4. Observations and Conclusions

4.1 Observations

The beams tested in this study had a wide range of observed composite efficiency values. Much of this observed variation in efficiency was explained in terms of ineffective notch connections or material failures. For example, deep beam 1 had an average composite efficiency of 59 percent for four repeated load tests. However, in load test DB1.1 a much lower composite efficiency (42%) was observed due to the notch anchor connections not having been tightened before the loading was initiated. As another example, wide beam 2 had observed efficiencies of 47 percent and 34 percent, respectively, for load tests WB2.1 and WB2.2. However, WB2.2 was conducted after an undetected flexural failure in the wood layer of wide beam 2 in the prior loading.

Thus many of the composite efficiency results of this study were strongly influenced by some type of ineffective connection or material failure. Because the presence of ineffective connections and material failures are localized effects, albeit very important ones, they do not allow the overall performance of the wood-concrete beam system to be studied. Thus it may be better to assess the gross wood-concrete beam system performance based on composite efficiency results for only those beam specimens where ineffective connections and material failures were not observed.

Table 4.1 lists the adjusted average composite efficiency results for all of the load tests. The adjusted average composite efficiencies were computed after excluding all of the load test results where known or suspected connection or material failures occurred. Note that the results of wide beams 3 and 4 were not adjusted because no ineffective connector or material failure effects were observed for those tests. The load tests excluded from the adjusted average composite efficiency computation were omitted for the following reasons; DB1.1 had an ineffective (not tightened) notch; DB1.5 was a failure load test; DB1.6 was a post failure load test; DB2.3 was a failure load test; DB2.4 was a post failure load test; WB1.2 had an ineffective (failed) notch; and WB2.2 had a flexural failure.

Table 4.1 Adjusted Average Composite Efficiency Results

The composite efficiency results in Table 4.1, although not appropriate for a statistical evaluation, show that the Deep Beam specimens had higher average composite efficiency than the Wide Beam specimens. There are many potential explanations for this observation as discussed in earlier sections. Another possible explanation is that the Wide Beams were significantly more difficult to construct than the Deep Beams. As a result the Wide Beams had many more construction flaws (e.g. poor consolidation of the notch) than the Deep Beams. Thus the Deep Beam and Wide Beam results in Table 4.1 may not be directly comparable considering that there was a large difference in the quality of the two beam types. Table 4.2 compares the geometries of the various specimens and other observed efficiencies.

Table 4.2 Compilation of Specimen Geometries and Test Results

As with the deep beam specimens, creep is an issue for wide beam specimens. Wide beam 1 and wide beam 2 (width = 96.48") were in place and subject to creep for only about two months from the time of casting to the time of load testing. Connections in wide beam 1 were also not retightened before the first load test. Wide beam 3 and wide beam 4 (width = 60.98") were in place and subject to creep for about 27 months and 28 months, respectively, from the time of casting to the time of loading. Despite retightening the connections before load testing, the long term creep is likely the dominant factor in the much lower composite efficiencies for wide beam 3 and wide beam 4 compared to wide beam 1 and wide beam 2. The difference in width may have been a factor, but there is no data basis to distinguish the effect of that variable. An additional factor in the results is the wide beam had a longer span (20') than the deep beam (12').

Specimens wide beam 1 and wide beam 2 had the same geometries and initial creep time. Thus the type of connector (glue vs. mechanical) was the only difference, and the specimens had similar efficiencies. Specimens wide beam 3 and wide beam 4 had the same geometries but were about two thirds the width of wide beam 1 and 2. They were also subject to much longer creep (27-28 months vs. 2 months). The efficiency of wide beams 3 and 4 were much lower than for wide beams 1 and 2, attributed in large part to the much longer creep time. However, the mechanically connected wide beam 4 had about 60 percent of the efficiency of glue connected wide beam 3. So the connector quality may have been a secondary contributor.

4.2 Conclusions

A very small sample of beams of different geometries was involved in the pilot test program, so statistical assessment is not meaningful. It is evident that creep of the specimens (all un-shored, except for the wood layer itself holding the concrete layer) was a significant phenomenon. For bridge applications this suggests shoring may be needed unless significant camber can be incorporated to offset the effect. In residential and commercial application, lightweight concrete could be considered.

Several loadings weinvolved, and when questionable or compromised loadings are eliminated the findings are not significantly altered from those of the overall set. The deep beam results suggest that the 1-year creep time may not have been a significant factor as reasonably high efficiencies resulted. The connections in these beams were retightened before loading, which may have contributed to the outcome. But without a test of specimens with no creep exposure, a definitive degree of effect of creep is not distinguishable. They may have had even higher efficiencies if creep had been prevented. Thus, high efficiencies are possible as evidenced by these beams even with one year of creep. The wide beams had definite creep effects compared to the deep beams. Because the wide beams were exposed to creep more than twice as long as the deep beams, the length of time was apparently a major factor. No definitive observation is possible for the comparison of glued connectors vs. mechanical connectors, as no consitent difference in efficiency was evident.

The results of this study show the complex behavior of composite wood-concrete beams. Composite efficiency is affected by many factors such as concrete consolidation in the notch, knots or other defects in the timber layer, the effectiveness of the notch anchor, and many others. Also, creep effects may have a strong effect on the composite efficiency of large wood-concrete composites. Thus it is recommended that the influence of all of these different factors on composite efficiency be studied in depth with a large number of specimens so as to allow parameter studies.