Lateral Resistance Reduction to Cold-Formed Steel-Framed Shear Walls under Various Fire Scenarios

Abstract

This paper examines the structural response of cold-formed steel-framed building lateral force-resisting systems under combinations of simulated earthquake and fire loading. Full-scale experiments with gypsum-sheet steel composite panel sheathed walls, oriented strand board sheathed walls, and steel strap braced walls are presented. Twenty-two test specimens are subjected sequentially to combinations of cyclic shear deformation and fires of varying intensity; some approximate temperatures in standard furnace tests, and most have characteristics of actual building fires. In select tests, the walls are predamaged to simulate fire following an earthquake. The results show a progressive decrease of postfire lateral load capacity with increasing fire intensity for all walls; however, each wall type exhibits varied sensitivity to the fire intensity as well as to predamage. By understanding the response of these structural systems in real fires, designers can better plan for situations in which multiple hazards, including fire, exist.

Introduction

Lightweight framing systems are commonly used to construct single-story buildings; however, they are also used for the structural system in multistory buildings. The global trend toward urbanization (UNDESA 2014) places an increased demand for high-density, urban housing that optimizes land use. Consequently, we see an increase in multifamily housing. Lightweight structural systems using cold-formed steel currently represent about 20% of the nonresidential construction market in the United States (multifamily housing with five or more dwellings is considered commercial real estate) and are touted for their cost advantages and rapid construction speed—in particular, for prefabricated systems. However, an important consideration for tall, lightweight structures is their performance in fire because as buildings get taller, evacuation times increase, and fighting a fire becomes more challenging.

In addition to the gravity load-resisting system that supports the weight of a building and its contents, nearly all buildings have lateral force-resisting systems to resist horizontal loads, such as those due to wind or earthquakes. Although extensive information exists about the structural performance and fire resistance of cold-formed steel construction, (e.g., Schafer et al. 2016; Sultan 1996; Takeda 2003; Wang et al. 2015), there is limited knowledge about the behavior of cold-formed steel lateral force-resisting systems under combined hazards.

In 2016, a series of experiments (Phase 1) was performed at the National Fire Research Laboratory (NFRL) at the National Institute of Standards and Technology (NIST) to investigate the performance of earthquake-damaged gypsum-sheet steel composite panel sheathed cold-formed steel shear walls under fire load (Hoehler and Smith 2016, 2018). The tests indicated that a change in failure mode could occur in the walls from local to global buckling of the sheet steel following a fire and highlighted the importance of the response of the gypsum for both the fire and structural behavior (Hoehler et al. 2017).

This paper extends the Phase 1 research and presents experimental investigations of the performance of three common cold-formed steel lateral force-resisting systems under combinations of earthquake and fire loading. Fire loads of varying intensity that represent the characteristics of actual fires, in addition to tests with temperatures similar to those in standard furnace tests, are used. This paper focuses on the structural response. The thermal response of the wall systems is discussed in a separate paper (Andres et al., forthcoming). Details can be found in the full test report (Hoehler et al. 2019a) and dataset (Hoehler et al. 2019b).

Experimental Program

Table 1 shows the test matrix for this study. Three lateral force-resisting systems were investigated: gypsum-sheet steel composite panel sheathed walls, oriented strand board (OSB) sheathed walls, and steel strap braced walls. Specimen names, including 01, were subjected only to load cycling. Specimen names, including 02, 03, or 04, were subjected to fire of varying intensity followed by cyclic loading. Specimen names, including 05 or 06, were predamaged with cyclic loading, subjected to fire, and then cycled to failure. Specimens with an R designation were test replicates, except for S01R, which was a redesign of wall S01 to use symmetric bracing. Specimen OSB01NG was an OSB sheathed wall with no gypsum board installed on one side of the wall. OSB Kitchen was a test in which the fire load was provided by burning kitchen furnishings; this test involved only thermal loading, so it is discussed in a separate paper (Andres et al., forthcoming).

Table 1.

