Look Out Below – A Curtainwall Case Study Jason Siwek, PE, and Jackie Byndas Prakhov Walter P Moore 1747 Pennsylvania Ave. NW, Suite 1050, Washington, DC 20006 202-365-6754 • jsiwek@walterpmoore.com • jbyndas@walterpmoore.com IIBEC 2020 Virtual International Conve ntion & Trade Show | June 12-14, 2020 Siwek and Prakhov | 21 Jason Siwek is a project manager in the diagnostics group of his company. With more than eight years of experience in the building enclosure field, Siwek has gained expertise in unitized curtainwalls, punched windows, storefronts, canopies, skylights, and handrails. His experience involves formulating test procedures to determine material properties, evaluating performance conditions, and recommending efficient processes for repair. Siwek has special insight into international building codes, standards, and market trends based on his work with global design teams. In the building enclosure industry, he specializes in enclosure fabrication and rope access assessments. Jackie Byndas Prakhov is a graduate engineer in the diagnostics group of her firm. She has more than three years of experience in the field of forensic engineering. Her expertise includes evaluating and designing repairs for distress related to clay masonry, stone façades, concrete structures, building enclosure moisture management, roofing systems, and belowgrade waterproofing on concrete substrates. She has also developed work scopes, repair details, repair procedures, and technical specifications for waterproofing, restoration, and rehabilitation projects. 22 | Siwek and Prakhov IIBEC 2020 Virtual International Conve ntion & Trade Show | June 12-14, 2020 ABSTRACT SPEAKERS Curtainwall systems have been used in commercial construction since the 1970s. Though the performance of these systems has improved, there is still much to learn about how design choices affect service life resilience. This presentation will discuss the fundamentals of commercial fenestration system types, performance, and proper detailing followed by a case study illustrating how these concepts were applied. The case study of Clements University Hospital, a fully operational 12-story healthcare facility, illustrates how typical detailing of newly incorporated elements could have adverse long-term effects. Beauty caps on a typical captured glass curtainwall system were falling and posed a life safety risk to the public. Walter P Moore was engaged to perform a peer review of the construction documents to identify possible causes of failure, as well as a visual survey to identify additional falling hazards. The Challenging Access Team mobilized to conduct the visual assessment of the tower curtainwall system by rope access. By implementing quality control processes for the repairs, performing hands-on observations, and examining the installed repairs, repair solutions would mitigate future falling hazards and reduce risks for the owner and building patrons. These and other challenges encountered throughout assessment, design, and construction oversight will be discussed. IIBEC 2020 Virtual International Conve ntion & Trade Show | June 12-14, 2020 Siwek and Prakhov | 23 The diversity in design of fenestration systems provides a creative foundation for the variety of building façade designs found today. Commercial fenestration systems include punched windows, storefront systems, and unitized curtainwalls, which are often utilized to achieve specific design aesthetics while also serving a functional purpose for occupants. Depending on the budget, construction methodology, and desired façade geometry, several different glazing systems can be implemented. This paper will assess the use of various fenestration systems for storefront systems, window wall, and curtainwall construction, and recap the common installation guidelines for a continuous enclosure. COMMERCIAL FENESTRATION SYSTEMS Window Wall When an exterior wall is composed of windows that sit between the floor slabs, the fenestration system is known as a “window wall” (Figure 1). Window walls typically span one story from the top of a slab on a lower level to the soffit of the slab on the next level, meaning each is anchored at both the top and bottom of every slab and has relatively shallow mullions due to their single-story span. Commonly, due to cost savings, window walls are used to mimic the appearance of curtainwalls by using a slab cover. Despite the aesthetic goal to mimic a curtainwall, the construction for window walls is similar to a storefront system, as a window wall can also be stick-built. With a stick-built construction, all the components are assembled and installed on site, or alternatively, hybrid units shipped in preconstructed cassettes can be set into carrier framing elements on site. One advantage of using a hybrid or cassette/carrier is greater quality control during the manufacturing process as compared to onsite conditions. For window walls, the airand water-resistive detailing is similar to a storefront system, as both require the interior seal to connect to the air/water-resistive barrier material at the rough opening to minimize air and water infiltration. The transitions to the adjacent air barrier are