3. Results of the Test Program
3.1 Calculation of Efficiency
For each cycle of each load test the following results were either recorded or calculated: the maximum measured deflection at mid-span (Δ), the applied load at mid-span corresponding to the maximum deflection, the ratio of clear span to maximum mid-span deflection, the theoretical non-composite deflection (ΔNC), the theoretical fully-composite deflection (ΔFC), the available composite action, the composite efficiency, and the composite action observed. A complete presentation of the results is given in Appendix A. A group of selected results is described in the subsequent sections.
The theoretical non-composite deflection was calculated using a composite flexural stiffness of the beam equal to the sum of the EI value for the wood and concrete layers, where EI is the product of the modulus of elasticity (N/cm2) and the moment of inertia (cm4). The moment of inertia used to compute EI was simply calculated from the dimensions of the sections. Fully composite deflection values were computed by a transformed sections approach. In this approach the concrete layer was transformed into an equivalently stiff wood layer and the beam was analyzed as a one-layer homogenous wood beam. The analysis uses the E of the wood layer and the moment of inertia of the transformed section. For the transformed concrete layer the depth was the same as the original concrete layer, but the transformed width was equal to the original concrete width scaled by the modular ratio of the E of the concrete to the E of the wood.
It should be noted, however, that computed fully composite deflections are smaller than actual fully composite deflections. In the fully composite deflection calculation using transformed sections it is assumed that the two layers of the beam form a continuous bond along the length of the beam. However, because in the actual beams interlayer slip is only resisted at the notch locations the computed fully composite deflection is under-estimated. Furthermore, for both the non-composite and the fully composite cases the deflection at mid-span was calculated using elementary beam deflection formula for simply supported beams according to Equation 1 below. In this calculation shear deformation is neglected, which also introduces a slight error into the efficiency calculations.
 | Equation 1 |
The composite action available, composite efficiency, and composite action observed was calculated using the measured maximum deflection and load values for each load test cycle. The equations used are given below.
Composite Action Available (CAA) =  | Equation 2 |
Composite Efficiency (EFF) =  | Equation 3 |
| Composite Action Observed = EFF * CAA | Equation 4 |
Detailed test results for all the specimens are provided in Appendix A.
3.2 Deep Beam Test Results
As a typical example, Table 3.1 lists the results observed for load test DB1.2. All results refer to mid-span deflections, and each row of the table corresponds to one cycle of the load test. The second and third columns of the table give the measured maximum mid-span deflection and the load recorded at the maximum deflection reading respectively. The fourth column shows the ratio of beam length to maximum mid-span deflection. The fifth and sixth columns give the predicted fully composite and fully non-composite mid-span deflections calculated using Equation 1. The remaining three columns of the result tables give the composite efficiency calculations based on Equations 2, 3, and 4.
Table 3.1 Deep Beam 1 Load Test 2
| Deep Beam 1 - Glued Dowel Connection | ||||||||
| Midspan Deflection Results - Load Test 2 | ||||||||
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
| Repetition | Max, Δ (cm) | Load (kN) | L/Δ | ΔNC (cm) | ΔFC (cm) | Composite Action Available | Composite Efficiency | Composite Action Observed |
|---|---|---|---|---|---|---|---|---|
| 1 | 0.90 | 34.9 | 405 | 1.74 | 0.46 | 73.76% | 65.08% | 48.00% |
| 2 | 0.93 | 35.7 | 395 | 1.78 | 0.47 | 73.76% | 64.79% | 47.79% |
| 3 | 0.94 | 36.6 | .88 | 1.82 | 0.48 | 73.76% | 65.36% | 48.21% |
| 4 | 0.94 | 36.4 | 389 | 1.81 | 0.48 | 73.76% | 65.23% | 48.11% |
| 5 | 0.94 | 36.2 | 390 | 1.80 | 0.47 | 73.76% | 65.04% | 47.97% |
| 6 | 0.94 | 36.3 | 389 | 1.81 | 0.47 | 73.76% | 65.00% | 47.94% |
| 7 | 0.95 | 36.6 | 386 | 1.82 | 0.48 | 73.76% | 65.08% | 48.00% |
| 8 | 0.95 | 36.8 | 387 | 1.82 | 0.48 | 73.76% | 67.91% | 47.88% |
| 9 | 0.95 | 36.4 | 387 | 1.81 | 0.47 | 73.76% | 64.65% | 47.69% |
| 10 | 0.94 | 36.0 | 389 | 1.79 | 0.47 | 73.76% | 64.33% | 47.45% |
| Average | 0.94 | 73.76% | 64.95% | 47.90% | ||||
The modulus of elasticity value for the deep beams was taken as 896.3 kN/cm2 (1300 ksi) for the wood, and as 2190 kN/cm2 (3177 ksi) for the concrete. The modulus of elasticity of the concrete was calculated from the average of six standard compressive strength test cylinders tested at 28 days after placement of the concrete. This was done in accordance with the ACI 318-99 standards. The modulus of elasticity of the wood was taken as the tabulated value in the American Forest & Paper Association's 1997 edition of the National Design Specifications for Wood Construction.