Test matrix

		Loading
Wall type	Specimen name	Cycling (before fire)	Fire	Cycling (after fire)
Gypsum-sheet steel composite	SB01	To failure	—	—
	SB02	—	Severe fire	To failure
	SB03	—	Mild fire	To failure
	SB03R	—	Mild fire	To failure
	SB04	—	Standard fire	To failure
Oriented strand board	OSB01	To failure	—	—
	OSB01R	To failure	—	—
	OSB02	—	Severe fire	To failure
	OSB03	—	Mild fire	To failure
	OSB03R	—	Mild fire	To failure
	OSB04	—	Standard fire	To failure
	OSB05	Drift level 2	Mild fire	To failure
	OSB06	Drift level 1	Mild fire	To failure
Steel strap braced	S01	To failure	—	—
	S01R	To failure	—	—
	S02	—	Severe fire	To failure
	S03	—	Mild fire	To failure
	S04	—	Standard fire	To failure
	S05	Drift level 2	Mild fire	To failure
	S06	Drift level 1	Mild fire	To failure
Additional	OSB01NG	To failure	—	—
	OSB Kitchen	—	Real furnishings	—

Specimen Geometries and Material Properties

A large percentage of lateral-force resisting systems used in cold-formed steel structures are shear walls located on the interiors of buildings. Consequently, we focused this investigation on interior walls. Fire-resistance requirements for exterior walls and gravityload bearing walls were not considered, nor were the effects of supplemental vertical (gravity) loads because these topics have been investigated by other authors (Alfawakhiri 2001; Ariyanayagam and Mahendran 2015; Chen et al. 2013; Feng and Wang 2005; Gunalan et al. 2013). Although the wall specimens were not designed based on a specific prototype building, the detailing was selected to be typical of that used in moderately-high seismic regions of the United States.

Fig. 1 shows photographs and cross-sections of the investigated wall types. All wall specimens were 2.7 m tall by 3.7 m long (9 × 12 ft). The walls were designed using the allowable stress design (ASD) following the American Iron and Steel Institute (AISI) standards S100 (AISI 2016a) and AISI S400–15/S1 (AISI 2016b). Certain wall elements, e.g., the steel strap braces, were capacity designed to achieve specified failure modes in the absence of fire. All cold-formed steel sections were structural grade 50 (Grade 340) Type H, conforming to ASTM A1003 (ASTM 2015) with a minimum specified yield strength of 345 MPa (50 ksi). The contribution of the gypsum boards to the shear capacity of the walls was neglected during design, as is typically done in current practice.

Fig. 1.

Photographs of partially-constructed test specimens and cross-sections: (a) gypsum-steel sheet composite panel sheathed walls; (b) oriented strand board sheathed walls; and (c) steel strap braced walls (1 = steel stud; 2 = structural sheathing or diagonal bracing; and 3 = gypsum board).

The scenario of a postflashover kitchen fire in a room adjacent to a corridor was used as a prototype for these investigations (Fig. 2). All wall specimens were designed to achieve a 1-h fire-resistance rating per ASTM standard E119 (ASTM 2016). The influence of insulation material in the wall cavity on the thermal and mechanical behavior of the walls was not investigated.

Fig. 2.

Investigated fire scenario.

Gypsum-Sheet Steel Composite Panel Sheathed Walls

The gypsum-sheet steel composite panels were a proprietary product (Sure-Board Series 200), which consisted of 0.69 mm (0.027 in.) thick sheet metal adhered to 16 mm (⅝ in.) thick Type X gypsum board. The specimen framing is shown in Fig. 3 with details provided in Table 2. Two 600S250–54(50) cold-formed steel studs, connected back-to-back with an unbraced length of 1.2 m (48 in.), provided sufficient axial design strength for the chord studs, assuming an expected strength factor of 1.2 for the composite panels based on the Phase 1 test results (Hoehler and Smith 2016). The chord studs were secured back-to-back with two rows of screws at 305 mm (12 in.) on center and stitched at the ends with screws spaced at 38 mm (1.5 in.) or less. All other frame joints had one screw where the flanges overlapped. The design calculations for all wall specimens are provided in Hoehler et al. (2019a).

Fig. 3.

Drawing of framing for the sheathed walls (1 ft = 30.5 cm).

Table 2.