critical to prevent water infiltration. However, window wall construction is more complex than storefront construction in anchoring the windows to the slabs and transitioning in front of the slab edge. A window wall is usually categorized by the following characteristics: • Spans a single story • Can be one of two construction methods: stick, or hybrid with stick frames and unitized cassette/carrier • Is anchored at top and bottom of every slab • Uses slab covers to mimic curtainwall aesthetic • Has relatively shallow mullions (typically 4–6 in. deep), due to span Curtainwall Curtainwall systems (Figure 2) are the most complicated, versatile, and expensive options of the commercial fenestration Look Out Below – A Curtainwall Case Study Figure 1 – Window wall terminology. Figure 2 – Curtainwall systems; stick-built (left) and unitized (right). Graphic by Elicc Group. systems discussed. Typically, curtainwall systems are more appealing aesthetically due to their clean, sleek appearance and their ability to allow natural light to flow into the interior space. Additionally, high-rise office and residential towers commonly use a curtainwall façade for the ability to accommodate design features, such as sunshades. Unlike storefront or window wall systems, which bear on the concrete slab at each floor, curtainwall systems sit outboard of the slab and connect to the exterior face of the concrete slab with anchors. As a result, curtainwall systems can span multiple levels. The vertical mullions tend to be deeper, depending on the span of the curtainwall component. Curtainwalls can be categorized based on their manufacturing methods as well as installation. Two of the most common curtainwall systems are stick-built systems and unitized systems. Curtainwall: Stick-Built A stick-built system (Figure 3) is assembled at the job site. For projects with unique curtainwall geometry, such as a sloped façade, stick-built systems are the most cost effective. Nonetheless, stick-built systems may have additional quality control issues if a stringent quality control plan is not in place. Stick-built systems are usually built to span two levels and are spliced at the vertical mullions between each unit. For units spanning more than one level, the vertical mullions are typically deeper in order to accommodate greater exterior loads applied to the façade system. Typically, the two-level unit is dead-loaded at one floor with a wind load anchor at the other. The key difference between the anchors is that the wind load anchor has slotted holes to allow for movement. In summary, the stick-built system is characterized by: • On-site erection and glazing • Conventionally glazed • Composed of long aluminum mullions Curtainwall: Unitized Unlike the stick-built system, a unitized curtainwall system (Figure 4) is assembled in a factory setting as a single unit that can be quickly installed on site. The unitized curtainwall anatomy has many of the same components as a stick-built curtainwall system, including vertical mullions, horizontal transoms, and anchors; however, there are a few key differences. Since the unitized curtainwall system must interlock to adjacent panels, the vertical mullions are composed of one male and one female component. Additionally, the top horizontal transom has a stack joint, also commonly referred to as a “chicken head,” that interlocks with the sill of the panel above. Typically located at every level, this stack joint can better accommodate differential movement throughout the façade and, thusly, is commonly used in seismic areas. Since unitized curtainwall panels are manufactured in a controlled environment, they are subject to more stringent assembly requirements. Within the factory, the critical seals are installed and tested to conform to industry standards rather than relying on the on-site construction method of the stick-built curtainwall system. This gives a higher level of quality assurance for moisture-critical spaces. In conclusion, unitized systems allow for faster installation, can better accommodate building movement due to the stack joint (chicken head), and offer 24 | SiwekEK and Prakhov IIBEC 2020 Virtual International ConveVEntion & Trade Show | June 12-14, 2020 Figure 3 – Stick-built construction. Figure 4 – Unitized system. better quality assurance due to controlled manufacturing. There are several factors to consider when determining the type of system to use. Stick-built systems offer customization for a lower price compared to unitized curtainwalls. Costs can rise exponentially for a manufacturer who needs to produce a customized, unitized curtainwall compared to a system that is already in production with a refined manufacturing process. Due to associated costs with customization, in the case of curtainwalls with sloped, curved, or other geometric irregularities, the owner or architect may choose a stick-built system instead. Typically, unitized systems are favorable for high-rise construction with repetitive designs. In a building where the façade is only partially composed of a curtainwall system, a stick-built system may be preferred. In contrast, with large surface areas, a unitized system will likely be favorable