The objective of this research study was twofold: (a) to determine at what beam depth specially graded 302-24 (2) tension laminations are required to attain a 2,400 lb/sq. in. design stress; (b) to determine the appropriate design stress for specification-type beams with the regular laminating grades of L1 Douglas Fir-Larch (DF-L) or No. 1D Southern Pine (SP) as the tension laminations. The 302-24 grade of lumber required for tension laminations on 24F4 deep beams is a limited resource and is not as readily available as L1 or No. 1D lumber now used as the tension lamination on 24F shallow beams-thus necessitating objective (b). The 180 glued-laminated beams evaluated were of near-minimum quality; each selected tension lamination had a near-maximum-size, strength-reducing characteristics positioned in the most highly stressed region of the test beams. (Author).

In bridge applications, it is often necessary to minimize the depth of the bridge structure to provide for the required hydraulic opening or reduce the volume of approach fill. For bridges that utilize structural glued-laminated (glulam) timber beams as stringers, reinforcement using thin strips of pultruded E-glass-fiber-reinforced plastic (GFRP) composites may permit reduced depth, because the reinforcement has the potential to increase stiffness and strength. This study is part of an overall effort aimed at evaluating the potential for commercial production of glulam-GFRP beams in current wood laminating plants and a wood adhesive compatible with existing equipment. Twelve Yellow-Poplar glulam GFRP beams were commercially manufactured, and their performance was evaluated. The GFRP panels were bonded to the wood with a resorcinol formaldehyde adhesive to provide the reinforcement. The simplicity of the process used to manufacture the test beams indicates that the commercial production of glulam-GFRP beams is feasible. Increases of 18 percent in stiffness and 26 percent in strength were achieved by adding 3 percent of GFRP by volume. The bending strength values of the beams predicted by the ASTM D3737 procedure correlate well with the experimental values. However, the observed delamination of the reinforcement indicates that improved bonding strength of wood--GFRP interfaces is needed. Results of this study will be useful to manufacturers interested in improving the performance of glulam timber beams.

A previous paper presented a reliability-based model to predict the strength of glued-laminated timber beams at both room temperature and during fire exposure. This Monte Carlo simulation procedure generates strength and fire endurance (time-to-failure, TTF) data for glued-laminated beams that allow assessment of mean strength and TTF as well as their variability. This paper reports an effort to validate model predictive capability through an independently fabricated set of 21 glued-laminated beams. Based upon the available data for the model input parameters on lumber strength and stiffness, finger-joint strength, and length of laminating lumber between sequential finger joints, the model of beam strength appears acceptable and possibly slightly conservative. Refinements in the beam strength model allow its use for predicting fire endurance. In this case, the fire endurance is measured by the TTF and is defined as the time the beam will support its design load while subjected to fire. The residual strength of the beam is analytically calculated by removing the char layer, plus a finite thickness of weakened wood, from the beam cross section as fire exposure time increases. Employing the input parameters for values of finger-joint strength and lamination grades of Douglas-fir, the fire endurance TTF was analyzed for a 5.12- by 16.50-inch 11-lamination Douglas Fir-Larch beam (24F-V4) carrying full allowable uniform load (47.7 lb/in.). (Three-sided fire exposure was assumed; however, four-sided exposure can also be accommodated.) A simulated random fabrication and analysis of the TTF under fire exposure for 100 beams was performed. The mean TTF was estimated as 35.2 minutes with a coefficient of variation of 13.7 percent. Lateral torsional buckling was never the cause of failure in any of the simulations. The results compared well (within a 65 pct confidence band) with the observations and predictions for timber beams reported by sources in other countries. A simulation for a single glulam beam test in cooperation with the National Forest Products Association was also conducted which predicted the result exactly.