Specimen details

Specimen properties	Gypsum-sheet steel composite	Oriented strand board	Steel strap braced
Shear resisting element	Sure-board series 200	7/16 in. thick OSB	3.25 in. × 0.0713 in.
Shear element fasteners	#8 × 1-¾ in.	#8 × 1-¾ in.	#10 × ¾ in.
Shear element fastener spacing	4 in. O.C. (edges)	4 in. O.C. (edges)	8 rows of 4 screws
	12 in. O.C. (field)	12 in. O.C. (field)
Framing—top/bottom track	600T125-54 (50)	600T125-54 (50)	600T150-68 (50)
Framing—chord studs	600S250-54 (50) BB	600S250-54 (50) BB	600S250-54 (50) BB
Framing—interior studs	600S162-54 (50)	600S162-54 (50)	600S162-54 (50)
Framing—bridging	150U50-54 (50)	150U50-54 (50)	150U50-54 (50)
Framing—fasteners	#10 × ¾ in.	#10 × ¾ in.	#10 × ¾ in.
Framing—holdowns	S/HD10S	S/HD10S	S/HD10S
Gypsum boards	⅝ in. thick type X	⅝ in. thick type X	⅝ in. thick type X
Gypsum fasteners	#6 × 1–⅝ in.	#6 × 1–⅝ in.	#6 × 1–⅝ in.
Gypsum fastener spacing	6 in. O.C. (edges)	6 in. O.C. (edges)	6 in. O.C. (edges)
	12 in. O.C. (field)	12 in. O.C. (field)	12 in. O.C. (field)

Note: O.C. = on-center; OSB = oriented strand board; BB = back-to-back. Diameter: #6 ≈ 3.5 mm; #8 ≈ 4 mm; #10 ≈ 5 mm; and 1 in. = 25.4 mm.

The design for 1-h fire-resistance of the gypsum-sheet steel composite panels walls was based on product approvals (Intertek 2010). The fire exposed side of the wall was sheathed using Type X gypsum boards 16 mm (⅝ in.) thick. The seams were sealed using a base coat of drywall joint compound and paper tape. The seams and fastener heads were then covered with a skim coat of the drywall joint compound. The vertical seams on the fire-exposed and fire-unexposed sides of the wall were staggered by one-stud spacing. No drywall joint compound or paper tape was used on the side of the wall not exposed to fire. The measured moisture content of specimens from three randomly sampled gypsum boards was (18.9 ± 0.1) percent by mass at the time of fire testing.

Oriented Strand Board Sheathed Walls

The design of the oriented strand board sheathed walls was analogous to the gypsum-sheet steel composite panel sheathed walls; however, the sheathing elements were 11 mm (⁷/₁₆ in.) thick wood structural panels rated for shear resistance. Because the maximum design chord force was smaller than that for the gypsum-sheet steel composite panel sheathed walls, the same framing could be used (refer to Fig. 3 and Table 2).

Three OSB panels were installed with vertical seams between the panels. The measured moisture content of specimens from three randomly sampled OSB boards was (7.4 ± 0.02) percent by mass at the time of fire testing. Type X gypsum boards 16 mm (⅝ in.) thick were then attached to the oriented strand boards to meet the 1-h fire-resistance rating requirements per Underwriters Laboratory (UL) Design No. U423 (UL 2017a), with the addition of wood panels as contemplated in UL (2017b). The vertical seams between the oriented strand boards and the gypsum boards were staggered. The gypsum boards were attached to the oriented strand boards using 3.5 × 41 mm (6 × 1–⅝ in.) bugle-headed drywall screws spaced 305 mm (12 in.) on center and positioned to avoid fastening to the framing. The authors elected not to attach the gypsum board through the oriented strand boards to the framing to maintain the fastener spacing on the oriented strand boards panels. No drywall joint compound or paper tape was used on this side (the unexposed side) of the wall. The opposite side of the wall (the fire-exposed side) was sheathed analogous to the gypsum-sheet steel composite panel sheathed walls, except for Specimen OSB01NG, which had no gypsum on the unexposed side of the wall.

Steel Strap Braced Walls

The steel strap braced walls were designed to achieve significant yielding of the straps before failure. Initially, the wall was designed with straps only on one side of the wall (asymmetric bracing) using the same stud and track dimensions as for the sheathed walls. However, during the first cyclic loading test, the wall failed before the strap yielded due to a combination of buckling and torsional failure of the top track, so the wall was redesigned using bracing on both sides of the wall (symmetric bracing) to eliminate torsion.

To keep the chord stud design the same as in the sheathed walls, a strap width of 83 mm (3.25 in.) and a thickness of 1.8 mm (0.0713 in.) was chosen. The specimen framing is shown in Fig. 4, and details are provided in Table 2. A notable difference to the framing for the sheathed walls is that the top and bottom track size was increased to 600T150–68(50) to increase the buckling strength. Furthermore, two additional 19 mm (¾ in.) diameter holes were drilled at each end of the top track outside of the chord stud to accommodate additional bolts to help distribute loads in the track near the gusset plates to the loading frame.

