Acknowledgements

The authors gratefully acknowledge the Mountain Plains Consortium for generously funding this project. Colorado State University provided the use of the facilities of the Structural Engineering Laboratory (SEL) at its Engineering Research Center. The authors thank Amber Appel, Misty Butler, Ryan Fast, and Martin Wieligmann for their assistance in the preparation of the specimens and conduct of the load tests. Thanks especially to TJ Schilling for his dedication to continually assisting in this project and in many other MPC supported projects either ongoing or completed ongoing at the SEL. Horacio Garza, Jr. provided shop services and invaluable advice related the preparation of the test specimens and test set up.

Disclaimer

This paper reflects the views of the authors, who are responsible for the facts and the accuracy of the information presented. This document is disseminated under the sponsorship of the Department of Transportation, University Transportation Centers Program, in the interest of information exchange. The United States Government assumes no liability for the contents or use thereof.

Abstract

Research at Colorado State University in wood-concrete layered composites for structural applications has been in progress for more than a decade. Key features of the research are experimental and analytical work to analyze notched shear key interlayer connections. These connections are being used to achieve composite action by transferring interlayer forces by bearing in a notch cut out of one material layer. This report documents a series of experimental load tests on thick-layered wood-concrete composite beams. The research of thick-layered beams is a precursor to possible future research of thick-layered wood-concrete bridge deck systems. Experimental results are given for two beams referred to as the deep beam specimens (12' long x 13.5" deep x 7.25" wide), and four beams referred to as the wide beam specimens (20' long x 15"deep x 2' and 3' wide). The results of the load tests are presented in terms of the measured beam deflections and the horizontal slip between the wood and concrete layers. Some load tests were eliminated due to various physical reasons which compromised them. Results for the loadings show that a low to moderate percentage of potential composite action was achieved. For the deep beam specimens the degree of composite action observed ranged from 64.22 percent to 84.22 percent of that theoretically possible if the layers were fully bonded to each other. The average value was 72.31 percent. These results are similar to those found by past researchers for narrow, shallow beams. Before the first loading, the deep beams were free to creep for approximately one year. This high average efficiency suggests creep effect did not significantly affect the deep beams. For the wide beam specimens the degree of composite action observed ranged from 11.81 percent to 46.86 percent and the average value was 27.35 percent. This is an unexpected low level. Before the first loading, two of the wide beams were free to creep for about two months. Two other wide beams were free to creep for about 28 months. The latter exhibited the lowest efficiencies. It appears that creep significantly affected the wide beams. Because creep itself was not measured, the effect of increased beam width versus creep could not be distinguished.

Executive Summary

Our nation's bridges are predominately on rural secondary roads and are critical to the movement of agricultural and mineral products. Low tax bases make saving every possible bridge repair and replacement dollar important. About 20 percent of the bridges use timber decking on either steel or wood stringers. Typically, timber decks in older bridges are the first component to wear and exhibit reduced load capacity. The usual consequence is the deck is completely replaced with a new one. Otherwise, the bridge using the deck would have to be either posted for lower load limits or closed.

In Europe, an innovative concrete overlay technique was recently developed to strengthen solid wood floors in apartment and office buildings. Interlayer shear transfer, needed to affect composite behavior, is achieved by a notched shear key. Wood-to-concrete bearing and shear in the notch materials affect the interlayer force transfer. A vertical anchor is glued into the wood and can be turned to tighten the concrete-to-wood bearing surfaces after hydration of the concrete has taken place. The shear key concept has potential as either a retrofit strengthening technique for old timber bridge decks or as new construction. However, because of much higher loads involved in bridges, the mixed material deck system must be much deeper than in residential and commercial floor systems. Consequently, in this study a series of load tests were conducted on large composite beam specimens as a low-cost approach in advance of possible load tests of full decks.

Two specimens (termed "deep beams") involved beams with a span and depth (overall and for each material) consistent with what might be required in a short bridge deck. One specimen used a glued anchor connection such as developed in Europe. The second specimen used an alternative mechanical anchor connection which passed through the wood and was anchored against the bottom of the wood member. A second set of specimens (termed "wide beams") was intended to examine the effect of increased span and member width on the load resistance on layered deep beams. Two specimens used a glued anchor connection and two used the mechanical anchor connection.

The effectiveness (efficiency) of the specimens was examined by comparing the measured displacements with those based on each of two theoretical bounds. One bound was with the two layers fully bonded to each other, the other bound was with complete absence of interlayer bond. Efficiency is an indicator of how close the behavior is to that of the fully bonded bound. For the deep beams with glued connections the average efficiency was 64.91 percent which is consistent with past studies of shallower beams. For the deep beams with mechanical connections the average efficiency was 83.55 percent. Thus, in these specimens, the mechanical connection performed at a much higher efficiency as compared to the glued connections and past studies of shallower members. Overall their average efficiency was 72.31 percent. For the wide beams with a glued anchor connection the average efficiency was 28.68 percent. For the wide beams with mechanical connections the average efficiency was 35.24 percent. Thus, the wide beams were not as noticeably different for each connection method. However they both exhibited a significant drop in efficiency (overall average efficiency was 27.35%) compared to the Deep Beams. Significant creep deflection likely occurred in the Wide Beam specimens and likely influenced the results for those specimens. Creep deflection was not measured, thus could not be considered quantitatively in this study. Consistency of construction and possible flaws (such as poor concrete consolidation at some locations) were other intangible factors in the low results for the wide beam specimens.

The results suggest that with proper construction, and control of creep, for example by showing the mixed wood-concrete construction for deep beams, is feasible. This is evidenced by the 72.31 percent efficiency of the deep beams. However, the tests were limited to ramp loadings and the long-term effect of repeated loading and possible fatigue of the connection should be examined. The wide beam specimens were compromised by the effects of creep and other physical effects. Thus they should be redone and no dependable conclusion is possible. Overall, quantitative studies of the creep phenomenon should be undertaken.