From Table 3.1 it is evident that a gradual decrease in composite efficiency was observed for successive repetitions of the loading. This gradual decrease was observed for both deep beam specimens. It is attributed to the notch connection loosening as the number of load test repetitions increases. Further, it has been noted that re-tightening the notch connections is necessary after allowing wood-concrete composite beams to cure in place for extended periods of time. Unfortunately, because of unexpected delays due to equipment failure the Deep Beam 1 was left to creep under its own weight for nearly one year before load testing began.
Table 3.2 presents a summary of the results for all the deep beam load tests. Large increases in composite efficiency (23%) and composite action observed (17%) were found in load test DB1.2 with respect to load test DB1.1. These differences are attributed to the fact that the notch connections in deep beam 1 were not re-tightened before initiating load test DB1.1. This was noticed after load test DB1.1 and the notch connections were then re-tightened by applying a 70 N-m (51.63 lb-ft) torque to the connectors. Also, as noted above, deep beam 1 was in place for nearly one year before testing. During that time creep deflections occurred. Thus load test DB1.1 shows the response for loosened notch connections after a long period of creep. The subsequent tests for deep beams, load tests DB1.2 - DB1.5, are more representative of the service behavior of deep beam 1 with tightened notch connections. However, about six weeks of creep occurred between tests DB1.3 and DB1.4. For deep beam 2, the connections were retightened before each load test, including DB2.1. Thus tightening was not an issue. However this specimen experienced creep from the time of casting until the load test of DB2.1. This initial period of creep was the same for deep beam 1, i.e. about one year.
Table 3.2 Deep Beam Test Results for All Load Tests
| Beam | Date Tested | Load Test | Connection Type | Number of Repetitions | Composite Action Available | Composite Efficiency | Composite Action Observed |
|---|---|---|---|---|---|---|---|
| Deep Beam 1 | 2/20/2002 | DB 1.1 | Glued-Dowel | 5 | 73.76% | 41.86% | 30.88% |
| Deep Beam 1 | 2/20/2002 | DB 1.2 | Glued-Dowel | 10 | 73.76% | 64.95% | 47.90% |
| Deep Beam 1 | 2/20/2002 | DB 1.3 | Glued-Dowel | 10 | 73.76% | 65.79% | 48.52% |
| Deep Beam 1 | 4/4/2002 | DB 1.4 | Glued-Dowel | 5 | 73.76% | 64.24% | 47.38% |
| Deep Beam 1 | 4/4/2002 | DB 1.5 | Glued-Dowel | 0 | 73.76% | 59.36% | 43.78% |
| Deep Beam 1 | 4/4/2002 | DB 1.6 | Glued-Dowel | 1 | 73.76% | 45.96% | 33.90% |
| Deep Beam 1 Average (excluding failure load tests DB1.5 - DB1.6) | 73.76% | 59.21% | 43.67% | ||||
| Deep Beam 2 | 2/18/2002 | DB 2.1 | Mechanical | 5 | 73.76% | 82.87% | 61.12% |
| Deep Beam 2 | 2/18/2002 | DB 2.2 | Mechanical | 8 | 73.76% | 84.22% | 62.12% |
| Deep Beam 2 | 2/19/2002 | DB 2.3 | Mechanical | 1 | 73.76% | 69.71% | 51.41% |
| Deep Beam 2 | 2/19/2002 | DB 2.4 | Mechanical | 1 | 73.76% | -28.97% | -21.37% |
| Deep Beam 2 Average (excluding failure load tests DB2.3 - DB2.4) | 73.76% | 83.54% | 61.62% | ||||
The small changes in composite action observed in going from load test DB1.2 to load test DB1.3 to load test DB1.4 for deep beam 1 seem to be the result of the gradual decrease in composite efficiency due to connector loosening. Load test DB1.6 shows sharply decreased values of composite efficiency and composite action. This is because this was a post failure load test conducted after the failure load test DB1.5. Similarly, a noticeable (14.5%) drop off in composite efficiency occurred in load test DB2.3 compared to DB2.4. This was because DB2.3 was a failure load test and DB2.4 was a post-failure load test. The negative efficiency value found from test DB2.4 simply means that the already failed specimen deflected beyond its theoretical fully non-composite limit.
Results for all of the load tests for deep beams 1 and 2 were similar to those shown in Table 3.1 and are provided in the Appendix A. Results for deep beam 2 showed a higher average composite efficiency (83.5%) as compared to deep beam 1 (59.2%). This is possibly due to the different connector type, but could also be from a number of other causes, e.g. better concrete consolidation in the notches or less knot defects or less cracks in the wood layer. The results of the failure load test (load tests DB2.3 and DB2.4) for deep beam 2 and for deep beam 1 (load tests DB1.5 and DB1.6) are not directly comparable. This is because deep beam 2 was forced to deflect beyond its theoretical fully non-composite limit, whereas deep beam 1 was not.