due to the ease of installation and ability to adhere to schedule demands. Stick-built systems are also more likely to be used in locations where site labor costs are lower, meaning façades assembled on site are cheaper than the manufactured curtainwall system. Glazing Conventional Glazing Another determining factor is choosing the appropriate glazing process. Typically, stick-built systems are conventionally glazed, meaning there are gaskets on both the interior and exterior sides of the glazing with a pressure plate to hold the assembly in place. In an exterior, conventionally glazed system, a small piece of aluminum extends out from the horizontal transom and is commonly referred to as a setting chair. A setting block is then placed on the setting chair. The contractor then installs the glass/exterior gaskets and holds the assembly together with a screw spline and pressure plate. Finally, a face or “beauty” cap connects with the pressure plate to hide the screws from public view. This system is considered an exterior, dry, conventionally captured glazed system (Figure 5). In addition to conventional glazing, there is an interior, dry, conventionally captured system (Figure 6). It is assembled in opposite order of the conventional exterior glazed system. Instead of a pressure plate, there is a molded piece of aluminum that is continuous from the exterior (where a pressure plate would be) to the interior mullion. The exterior gaskets are placed first, followed by the glazing, and finally the installation of the interior gaskets and glazing stop. The glazing stop forms part of the mullion that is visible from the interior and, similar to the exterior conventional system, holds the components together. A conventional captured system—whether it is internally or externally glazed, relies on a clamping mechanism to hold the glass in place. Unitized curtainwalls can also be captured systems, as described above, and arrive at the site with the glazing already installed. Structural Silicone Glazing Curtainwall systems—typically unitized—can be structurally glazed as well IIBEC 2020 Virtual International ConveVEntion & Trade Show | June 12-14, 2020 SiwekEK and Prakhov | 25 Figure 5 – Exterior, dry glazed. Figure 6 – Interior, dry glazed. Figure 7 – Structurally glazed. (Figure 7). The primary difference between structural glazing and a captured system is that the former relies on structural silicone sealant instead of a compression mechanism to hold the glass in place and transfer environmental loads from the glass to the mullions. Setting chairs are installed at quarter points of the horizontal transom to support the self-weight of the glass. Gaskets and a bed of structural sealant between the glazing and mullion, are installed on all four sides of the piece of glass. Finally, an exterior sealant is applied at the exterior between lites of glass in order to prevent water infiltration at these joints. The integrity and quality of the sealant is responsible for maintaining the integrity of the glazing. Therefore, most structurally wet-glazed systems are unitized systems, providing additional quality assurance. These systems are typically used when the architect wants to eliminate the exterior pressure plate or portion of the mullion and associated face, giving the structure a seamless appearance. Thermally Improved Systems Curtainwall systems fall into three categories: not thermally improved, thermally separated, or thermally broken systems (Figure 8). In order to prevent thermal bridging, a curtainwall mullion can be thermally improved. To be thermally broken, the exterior metal and interior metal will be disconnected by using polyester reinforced nylon connectors between the exterior portion of the mullion or pressure plate and the interior portion of the mullion. These components, commonly referred to as “dog bones,” are secured by crimping the hammer to the anvil elements created in the aluminum extrusion. By providing connections of a nonconductive material such as polyester-reinforced nylon, heat transfer is minimized and therefore considered thermally broken. A thermally improved system uses a clip composed of a nonconductive material to connect the exterior face, or beauty cap. CASE STUDY Clements University Hospital (CUH) is a fully operational, 12-story hospital on the University of Texas Southwestern (UTSW) Medical Center 26 | SiwekEK and Prakhov IIBEC 2020 Virtual International ConveVEntion & Trade Show | June 12-14, 2020 Figure 8 – Nonthermally broken (left), thermally broken (center), and thermally separated (right). Figure 9 – Overview of UTSW Medical Center fenestration system. campus located in Dallas, TX. The building is concrete-framed and clad with precast architectural concrete panels and an aluminum-framed curtainwall product in a window wall configuration. Its corner sections are clad with a continuous unitized curtainwall product. The ground level has a stick-built configuration (Figure 9). The window wall system is an interior-glazed system with thermally separated aluminum extrusions. The shop drawings illustrate horizontal and vertical field-applied aluminum beauty caps that engage shop-installed, 1-in.