fire resistance of unprotected cross-laminated timber wall...

Fire Resistance of Unprotected Cross-Laminated Timber Wall Assemblies Made in the U.S.A.

Seung hyun Hong1, Lech Muszyński1, Rakesh Gupta1, Brent Pickett2Oregon State University 1Department of Wood Science and Engineering, 2Western Fire CenterInc.

Abstract

Fire hazards are one of the major concerns for wooden constructions. U.S. Building Code restricts wooden buildings to four stories, despite data on the fire resistance of cross-laminated timber (CLT), generated in Europe, Japan, Canada, and in the US, and a growing body of evidence that CLT is an acceptable materials for tall construction. The major reason for code restrictions in the U.S. is the scarcity of full-scale tests conducted on “made in the US” structural CLT. Therefore, the objective of this project was reducing this barrier by testing full-scale unprotected CLT according to the ASTM E119 fire resistance on CLT wall assemblies produced by US manufacturers. The assemblies represented two species groups (SPF and DF-L), and two adhesive systems (PUR and MF). All assemblies were conducted in a loaded condition. Two walls (DF-L/PUR and DF/MUF) assemblies met ASTM E119 standard qualifying criteria for 2-hours Time-Temperature Area, but SPF/PUR wall assembly passed 101 min. The statistical significant difference was observed depending on the adhesives type. MUF held char layer more efficient than PUR. The major driving force of char rate was furnace temperature and adhesive types than wood species. The unprotected half-lap joints provided an adequate barrier against transmission of hot gases and flames through the assemblies before char depth reached to half of total CLT thickness.

KEYWORDS: Fire resistance; cross-laminated timber; CLT; walls; half-lap joints; PUR; MF; SPF; Douglas-fir;

Introduction

Historically, fire hazard situation is one of the major concerns for 4-stories cap on wooden constructions, despite a predictable rates of a solid wood burning (e.g. about 0.63 mm/min or 3.81 cm/hour for spruce) in standard fire tests [4, 5]. It causes tall wooden building code restrictions in the U.S. despite data on the fire resistance of cross-laminated timber (CLT), which are already tested in Europe, Japan, and Canada, and proved that CLT is acceptable to use materials of tall construction against fire hazard. Another reason for its building code restrictions in the U.S. is the lack of full-scale tests which are conducted using “made in the US” structural CLT [2].

Materials and Methods

The fire performance of three unprotected CLT assemblies was determined in full-scale fire tests following the ASTM E119 test procedures for fire tests of buildings and construction materials[1], at the Western Fire Center (WFC) in Kelso, Washington. All CLT assemblies were structural grade 5-ply Layers were made up of visual grade #2 or better [3] 35 mm x 140 mm laminations. The assemblies differed by species of the laminations and adhesive systems used to non-edged glue, and represented two species groups (SPF and combination of Pseudotsuga menziesii and Larix spp.), and two adhesive systems (polyurethane HBE adhesive and melamine formaldehyde) The origins and material composition of the assemblies are summarized in Table 1. The diagrams showing the external dimensions, the connector spacing in the wall assemblies in figure 1a.

The assemblies were instrumented with 9 type K ceramic insulated 0.51 mm thermocouples placed on the unexposed surface covered with a 152 mm x 152 mm ceramic fiber pads. Each group was placed in a square configuration 102 mm apart. These thermocouples are marked as grey dots in figure 1a. In addition, three groups of two thermocouples were positioned along the lap joint: one the pair would be located at the joint surface groove, the other 5 cm (2 in) away from the joint surface groove at the half-lap. These thermocouples are marked as orange dots in figure 1a. Diagrams illustrating through-the-thickness positions of the thermocouples are explained in figure 1b.

Out of plane deformation of the wall panels was monitored with four LVDT sensors distributed along the joint in walls. The in-plane deflection of the wall assemblies was monitored with one LVDT positioned in the bottom left and right corner of the assembly frame.

The progress of the char limit layer in the assemblies was initially assumed from threshold temperature of 300º [4] recorded with the embedded thermocouples, and later confirmed by the measurement of the final char depth at the conclusion of the tests.

