Canopies There was extensive sign and canopy damage throughout the areas surveyed (Figure 13-1). In areas where the winds were at or near hurricane velocities, almost all signs and canopies incurred major damage. If the plastic or lightweight metal panels did not fail, then the structural system or foundation failed. The plastic or metal or other cladding material from each sign or canopy failure became airborne and caused damage to utility lines and buildings. When the structural frame or foundation failed, the falling canopy or sign in several cases caused major damage to an adjacent building.

OCR for page 247
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA FIGURE 13-1 Typical damage associated with signs. MARGINALLY ENGINEERED BUILDINGS AND STRUCTURES A number of wind failures were noted in lowrise apartments, motels, and businesses. The predominant failures observed were roof failures (Figure 13-2). Inspection of the damaged buildings indicated that in most cases the roofs had been tied to the walls with hurricane clips that were inadequately sized to support the design wind load. Other failures noted were end-wall failures in the gable area because of inadequately sized members, inadequate horizontal support for the vertical members, poor framing techniques, or a combination of the three. Typically, many lowrise buildings lose roof coverings in hurricanes. Typical roof coverings on these buildings were either galvanized metal, single-ply membranes, or built-up roofing (Figure 13-3). The failures of both types of roofing resulted from inadequate ballast or attachments of the roofing to resist the design wind loads. In several cases, the builder used components from an engineered metal building system along with nonengineered components such as unreinforced concrete masonry, or made field changes to a preengineered building. These combinations frequently resulted in a number of failures in various components of the building. FULLY ENGINEERED STRUCTURES No damage was observed to main structural systems of engineered buildings. There were, however, a number of roofing, wall panel, and cladding failures. The wall panel failures occurred because winds were allowed to penetrate the building and cause changes in the internal pressures, which led to additional failures of

OCR for page 247
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA FIGURE 13-2 Typical damage to roofs. FIGURE 13-3 Typical damage to metal roof coverings.

OCR for page 247
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA components such as roofs and interior partitions. These failures significantly increased the damage to the building. Wall panel failures were observed in engineered buildings as far inland as Charlotte, North Carolina Inspection of the wall panel attachment to the buildings' structural systems indicated that attachment systems were inadequate for the design wind loads. Where the wall claddings were made of materials such as brick veneer or stucco panel, failures did not tend to portend additional damage. Cladding failures were noted as far north as North Myrtle Beach, where the recorded winds were well below design wind velocities. Roofing failures, on the other hand, resulted in extensive additional damage, as water was able to penetrate the interior building, causing extensive damage to finishes and contents. These roofing failures occurred in both builtup and membrane roofing systems. As with the lowrise buildings, the failures were due to inadequate ballast or roofing attachment for the design wind loads. A significant number of mechanical equipment component failures were noted (Figure 13-4). Most of the failures were located on the roofs and were associated with wind failures. The equipment included rooftop air conditioners, satellite dishes, and other communication components. In a number of cases, the equipment was blown over, causing damage to portions of the roofing system. This damage allowed water to penetrate the building, causing additional damage to the interior. The primary reason for the failure was that the equipment supports were not designed and constructed to resist the design wind forces acting on the equipment. FIGURE 13-4 Damaged satellite dish is typical of damage associated with roof-mounted mechanical equipment.

OCR for page 247
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA Other Structures Water Tank In the historical area of Charleston, the top of a 60-ft water tank was blown approximately 200 yards from the tank. The top was constructed of approximately 5/8-inch steel welded to the sides of the tank. Because of inadequate maintenance, many areas around the welds were corroded completely through the metal. Light Standards Many light standard failed in the face of the high winds. In most cases, the foundation of the light standard was inadequate for the design wind loads. Bridge The Ben Sawyer Memorial Bridge, the only bridge connecting Sullivans Island and the Isle of Palms to the mainland, was blown out of position and tilted at a 30-degree angle. Several possible reasons were offered for the failure. The most plausible reason is that the bridge was left in an open position to allow boat traffic to pass on the Intracoastal Waterway, and the braking system was not properly activated. Crane The southernmost of several track-mounted cranes in an industrial complex adjacent to the Dockside Condominium totally collapsed ( Figure 13-5). Since the other cranes appear to have no damage, it is believed the failure was caused by storm surge interacting with the foundation system of the crane. CODES A telephone survey was made of the 19 town, city, and county building departments on the South Carolina coast. This survey indicated that the earliest building code was adopted in 1929, when the city of Charleston developed and began enforcing its own building code. The latest local code adopted was 1985, when the town of Pauleys Island was incorporated and adopted the 1985 edition of the Standard Building Code and the 1983 edition of the One- and Two-Family Dwelling Code. All of the local governments surveyed had adopted the Standard Building