-long polyvinyl chloride (PVC) thermal isolators. The thermal isolators engage the horizontal and vertical aluminum extrusions at 12 in. on center and are intended to provide thermal separation between the interior and exterior. The window wall system also features vertical sunshades, or “fins,” in regularly spaced stacks that extend the full height of the tower. Each individual vertical fin spans the length of a single floor and includes a two-part installation that snaps into the aluminum mullion. The fins remain in place via friction and a spring pin installed at the top and bottom of each fin. A vertical alignment pin at the exterior edge of the fin aligns adjacent fins within a stack. In February 2016, exterior aluminum beauty caps on the window wall horizontal mullions disengaged from the mullions and fell to the lower-level triangular roof. Similar subsequent failures were reported in March and April 2016, and in July 2016, a vertical architectural face cap fin fell from the window wall at the upper levels during a high-wind event. In total, nine face caps and one fin had disengaged from the exterior face of the existing fenestration systems. A review of construction documents pertaining to the curtainwall systems and a visual assessment of the existing fenestration systems were required to determine the extent of failures and determine an appropriate remedial action, restoring integrity and serviceability and reviewing proposed repairs by the curtainwall manufacturer. Document Review To understand existing conditions and methods of construction, the existing architectural drawings and shop drawings were reviewed. They revealed the following: • The typical window wall head detail shows a face cap engaged by a thermal isolator on each end. The face cap cross section includes receiver tracks that engage the isolator clips. • The typical vertical face cap extension fin shows an inner extrusion of the face cap that is mechanically fastened to the mullion and the outer extrusion, which is snapped to the inner part. • The window wall transom detail shows a face cap that is engaged to the mullion extrusion via a thermal isolator. A comment from the design team about the intersection of the face cap with the thermal isolator states, “Confirm size of isolator clip to assure face caps are/remain securely affixed to extrusion?” • A vertical face cap extension fin is mechanically attached on one side to the mullion (Figure 9). Design team comment: “Verify span (length) of face cap that can be installed and anticipated thermal movement.” • The primary vertical face cap (Figure 10) extension piece is shown unfastened to the secondary extension fin. Design team comment: “Typical. Provide mechanical fastening of face cap extensions.” On August 5, 2016, the manufacturer issued a letter stating that they reviewed the dislodged covers on site. The letter asserts the following: • Repairs in May and June of 2015 included replacement of thermal isolators to allow more thermal expansion as well as verification of engagement of the thermal isolators with the face caps. • The possible causes for dislodged face caps include wind, rain, and window washing activity. Based on field dynamic testing performed on the stick-built system in March of 2014, as well as a demonstrative face cap removal on the adjacent unitized system, the manufacturer concluded that window washing activity was the likely source for the face cap failure. IIBEC 2020 Virtual International ConveVEntion & Trade Show | June 12-14, 2020 SiwekEK and Prakhov | 27 Figure 10 – Fin assembly. Analysis Analysis and associated calculations were performed to evaluate the potential causes for the face cap disengagement, with wind events and window washing activities considered as potential causes. Although the manufacturer also asserted rain as a possible cause of disengagement, calculations were unable to confirm any significant force that rain could have exerted on the face caps, and so it was not included as a potential cause in the analysis. Design Wind Loads The team calculated the negative pressure from leeward wind loads that the face plates would need to resist. In the case of a 90-mph wind speed for a components-and-cladding design load case, it was determined that a typical full-length transom beauty cap must resist 80 lbs. of negative pressure in a design wind event. During on-site assessments, 14 thermal isolators were observed at failed mullion locations. If this installation is typical, the design force per thermal isolator connection is 5.7 lbs. per connection. Worker Loads A 250-lb. worker exerting his or her full weight on the edge of a horizontal face cap could potentially generate a force couple resulting in a horizontal load of 3.6 lbs. per connection. However, this is a very conservative estimate, assuming the worker is balancing his or her entire weight on the edge of the ½-in.