Mock-up Construction/Testing Configurations

Three full-scale identical curtain wall mock-ups, including the corner section (three face units and one corner unit), were fabricated. The elevation and plan views are shown in Figs. 3 and 4. Details of the typical vertical mullion, corner vertical mullion, and typical horizontal mullion are show in Figs. 5, 6, and 7. The structural sealant bead along the vertical mullion has a glue line thickness of 9/16 in. (14 mm) and a width of 9/16 in. (14 mm). The structural sealant bead along the horizontal mullion has a glue line thickness of 5/16 in. (8 mm) and a width of 9/16 in. (14 mm). Sealant joints were designed taking into considera­tion a typical wind load of 30-50 psf (1.4-2.4kPa) and a maximum glass width of 5 ft. (1524 mm), yielding a required sealant bite of 7/16 in. (11 mm) using the commonly accepted “trapezoidal loading theory" for calculating sealant bite dimension based on windload (ASTM C 1401 [8]). The equation for calculating sealant bite for rectangular lites of glass, for windload conditions, is as follows:

Sealant bite in inches = {0.5 • short span length (ft) – windload (psf)}/

{12in./ft • sealant design strength (20 psi)}

Sealant bite in mm = {0.5 • short span length (mm) – windload (kPa)}/ sealant design strength (138 kPa)

Although there is no published seismic sealant bite equation per se, it was good engineering judgment to increase the sealant bite as the curtain wall design

14115 mm)

allowed to provide a sealant thickness that could accommodate more shear deflection, while still maintaining a joint dimension that would typically be found in a real curtain wall design. Per industry standard structural glazing guidelines, sealant glueline thickness is not to exceed the sealant width or “bite" (ASTM C1401 [8]), thus the resultant sealant thickness was 9/16 in. (14mm) and 5/16 in. (8 mm) as noted above.

Three complete curtain wall mock-ups were constructed to have a repeat­able system to test, while allowing for varying boundary conditions to be stud­ied. The method by which the mullions are attached to the building can have a significant impact on the curtain wall performance independent of the glazing attachment method. For that reason it was decided to test these mock-ups in three distinct attachment configurations or boundary conditions. The three boundary conditions that were tested include: sway (boundary condition 1), rack with an allowance for vertical slip between vertical mullions (boundary





BEAD 0.56′(h) X 0.56 (w)

(14 mm x 14 mm)

(76 mm)

FIG. 5—Typical vertical split mullion.

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014 Downloaded/printed by



FIG. 6—Corner vertical split mullion.

condition 2), and rack (boundary condition 3). Refer to Fig. 8 for a description of these three boundary conditions.

Boundary condition 1, “sway," is commonly referred to as a “unitized" cur­tain wall system. Curtain wall units in this type of system are typically fabri­cated one building story in height. As they are attached to the building, the bottom of the unit is “nested" into the top of the unit below. This “nesting" allows the units above and below each other to drift in-plane independently. The only source of restraint that is expected will likely come from the corner

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014 Downloaded/printed by

condition. Boundary condition 3, “rack," is commonly referred to as a “stick – built" curtain wall system. In this type of system, the vertical mullions are attached to each building level to restrain movement of the mullion (with respect to the building structural system) in both the lateral (in the plane of the curtain wall) and transverse (normal to the plane of the curtain wall) directions. When the building displaces in a seismic event the vertical mullions will lean or rack with the building. Boundary condition 2, “rack with vertical slip," is a hybrid attachment system. The vertical mullions are attached at each building level, and therefore rack as the building drifts. During this racking, the vertical male-­female mullions are able to slip vertically slightly relative to each other. It is believed that this movement between the split vertical mullions will help to reduce some of the stress the sealant would normally experience during seismic racking behavior (boundary condition 3). It is therefore expected that this boundary condition will result in better performance than boundary condition 3.

Because the amount of sealant movement during boundary condition 1 is expected to be low (i. e., no sealant damage), it was assumed that it would be possible to test a second boundary condition (2 or 3) on the same physical mock-up unit. Many systems are attached such that sway is the primary expected movement during a seismic event. As the Zarghamee et al. [2] work attests to, sway does not allow much of the stress generated from the seismic movement to transfer to the glass-attachment system (regardless of its construc­tion type). Boundary condition 3, “rack," is expected to transfer the maximum amount of loading to the glass-attachment system, in this case, the silicone seal­ant on all four sides. This is the most severe case for a four-sided SSG system with regard to stress being transferred through the sealant. Boundary condition 2, “rack with vertical slip," is expected to produce results somewhere in between boundary conditions 1 and 3.