OCR for page 247
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA FIGURE 13-5 Damage to track-mounted cranes in industrial complex. Code, with 16 having adopted the Council of American Building Officials One- and Two-Family Dwelling Code. The 1985 edition of the Standard Building Code is the earliest edition being enforced, although the 1988 edition is the latest. The 1983 edition of the One- and Two-Family Code is the earliest edition being enforced, with the 1986 edition being the latest adopted. No jurisdiction surveyed had adopted the 1989 One- and Two-Family Dwelling Code. The 1985 and 1988 editions of the Standard Building Code have prescriptive provisions for masonry (Chapter 14) and wood frame constructions (Chapter 17). These provisions were intended to be limited to light frame construction structures having light loads. A number of builders and code officials misinterpreted the intent of these chapters and used these prescriptive provisions for one- and two-family dwelling constructions on the South Carolina coast. The 1983 and 1986 edition of the One- and Two-Family Dwelling Code, based on the wind probability map in Appendix A of the code, recommends that buildings less than 30 ft in height located along the South Carolina coast be designed to resist wind pressure of 25 psf. This code requires special design consideration for wood frame walls and related connections only when the wind pressure exceeds 30 psf. This provision resulted in nailed connections at rafter-top plate-studs and at stud-bottom plate joints for a number of residences on the South Carolina coast. The membership of the Southern Building Code Congress considered code changes during 1990 to Chapter 14 (Masonry Construction) and Chapter 17 (Wood Construction) to clearly indicate that these chapters are intended only for areas with design wind velocities not exceeding 80 mph. However, these changes failed to gain approval because the membership did not believe the proponent justified the

OCR for page 247
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA proposed threshold. Additionally, the membership seems to favor the generalized threshold statement currently in the code over more specific guidelines. The 1989 edition of the One- and Two-Family Dwelling Code has been revised to reflect design wind pressures above 30 psf for the entire United States coast. The code now requires walls and related connectors of all wood framed one- and two-family dwellings located in coastal areas be designed. Additionally, the revisions will require greater amounts of reinforcement in masonry walls, and connections to the masonry walls will be required to have higher capacity. CONCLUSIONS The following conclusions were drawn from the ground and air surveys of Hurricane Hugo damage found in meteorological reports with respect to measured wind speeds (Table 13-1), as well as independent ground and aerial surveys: The wind velocities were at or below the design wind velocity requirements of the Standard Building Code. A comparison between hurricanes Alicia and Hugo indicated that the velocity of Hugo was only slightly greater than that of Alicia. The inland extent of the effects of this hurricane was unusual. Near-hurricane-force winds extended inland to Charlotte, North Carolina, causing extensive tree and minor building damage for distances up to 180 mi from the South Carolina coast. Hugo had a much faster forward movement than most storms, and it is believed that this accounts for the higher than normal inland wind speeds. The faster forward speed may have actually limited damage in the coastal regions because of the shorter time of exposure; on the other hand, it did not weaken the storm as much as would have been expected as it went inland. The storm did not generate any tornadoes; however, the high number of wind cells (downbursts) in the northern quadrant caused a number of isolated areas of major tree and minor building damage as far north as Myrtle Beach. Building code compliance varied, from good in most urban areas to non-existent in some rural areas. The intent of construction requirements is not clear—or special construction requirements are not specified—in codes being enforced by local governments on the South Carolina coast. The sign codes and ordinances did not contain wind-load provisions, or enforcement of these documents was inadequate.

OCR for page 247
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA TABLE 13-1 Measured Wind Speeds for Hurricane Hugo Location Sustaineda (mph) Fastest-Mileb (mph) Beaufort, S.C. 54 53 Folly Beach, S.C. 85 88 Charleston, S.C. 88 91 Mt. Pleasant, S.C.c 82 84 Myrtle Beach, S.C. 52 50 Charlotte, N.C. 69 73 Notes: aBased on sustained wind speeds (1-min averaging time) reported from anemometer readings taken by the NWS. bBased on approximate conversion methods that adjust sustained wind speeds to fastest-mile wind speeds at 10 m (33 ft) above the ground. cWind speeds from Mt. Pleasant may not be representative ov actual wind speeds, since the anemometer was well shielded. RECOMMENDATIONS General Local governments enforcing the Standard Building Code should interpret the provisions of Chapter 14 (Masonry Construction) and Chapter 17 (Wood Construction) as not being applicable to coastal construction (e.g., wind velocities in excess of 80 mph fastest-mile wind). Local governments enforcing the One- and Two-Family Dwelling Code should adopt the 1989 edition immediately. All local governments should adopt and enforce the latest edition of a model building code and federal flood-plain-management regulations. All local governments should adopt or revise and enforce a sign code that has wind-load requirements that comply with the building code. Develop and implement a system to evaluate and rate the effectiveness of local code enforcement. Base both wind and flood insurance premiums on this local government rating.

OCR for page 247
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA Nonengineered Buildings and Structures Develop a prescriptive document that clearly details construction techniques for compliance with wind-load provisions of the building code. All local governments should adopt the document as a mechanism for complying with the code. Develop and implement training and certification programs on hurricane-resistant construction for local building department inspectors. Be more concerned with the wind resistance of roof coverings, particularly asphalt shingles and metal roofs. A viable testing standard and evaluation procedure needs to be developed for accessing the wind resistance of asphalt shingle roof coverings. Marginally Engineered Buildings and Structures Develop and implement a training and certification program on hurricane-resistant construction for contractors and local building department plan reviewers. All building departments should require the designer to submit calculations for sizing all connectors. Pay greater attention to systems engineering details in designing buildings that use “mixed” components. Fully Engineered Building and Structures Inform designers (architects and engineers) through the states' registration systems of the importance of proper wind design and consideration for wall panels, glazing, roofing, and mechanical equipment located on roofs. Better coordination between members of the design team, and between the design team and material vendors, is necessary to make sure that these components are being designed. All building departments should require the designer to submit calculations not only for main framing systems but also for cladding and accessory items such as wall panel connections, glazing, roofing, and mechanical equipment supports.