-wide top surface of the cap without bearing any weight on the adjacent window surface. The load assumption also neglects the fact that the window washer is suspended with rope access equipment. IN-SERVICE WIND LOADS AT TIME OF FAILURES Wind speed data were gathered from the closest National Oceanic and Atmospheric Administration (NOAA) weather station, Dallas Fort Worth Airport, with equipment placed at a 33-in. height in design exposure category C. Background information from UTSW indicates that mullion face cap failures occurred during the following time periods with the highest recorded wind speeds from NOAA: 1. Horizontal mullion face cap failure occurred in February 2016. The highest wind speed in February 2016 in Dallas was 37 mph, with the highest gusts at 46 mph on February 8. 2. Vertical mullion face cap failure occurred in March 2016. The highest wind speed in March 2016 in Dallas was 40 mph on March 8. 3. UTSW estimates the vertical extension fin fell off in March or April 2016. The highest wind speed in April 2016 in Dallas was 41 mph on April 26. 4. Two vertical face caps fell off on July 9 at approximately 5 PM. The wind speed reported for July 9 was 46 mph. The design wind speed for the building was calculated for a 120-mph, Risk Category IV (hospital) according to the International Building Code (IBC) and ASCE 7. In each of the above time periods during which failures occurred, the reported wind speeds in the Dallas area did not approach the basic (design) wind speed. The vertical mullion extension fins are secured in place by means of an extruded clip connection with two pins through the extension—one at the top and one at the bottom of the extension—presumably to resist the gravity loads. The design wind loading on a vertical fin was calculated and 28 | SiwekEK and Prakhov IIBEC 2020 Virtual International ConveVEntion & Trade Show | June 12-14, 2020 Analysis and associated calculations were performed to evaluate the potential causes for the face cap disengagement, with wind events and window washing activities considered as potential causes. Figure 11 – Face cap assembly. it was determined that the vertical extension connection would have to resist a combined load of 347-lb. laterally and an outward force of 729 lb., resulting from the moment due to wind loads. The calculation package for the vertical extension that was submitted by the curtainwall installer did not check the capacity of the extension connection. Face Cap Thermal Expansion In general, failures manifested as either displacement at the ends of the face cap immediately adjacent to the end joint or as “bowing out” at the midspan of the face cap. These conditions suggest that the mullion face caps are losing connection with the thermal isolator clips because of inadequate allowance for face cap thermal movement. This possible failure mechanism is substantiated by observation of closed end joints between the ends of face plates and adjacent face caps. Previous Test Results Field testing according to AAMA 501.1, Standard Test Method for Water Penetration of Windows, Curtain Walls and Doors Using Dynamic Pressure, occurred only on the lower podium stick-built curtainwall system, and not the window wall system on the tower. Given that a different window system was tested, results from the field dynamic testing may not be directly applicable to the face cap failures. The testing performed on the stick-built system was at a much lower pressure (8 psf) than design wind pressures (32.8 psf). Further, this testing does not provide any information related to the capacity of the vertical mullion extension fin, as the stick-built system does not have the same architectural fin feature as the window wall system. Observations A visual evaluation of the installed window wall system was performed to reasonably assess the general nature and extent of the face cap and fin attachment failures and to determine an appropriate scope of remedial action to restore integrity and serviceability. Visual evaluations were performed by rope access at the tower and accessible parts of the podium (Figure 12). Observations were intended to confirm construction conditions and to obtain information not found in previous document review. Quantities and location of elements that appeared to be at imminent risk of failure were logged. Survey Methodology A hands-on survey of the installed window wall system was performed at 100% of the tower’s accessible locations and portions of the podium stick-built system. There were three major objectives in performing the survey: 1. Identify at-risk face caps and fins in IIBEC 2020 Virtual International ConveVEntion & Trade Show | June 12-14, 2020 SiwekEK and Prakhov | 29 Figure 12 – Performing visual assessment via rope access. imminent risk of failure. 2. Make detailed observations of the face caps and fin that had previously failed. 