Fig. 4.

Drawing of framing for the symmetric strap braced walls (1 ft = 30.5 cm).

A load was transferred from the braces to the framing through gusset plates using the same 5 × 19 mm (10 × ¾ in.) self-tapping, flat pan headed sheet metal screws used to connect the framing. Eight rows of four screws (32 screws total) were used to attach the strap to the gusset plate. For the gusset plate, a similar material to that used for the strap was selected with a strength of 345 MPa (50 ksi) and a thickness of 1.72 mm (0.068 in.). The required number of screws connecting the gusset plates to the chord and track flanges was determined using the conservative assumption that the vertically placed screws carry the vertical component of the expected brace force, and the horizontal screws carry the horizontal component.

Both sides of the wall were sheathed with 16 mm (⅝ in.) thick Type X gypsum boards to meet the 1-h fire-resistance rating requirements per UL Design No. U423 (UL 2017a). The vertical seams on the front and back of the wall were staggered by one-stud spacing. No drywall joint compound or paper tape was used on the unexposed side of the wall. On the fire exposed side, the seams were sealed using a base coat of drywall joint compound and paper tape, and the seams and fastener heads were then covered with a skim coat of drywall joint compound.

Holdowns

Uplift forces were transferred from the chord studs to the loading frame by holdown devices. Simpson strong-tie S/HD10S holdowns with an allowable tension capacity of 54.4 kN (12.2 kip) when used with back-to-back 1.37 mm (0.054 in.) thick studs were used (Simpson Strong-Tie 2017). Two holdowns were attached to the bottom of each chord stud with twenty-two 6 mm (¼ in.) diameter screws.

Test Setup

The test setup was informed by ASTM E2126 Standard Test Methods for Cyclic (Reversed) Load Test for Shear Resistance of Vertical Elements of the Lateral Force Resisting Systems for Buildings (ASTM 2011), but deviations were made to accommodate a burn compartment on a rolling platform. Specifically, this necessitated unobstructed access to the specimen, resulting in a 1.9 m (6 ft 4 in.) free span of the top-loading beam between the specimen and the actuator.

Mechanical

The test specimens were loaded mechanically by holding the base of the wall fixed and applying a prescribed in-plane deformation to the top of the wall (Fig. 5). The out-of-plane movement of the wall was limited by four steel guide frames placed perpendicular to the wall. The wall was attached to the bottom beam by two rows of 16 mm (⅝ in.) A325 structural bolts each pretensioned to 162 Nm (120 ft · lbs). Each holdown at the bottom of the chord studs was attached to the bottom beam using a 22 mm (⅞ in.) diameter A325 structural bolt pretensioned to 542 Nm (400 ft · lbs). The attachment of the top track to the loading beam was the same as at the bottom track; however, no holdowns were present.

Fig. 5.

Mechanical loading setup: (a) photograph of test rig; and (b) schematic drawing.

A mechanical load was applied to the specimen using a servohydraulically controlled actuator (MTS Series 201) with a load capacity of 240 kN (54 kips) in tension and 365 kN (82 kips) in compression. The maximum stroke of the actuator was ±381 mm (±15 in.).

Fire

The thermal load was provided by a natural gas diffusion burner in a three-sided compartment with interior dimensions 3.2 m long × 1.2 m deep × 2.9 m high (10.5 × 4 × 9.5 ft) that could be rolled against the specimen (Fig. 6). The inside of the compartment was lined with two layers of a thermal ceramic fiber blanket, each 25 mm (1 in.) thick. Two vents, 1.7 m tall × 1.2 m wide (5.5 × 4 ft) each, were located at the ends of the compartment. A window was made of fused silica glass in the back wall of the compartment to allow viewing of the test specimen.

Fig. 6.

Fire loading setup: (a) photograph of compartment; and (b) schematic drawing of cross-section with specimen in place.

The burner was 1.4 m long × 0.8 m wide (4.6 × 2.6 ft). The gas entered near the bottom of the burner and percolated up through a thermal ceramic fiber blanket. A positive displacement rotatory flowmeter, pressure gauge, and thermistor were used to measure the mass flow rate of the gas to the burner. The composition of the gas was determined using in-line gas chromatography (Bryant and Bundy 2019).