Table 1: Description of the test assemblies

Assembly ID SourceWood species(lumber grade) PRG320 CLT grade

Wall 1 SPF/PUR Smartlam Spruce-pine-fir (SPF, #2 btr) V2

Wall 2 DF-L/PUR Smartlam Doug fir-larch (DF-L, #2 btr) V1

Wall 3 DF-L/MUF DR Johnson Doug fir-larch (DF-L, #2 btr) V1

102 mm

(4 in)

175

mm

(6

7/8

in)

JGJH

BL1BL2

BL3BL4

S

(b)

3048

mm

(10

ft)

1575 mm (5 ft 2 in) 1473 mm (4 ft 10 in)3048 mm (10 ft)

(a)

figure 1: Wall assembly diagrams including the position of connectors and in-plane thermocouple position (a), and the placement of the thermocouples through the thickness of the panel and in the lap joint (b).

Results (cont’d)

Wall1 (SPF PUR) Wall 2 (DFL PUR) Wall 3 (DF MUF)


Wall1 (SPF PUR) Wall 2 (DF-L PUR) Wall 3 (DF-L MUF)

Walls

Assembly LayersMean char rate

mm/minWall 1SPFPUR

B1 0.59B2 1.30

cumulative 0.81Wall 2DF-LPUR

B1 0.55B2 0.82

cumulative 0.65Wall 3DF-LMF

B1 0.49B2 0.93

cumulative 0.58

Walls

Figure 3: Development of bond line temperatures (blue lines, left scale) and the deflection in floors and walls (black line, right scale).

Figure 4: Development of bond line temperatures depending on distance from exposed surface

Figure 7: The effect of softening of the PUR bonds.

Table 2: Average char rate of walls assemblies

Conclusions

References

Two of the three wall assemblies produced by American manufacturers passed the 2-hour fire resistance test following ASTM E119 standard procedure. One assembly (SPF/PUR) passed 101 min mark before the diaphragm was breached.

Statistical significant difference was observed depending on adhesives type. MUF held char layer more efficient than PUR. Thus, the amplitude of thermocouples plot of MUF CLT was smaller than PUR. It was caused by increasing flame area by falling char caused increasing temperature in the furnace.

The char rate of SPF PUR second layer was increasing, and its value was the highest than others because of increasing furnace temperature. Therefore, the major driving force of char rate was furnace temperature and adhesive types than wood species.

The heat energy exposed time caused softening of the PUR bonds because the temperature of softening of the PUR bonds was decreasing depending on distance from exposed surface.

The unprotected half-lap joints during the 2-hour standard test exposure provided adequate barrier against transmission of hot gases and flames through the assemblies before char depth reached to half of total CLT thickness.

[1] ASTM International WC, PA.ASTM E119-16a Standard Test Methods for Fire Tests of Building Construction and Materials. by ASTM International.R. Mahnken. A Newton-multigrid algorithm for elasto-plastic/viscoplastic problems. Comp. Mechs., 15:408-425, 1995.

[2] Barber D. Fire Safe Design of Exposed Timber in Mass Wood Buildings. . Mass Timber (CLT) Research Workshop, 2015.[3] West Coast Lumber Inspection Bureau W.Grading rules for Western Lumber. by, Portland, Oregon, 97281.[4] White; RH, Dietenberger MA. Fire Safety of Wood Construction. In. Wood Handbook: Wood as an engineering material. Madison, WI, U.S. Department of Agriculture, Forest Service, Forest Product Laboratory, 2010, pp. 18.1-18.22.[5] White RH, Nordheim EV, Charring rate of wood for ASTM E 119 exposure, Fire Technology, 1992;28: 5-30.

figure 2: Charred surfaces of the three test floor assemblies after the 2 hour exposure tests.

Result


BL1 BL2 BL3 BL4 S

0:15

0:30

0:45

1:00

1:15

1:30

1:45

2:00

Char LimitBL1 BL2 BL3 BL4 SBL1 BL2 BL3 BL4 S

Figure 6: Temperature distributions as measured by thermocouples embedded in the four bond lines

Figure 5: Char rate diagrams for the three test assemblies.


fire resistance of unprotected cross-laminated timber wall...

Documents