3. Record other typical conditions at the fenestrations that may be related to the face cap issues. A vertical or horizontal face cap was determined to be at risk of failure using the following criteria: • Visible displacement of the face cap from the curtainwall extrusion, including bowing of the face cap, a visible gap between the face cap and the extrusion, or thermal isolators that were visible behind the face cap, indicating outward movement of the face cap from the extrusion • Movement of the face cap when a very light outward pulling force was applied by hand A vertical fin was determined to be at risk of failure using the following criteria: • Visible displacement of the spring pins at the vertical fin, including misaligned or disengaged pins, loose pins, or missing pins • Movement of the vertical fin when a very light pulling force was applied by hand Site Visit Observations From October 27 to November 11, 2016, engineers certified by the Society of Professional Rope Access Technicians (SPRAT) performed drops via rope access at the main elevations of the tower. Portions of the podium tower were reviewed via aerial lift and from the ground on December 9 and 10, 2016. At-Risk Face Caps From the visual observation, it was found that 188 of 5,747 horizontal face caps were at risk and 49 of 624 vertical face caps were at risk. The following conditions were observed at the at-risk face cap locations: • The at-risk face caps exhibited either visible displacement or bowing, visible isolators behind the face cap, or movement with very light pulling force, applied by hand. — Less than 10% of the at-risk face caps exhibited a visible 30 | SiwekEK and Prakhov IIBEC 2020 Virtual International ConveVEntion & Trade Show | June 12-14, 2020 Figure 13 – Face cap disengagement. Figure 14 – Spring pin disengaged. displacement of the face cap from the extrusion, exposing the thermal isolators to view. — The majority (over 90%) of the at-risk face caps reviewed did not exhibit visible displacement of the face cap from the extrusion; however, once a light pulling pressure was applied by hand, the face cap easily became disengaged from the thermal isolator (Figure 13). All disengaged face caps were pushed back into their original connected position after the review was complete. • When pulled, the at-risk face caps disengaged from the thermal isolator. Disengagement of the thermal isolators from the extrusion was not observed. At-Risk Vertical Fins It was determined that 230 of 764 vertical fins were at risk at the tower areas (there were no fins installed at the podium curtainwall). In general, the following conditions at the at-risk fin locations were observed: • The majority of the at-risk vertical fins exhibited either missing pins, misaligned pins, or partially disengaged pins (Figure 14). • Two of the 230 at-risk fins exhibited movement of the fin when manipulated with very light pulling force, applied by hand. • Partially disengaged or missing spring pins were located either in the outer fin extrusion or the “middle” fin extrusion. • The spring pins did not have a head and could be moved inward or outward from the aluminum fin by hand in many locations. Previous Face Cap Failure Locations The team observed locations where eight horizontal face caps and three vertical face caps previously fell from the tower. And while no failures were observed at the podium locations, the following were observed at the previous failure locations at the tower: • The PVC thermal isolators were typically spaced 8 to 10 inches on center. In some locations, several isolators were grouped side by side. • The isolators could easily slide along the track of the extrusion. • Buildup of sealant on top of the PVC thermal isolators appears to have been added in a failed effort to adhere the face caps to the clips. • Sealant was installed along the entire length of the extrusion in some locations. This sealant appears to have been added in a failed repair effort to adhere the caps to the extrusion. Previous Fin Failure Locations Close-up observations at the locations where the outer piece of one vertical fin had fallen from the tower revealed the following: • The spring pins installed at the top and bottom of the fin remained in place on the inner part of the vertical fin that remained attached to the building. The pins appeared to have moved laterally within the extrusion, resulting in disengagement from the outer part of the fin that had fallen. The pins did not exhibit failure in shear. • The edge distance between the pins and the outer edge of the inner extrusion was minimal. Summary of Observations During several site visits to visually assess the existing curtainwall systems, it was determined that 182 horizontal beauty caps and 46 vertical fins were at risk of disengaging from the curtainwall mullions. At-risk face caps exhibited one or more of the following: visible displacement/bowing, visible thermal isolators behind the face caps, or movement with very light pulling applied by hand. Approximately 10% of at-risk face caps were visibly displaced, with the remaining 90% moving with a light pulling pressure by hand. When the light pressure was applied, the face cap disengaged from the thermal isolator while the isolator remained engaged with the mullion. In addition, many of the vertical fins were either missing or had partially disengaged. During the survey, thermal isolators were typically spaced 8 to 10 in. with areas of bunching, resulting in a much larger spacing. Discussion The following were determined from the extensive review of the curtainwall: 1. Wind loads are higher than worker loads, so a properly designed system would be able to handle worker loads. 2. The system has not experienced design wind loads during any of the documented face cap failures. 