Testing Procedure

Cyclic Loading

ASTM E2126 Method C (the CUREE basic loading protocol) was used with a reference deformation Δ based on expected story drift ratios (SDR) for each wall at the peak load. This protocol is widely used in the United States to characterize the seismic resistance of the lateral force-resisting system for buildings. The reference deformations were 1.5% SDR and 2.5% SDR for the sheathed and braced walls, respectively. The loading procedure involves displacement cycles grouped in steps at incrementally increasing displacement levels. The loading history starts with six, small (0.05 · Δ), equal-amplitude cycles. Subsequently, each step consists of a primary cycle with the amplitude expressed as a fraction (percent) of the reference deformation Δ and subsequent trailing cycles with an amplitude of 75% of the primary cycle. The rate of displacement was 1.524 mm/s (0.06 in./s) and 2.54 mm/s (0.1 in./s) for the sheathed and braced walls, respectively. These rates minimized inertial influences while maintaining efficient test durations.

Drift levels 1 and 2 were selected based on test data (Ayhan et al. 2018) to achieve displacements intended to represent damage in earthquakes smaller or larger than the design earthquake, respectively. Drift level 1 was 0.45% and 0.50% SDR for the sheathed and braced walls, respectively. Drift level 2 was 1.50% and 1.75% SDR for the sheathed and braced walls, respectively. Before the fire, the trailing cycles at a given drift level were completed. After the fire, cycling was restarted at the primary cycle for that drift level, i.e., the prefire level was repeated.

Fire Loading

Fig. 7 shows the investigated fire scenarios, defined as temperature-time curves, described in this study as follows:

• Standard fire: 1-h of temperature-time exposure similar to ASTM E119.

• Severe fire: A postflashover fire of relatively long duration (35 min) and a peak upper layer gas temperature of 1,100°C.

• Mild fire: A postflashover fire of relatively short duration (15 min) and a peak upper layer gas temperature of 900°C.

Fig. 7.

Target temperature-time curves for the fires.

The Severe fire and Mild fires encompass a range of postflashover fire conditions that could occur in residential kitchens. A combination of statistical data, empirical formulations, and engineering judgment were used to define these scenarios. The duration of the fire was estimated using typical North American kitchen fire load densities and dimensions (Bwalya et al. 2011). The Standard fire does not fulfill all requirements in ASTM E119; however, the upper gas layer temperatures approximate those specified in the standard. A detailed discussion of the fire load development, including the data and equations used to develop these curves, is provided in Hoehler et al. (2019a).

Instrumentation

Fig. 8 shows the locations and orientations of the sensors used for mechanical loading tests. The load applied by the hydraulic actuator (ActuatorForce) was measured with an expanded uncertainty of ±2.0% of the full-scale output (FSO) and the displacement of the actuator (ActuatorDisp) with an expanded uncertainty of ±0.2% FSO (see the “Mechanical” section for actuator FSOs). A longitudinal displacement at the top of the wall specimen (Disp_Long) was measured using a string potentiometer with an expanded uncertainty of ±1.0% FSO, where the FSO was ±38.1 cm (±15 in.). The out-of-plane (Disp_Tran) and vertical (Disp_Vert) movement of the top-loading beam were measured using string potentiometers with an expanded uncertainty of ±1.0% FSO, where the FSO was ±12.7 cm (±5 in.). The uplift (Disp_Uplift_N/S) and slip (Disp_Slip) of the wall relative to the bottom beam were measured using linear potentiometers with an expanded uncertainty of ±1.1% of the reading (RD).

Fig. 8.

Sensor locations on specimen and loading frame during mechanical loading.

The instrumentation in the fire compartment is shown in Fig. 9. There were two thermocouple (TC) arrays, each with five Inconel-sheathed Type K grounded-junction thermocouples with an expanded uncertainty of ±2.8% RD. One additional Inconel-sheathed Type K grounded junction thermocouple was placed 2.5 cm (1 in.) from the ceiling at the center of the compartment. Nine plate thermocouples were placed 10 cm (4 in.) from the back wall of the compartment facing the test specimen. Each plate thermocouple probe plate was 100 × 100 mm (4 × 4 in.). A manufacturer-reported standard uncertainty was not available. Details about the thermal instrumentation in the wall specimens are provided in Hoehler et al. (2019a).

Fig. 9.

Compartment sensors for gas fueled fires: (a) longitudinal view; and (b) transverse view (cm).

Results and Discussion

Fire Exposure

The mean (average of the three top thermocouples on the array) upper layer gas temperatures achieved in the fire tests and a photograph of a typical test are shown in Fig. 10. The temperatures generally tracked the target fire curves (Fig. 7), with exceptions being that the rate of temperature decrease after the fire was extinguished for the Severe and Mild fires was more rapid than the targets, and the temperature rise rate from 20 to 60 min in the Standard fire was slightly lower than in an ASTM E119 test. For a detailed discussion of the temperature variation in the burn compartment, as well as a measured spatial variation of the incident heat flux to the wall specimens, the reader is referred to Andres et al. (forthcoming).

Fig. 10.

(a) Mean values (lines) and standard deviations (shaded areas) of the measured upper layer temperatures for the Standard, Severe, and Mild fires; and (b) photograph of the back of the compartment during the fire test.

Characterization of the intensity of fire as experienced by structural members in a building depends on many factors. In this study, we compared the areas under the σ · T⁴ curves, where σ is the Stefan-Boltzmann constant, and T is the mean upper layer gas temperature in the compartment. This provided an approximation of the integrated incident heat flux to the wall specimens. Per this metric, the Mild fire is roughly half as intense as the Standard fire, and the Severe fire is roughly twice as intense.

Gypsum-Sheet Steel Composite Panel Sheathed Wall Performance

Fig. 11 plots the applied actuator (lateral) force versus the top-of-wall drift during mechanical loading of the gypsum-sheet steel composite panel sheathed walls for a representative case for each investigated fire loading. The peak forces achieved at each step of the loading protocol for positive (circle) and negative (square) actuator excursions are indicated on the plots. The peak force versus story drift ratio envelopes for all of the walls are shown in Fig. 12; only the positive excursions because the behavior is approximately symmetric with respect to the loading direction.

Fig. 11.

Lateral force versus drift during mechanical loading of gypsum-sheet steel composite panel sheathed walls: (a) cycling without fire (SB01); (b) cycling after Mild fire (SB03); (c) cycling after Standard fire (SB04); and (d) cycling after Severe fire (SB02).

Fig. 12.

Lateral load versus story drift ratio (positive excursions) during mechanical loading of gypsum-sheet steel composite panel sheathed walls.

This wall system exhibited increasingly diminished postfire capacity with increasing fire severity. The reduction in the peak force capacity (based on positive excursions) was 10%–23% (test repeat variation), 58%, and 68% for the Mild, Standard, and Severe fire, respectively. The primary failure mode during cycling at ambient temperature (without fire) was the failure of the shear panel connections. Connection failures were a combination of fastener failure and edge tear out from the sheet steel or pull-through of the fastener head through the sheet steel [Fig. 13(a)]. The Mild fire severely degraded the mechanical strength of the gypsum on the fire-exposed side of the wall and locally degraded the adhesive bonding of the gypsum to the sheet steel on the unexposed side of the wall [Fig. 13(b)]. Damage to the adhesive between the sheet steel and the gypsum reduces the stiffness of the panels out-of-plane, in effect, changing the specimen from a composite panel sheathed wall to a plain sheet steel shear wall with an accompanying capacity reduction. For more examples of failure mode transitions of these panels following a fire see Hoehler et al. (2017). The smaller force reduction in the Mild fire test repeat SB03R compared to SB03 is due to the lower temperature (and, consequently, less damage to the adhesive) experienced by the sheet steel in SB03R—approximately 175°C versus 250°C near the top of the wall. The Standard fire further degraded the adhesive, and more widespread buckling of the sheet steel occurred [Fig. 13(c)]. In the Severe fire, the fire damaged nearly all the adhesive on the composite panels, oxidized several screws along the top the wall, and burned through the sheet steel at a few locations [Fig. 13(d)]. Nevertheless, the applied load redistributed, and the system continued to resist the lateral force. The loss of galvanization on the sheet steel (dull versus shiny) in Fig. 13 shows the extent and spatial distribution of the heating of the sheet steel.

Fig. 13.

Photographs of gypsum-sheet steel composite panel sheathed walls after: (a) cycling without fire (SB01, detail of failure mode); (b) cycling after Mild fire (SB03); (c) cycling after Standard fire (SB04); and (d) cycling after Severe fire (SB02).

Oriented Strand Board Sheathed Wall Performance

Fig. 14 plots the lateral load versus drift during mechanical loading for representative cases of OSB sheathed walls with no predamage prior to the fire. For ambient temperature cycling, the failure mode was pull-through of the screw heads through the OSB board with some cases of edge breakout of the screw from the OSB [Fig. 15(a)]. The Mild fire effectively eliminated the gypsum on the fire-exposed side of the wall, caused surface charring on some OSB panels [Fig. 15(b)], and reduced the residual lateral capacity by 26%. Both the Standard fire and Severe fire consumed the OSB [Figs. 15(c and d)] and damaged the interior framing studs. The reduction of the lateral load capacity in both cases was nearly 100%.