3. Concerns were raised by the original design consultant with respect to face cap engagement, but no documented resolution of these concerns was found. 4. The dynamic, on-site testing was performed on a different window system and with wind pressures much lower than pressures required for design and, therefore, does not demonstrate performance of the face caps in the window wall under design wind loading. 5. Calculations or testing documents that demonstrated the design capacity of the face cap connections were not found. Regardless of the cause of the failures, the condition posed an overhead safety hazard, and actions had to be taken to IIBEC 2020 Virtual International ConveVEntion & Trade Show | June 12-14, 2020 SiwekEK and Prakhov | 31 During several site visits to visually assess the existing curtainwall systems, it was determined that 182 horizontal beauty caps and 46 vertical fins were at risk of disengaging from the curtainwall mullions. protect the public. Based on the review of the available documents, engineering analysis, and observations in the field, the observed face cap failures are systemic and are not conclusively related to window washing activities. The presence of other loose face caps on the curtainwall represents an overhead safety hazard. Previous attempts to repair the face caps with sealant did not adequately address the systemic issues that led to their failure. RECOMMENDATIONS Horizontal and Vertical Face Caps: • Engage a contractor to immediately remove all vertical and horizontal face caps that have been identified as being at imminent risk for failure. — Install overhead protection to address risk from potential additional failures until repairs can be made. • Design repairs for all vertical and horizontal face caps that will provide continuous engagement. Repair options include properly detailed adhesion with structural silicone and/or concealed fasteners. — Provide quality control measures (e.g., load testing) to verify attachment of the designed solution. — Perform periodic site visits during installation to provide quality assurance. Vertical Fins: • Due to their interlocking, stacking nature, the possibility of unintentionally causing additional failures of repair fins is likely, in our opinion. A repair-in-place approach can be installed with minimal risk and at a lower cost compared to temporarily removing the at-risk fins and subsequently stabilizing fins once a portion has been removed. • Design repairs for all vertical fins will provide additional dead-load support (e.g., self-tapping screws installed in new holes adjacent to the existing spring pins). — Provide quality control measures (e.g., load testing) to verify attachment of the designed solution. — Perform periodic site visits during installation to provide quality assurance. Final Repair Solution A positive mechanical attachment was determined to be the best repair solution. A jig was developed to prevent a blind application of a fastener into the glazing behind the face caps and to limit damage to the existing glass if a new anchor is not installed in the correct position. The jig 32 | SiwekEK and Prakhov IIBEC 2020 Virtual International ConveVEntion & Trade Show | June 12-14, 2020 Figure 16 – Vertical face cap extension fix with fasteners (red). Figure 15 – Horizontal fix with fasteners (red). allows a self-drilling/tapping fastener to penetrate the glazing pocket without damaging the glass. The fastener was inserted at the bottom edge of the glass where the glass dead-load supports were located to ensure they did not interfere with glass movement in the glazing pocket (Figures 15 and 16). Based on observations during the mock-up and repair installation, the drill jigs adequately addressed the risks of mechanical attachment. Glass breakage or fogging of the IGUs in the areas observed during the mock-up phase and subsequent site visits during the repairs indicate a successful mechanical attachment. The Clements University Hospital case study identified the importance of ensuring all architectural features comply with the components and cladding wind loads according to IBC and ASCE 7 standards. The falling face caps and architectural fins could have seriously injured hospital patrons. It illustrates the necessity of ensuring all aspects of the curtainwall system are designed and constructed for loading and are integrated with adjacent systems. The architectural fins created a distinctive feature and provided shade, but the desired length exceeded the limits of die casting for the aluminum extrusion. What resulted is the need for a two-part cantilever fin with friction fit snap caps. Concerning the transom beauty caps, the intent was to provide a thermally improved system for occupant comfort and the thermal performance of the building, but the building experienced unintended consequences as a result of the periodically spaced thermal isolators that separated the exterior beauty cap and the mullion. Aspects of the fenestration system need to be thought through with a holistic understanding of how design choices affect production, performance, and service life resilience. IIBEC 2020 Virtual International ConveVEntion & Trade Show | June 12-14, 2020 SiwekEK and Prakhov | 33