Fig. 14.

Lateral force versus drift during mechanical loading of OSB sheathed walls: (a) cycling without fire (OSB01R); (b) cycling after Mild fire (OSB03R); (c) cycling after Standard fire (OSB04); and (d) cycling after Severe fire (OSB02).

Fig. 15.

Photographs of OSB sheathed walls after: (a) cycling without fire (OSB01R, detail of failure mode); (b) cycling after Mild fire (OSB03); (c) cycling after Standard fire (OSB04); and (d) cycling after Severe fire (OSB02).

The peak force versus story drift ratio envelopes for all tests are shown in Fig. 16. To check whether the 26% force reduction following the Mild fire could be attributed to the loss of the gypsum board on the fire-exposed side of the wall, specimen OSB01NG was constructed without a gypsum board that could carry a lateral load and subjected to load cycling without fire. An approximately 25% reduction in force capacity was observed, which suggests that the capacity reduction in the case of the Mild fire was likely due to the damage to the nonstructural gypsum boards on the fire-exposed side of the wall.

Fig. 16.

Lateral load versus story drift ratio (positive excursions) during mechanical loading of OSB sheathed walls.

Cycling the OSB sheathed wall to 0.45% story drift prior to the fire resulted in only minor damage to the skim coat on the gypsum board joints and had no significant effect on the subsequent fire or postfire cyclic performance [compare Figs. 14(b)–17(a) (solid line = cycling prior to fire; dashed line = cycling after fire)]. The fire still reduced the postfire capacity as in the case of undamaged walls, but the reduction is not worsened by this level of predamage. This is also illustrated by Fig. 16 in which the postfire response of OSB06 (Mild fire with 0.45% SDR predamage) closely follows that of OSB03 and OSB03R (Mild fire without predamage).

Fig. 17.

Lateral load versus drift during mechanical loading of OSB sheathed walls: (a) cycling to 0.45% drift before Mild fire; and (b) cycling to 1.5% drift before Mild fire.

Cycling to 1.5% story drift prior to the fire tore the tape along the joints, and one of the OSB panels ignited during the Mild fire. The fire was suppressed 10 min after the burner was extinguished. The joint damage increased the interior stud temperature and degraded the postfire capacity of the wall [compare Figs. 14(b)–17(b)]. This is perhaps better illustrated by Fig. 16 in which the postfire response of OSB05 (Mild fire with 1.5% SDR predamage) drops below that of OSB03 and OSB03R (Mild fire without predamage). The capacity would have gone to zero had the fire not been suppressed after 10 min.

Steel Strap Braced Wall Performance

Fig. 18 plots the lateral load versus drift during mechanical loading of steel strap braced walls with no predamage prior to the fire for representative cases of steel strap braced walls. The baseline hysteretic behavior Fig. 18(a) (ambient temperature) shows a peak near the maximum load followed by a long plateau as the steel straps yielded. This initial peak is caused by the contribution of the gypsum boards. The ultimate failure mode was a rupture of the straps at the gusset plate connections [Fig. 19(a)] and crippling of the chord stud just above the holdown after significant yielding of the straps (>6% story drift). The Mild fire effectively eliminated the gypsum on the fire-exposed side of the wall and reduced the residual lateral capacity by 15% [Fig. 18(b)], but the failure mode was the same as without fire [Fig. 19(b)]. The reduction of 15% load capacity appears consistent with the loss of gypsum on the fire-exposed side of the wall. The response during the Standard fire was similar to that during the Mild fire; however, the gypsum paper on the inside of the wall on the unexposed side was blackened, indicating higher wall temperatures. The reduction to the residual capacity [Fig. 18(c)] and failure mode [Fig. 19(c)] were similar to the Mild fire results. The Severe fire burned through the gypsum on both sides of the wall at the top center of the wall (slightly shifted south) toward the end of the heating phase. During subsequent cyclic loading, when cycling in the direction that put the oxidized straps in tension [Fig. 19(d)], the wall had limited residual load capacity [Fig. 18(d), negative], while in the opposite loading direction close to the full ambient postyielding, load capacity was reached [Fig. 18(d), positive]. Interestingly, the postfire ductility in this direction increased significantly [note axes scale change in Fig. 18(d)], and there was a more pronounced postyielding hardening behavior. This may be due to the annealing of the cold-formed steel strap during the fire, but further study is required. This asymmetry in the response can also be seen in Fig. 20, which plots the peak force versus story drift ratio envelopes for all tests for load cycling in both directions.

Fig. 18.

Lateral load versus drift during mechanical loading of steel strap braced walls: (a) cycling without fire (S01R); (b) cycling after Mild fire (S03); (c) cycling after Standard fire (S04); and (d) cycling after Severe fire (S02).

Fig. 19.

Photographs of steel strap braced walls after: (a) cycling without fire (S01R); (b) cycling after Mild fire (S03); (c) cycling after Standard fire (S04); and (d) cycling after Severe fire (S02).

Fig. 20.

Lateral load versus story drift ratio during mechanical loading of steel strap braced walls.

Cycling the steel strap braced wall to a 0.5% story drift prior to the fire resulted in only minor damage to the skim coat on the gypsum board joints and had no significant effect on the subsequent fire or postfire cyclic performance [compare Figs. 18(b)–21(a)]. Cycling to 1.75% story drift prior to the fire tore the tape along the joints, and the paper on the back of the unexposed gypsum board was burned off near the top center of the wall. The joint damage increased the interior stud temperature but did not affect the postfire load capacity of the wall; however, the straps ruptured at a slightly smaller drift [compare Figs. 18(b)–21(b)].

Fig. 21.

Lateral load versus drift during mechanical loading of steel strap braced walls: (a) cycling to 0.5% drift before Mild fire; and (b) cycling to 1.75% drift before Mild fire.

Conclusions

This paper investigated the interplay between the thermal (fire) and the mechanical (cyclic) response of cold-formed steel lateral force-resisting systems with a focus on the mechanical response of the walls; the thermal response is presented in a companion paper (Andres et al., forthcoming). The investigated fire loads represented varying fire severities, including 1 h of temperature-time exposure similar to ASTM E119 (Standard fire), a postflashover fire of a relatively long duration (35 min) and a peak upper layer gas temperature of 1,100°C (Severe fire), and a postflashover fire of a relatively short duration (15 min) and a peak upper layer gas temperature of 900°C (Mild fire). The results are a limited set of data, and the findings presented should not be extrapolated beyond the conditions tested. It is important to note that the absence of a vertical load in this study may affect the relative capacity reductions reported for the various wall systems.

The gypsum-sheet steel composite panel sheathing exhibited an increasingly reduced postfire capacity with increasing fire severity. The maximum reduction in the peak force capacity was 23%, 58%, and 68% for the Mild, Standard, and Severe fire, respectively. A nontrivial portion of this force reduction (>10%) was attributed to the severely degraded mechanical strength of the gypsum on the fire-exposed side of the wall, which no longer contributed to the load capacity after the fire. The remainder of the capacity reduction was due to damage to the adhesive between the sheet steel and he gypsum in the composite panels. This reduced the out-of-plane stiffness of the panels, changing the specimen from a composite panel sheathed wall to a plain sheet steel shear wall with an accompanying capacity reduction. However, the sheet steel helped to prevent flame spread out of the compartment and allowed the walls to maintain lateral load capacity under reversed load cycling following the most severe fire investigated. Phase 1 tests reported previously (Hoehler et al. 2017) suggest that the postfire mechanical response of the composite panel system is insensitive to cyclic damage prior to the fire.

The strap braced walls were the most ductile of the investigated systems, achieving story drift ratios over 6%, and were insensitive to the lower investigated levels of thermal loading. The maximum reduction in the peak force capacity was 15% and 17% for the Mild fire and Standard fire, respectively. The reduced capacity is attributed to the degraded mechanical strength of the gypsum boards following the fire. However, in the case of the Severe fire, the residual lateral load capacity was reduced to essentially zero. The postfire mechanical response of the strap braced wall was insensitive to cyclic damage prior to the Mild fire.

For this limited set of data, the oriented strand board sheathed walls demonstrated a significant impact from the fire. Both the Standard fire and Severe fires caused the OSB to ignite, resulting in a near-total loss of residual lateral capacity. Moreover, cycling to a 1.5% drift, as might occur in a major earthquake, prior to the fire allowed the Mild fire to penetrate the wall and ignite the OSB.