Figure 3.1 Load Test DB1.1 Load vs. Mid-Span Deflection Response
Figures 3.1 and 3.2 show the observed load-deflection and load-slip characteristics, respectively, for load test DB1.1. These plots are representative of the typical results for all the deep beam load tests. In general, after the first loading an approximately linear elastic load-deflection was observed. The typical load-slip relationship, as shown in Figure 3.2, exhibited much more non-linearity and in-elastic deformation than the load-deflection characteristic. The overall load-deflection and load-slip response of deep beam 2 had similar properties as those of deep beam 1. However, the similarity of the responses of deep beam 1and deep beam 2 was not evident in the failure load tests. Deep beam 1 and deep beam 2 experienced completely different types of failure. Deep beam 1 failed in a mid-span flexural type failure mode in the wood, (Figure 3.3). Deep beam 2 failed due to cracks propagating in the wood layer behind the outside notch, effectively failing the notch (Figure 3.4).
Figure 3.2 Load Test DB1.1 Load vs. Relative Interlayer Slip Response
Figure 3.3 Deep Beam 1 Failure at Mid-Span
Figure 3.4 Deep Beam 2 Failure at Notch Connections
3.3 Wide Beam Test Results
Again, the modulus of elasticity of the concrete was based on standard compressive strength test cylinders tested at 28 days. However load tests were conducted long after the 28-day cure period and thus the value used under estimates of the actual value at the time of the test. For the wide beams the modulus of elasticity of the wood was taken as 1103 kN/cm2 (1600 ksi) and the modulus of elasticity of the concrete was taken as 2190 kN/cm2 (3177 ksi). Note that after being constructed with a 15.25 mm (6 inch) layer of concrete placed on top of the wood layer, the wide beam specimens were un-shored until the load tests were conducted. However, the deflections due to creep were not measured. Deflection measurements recorded in the various load tests were due to the applied load only. The modulus of elasticity of the wood layer (1103 kN/cm2) was again taken as the tabulated value in the American Forest & Paper Association's 1997 edition of the National Design Specifications for Wood Construction. The results of the load tests on the four wide beam specimens are shown in Table 3.3.
Table 3.3 Wide Beam Load Test Results
| Beam | Date Tested | Load Test | Connection Type | Number of Repetitions | Composite Action Available | Composite Efficiency | Composite Action Observed |
|---|---|---|---|---|---|---|---|
| Wide Beam 1 | 5/17/2001 | WB 1.1 | Glued-Dowel | 5 | 74.93% | 43.92% | 32.91% |
| Wide Beam 1 | 5/21/2001 | WB 1.2 | Glued-Dowel | 8 | 74.93% | 21.07% | 15.79% |
| Wide Beam 1 Average | 74.93% | 32.50% | 24.35% | ||||
| Wide Beam 2 | 5/17/2001 | WB 2.1 | Mechanical | 5 | 74.93% | 46.86% | 35.11% |
| Wide Beam 2 | 5/24/2001 | WB 2.2 | Mechanical | 8 | 74.93% | 33.60% | 25.17% |
| Wide Beam 2 Average | 74.93% | 40.23% | 30.14% | ||||
| Wide Beam 3 | 7/29/2003 | WB 3.1 | Mechanical | 8 | 74.98% | 24.63% | 18.46% |
| Wide Beam 3 | 7/30/2003 | WB 3.2 | Mechanical | 8 | 74.98% | 22.63% | 16.97% |
| Wide Beam 3 Average | 74.98% | 23.63% | 17.72% | ||||
| Wide Beam 4 | 8/4/2003 | WB 4.1 | Glued-Dowel | 8 | 74.98% | 15.05% | 11.28% |
| Wide Beam 4 | 8/5/2003 | WB 4.2 | Glued-Dowel | 8 | 74.98% | 11.81% | 8.86% |
| Wide Beam 4 Average | 74.98% | 13.43% | 10.07% | ||||
Figure 3.5 Wide Beam Composite Efficiency Results
In general the wide beam specimens showed a low degree of composite efficiency in general and compared to the deep beams. The range was 11.81 percent to 46.86 percent compared to 41.86 percent to 84.22 percent for the deep beams. Figure 3.5 shows the observed trend of gradually decreasing composite efficiency with respect to increasing number of load test repetitions. Wide beam 1 exhibited a significant drop in efficiency after load test WB1.1. However during load test WB1.2 the drop in efficiency was low and very gradual. Wide beam 2 exhibited a low and gradual drop in efficiency during load test WB2.1. However, similar to wide beam 1 a large drop in efficiency occurred after load test WB2.1 and a gradual drop in efficiency was observed during the loading sequences of load test WB2.2. The large drop in composite efficiency observed for wide beams 1 and 2 appear to be isolated cases. Considering the large (15-20%) and sudden drop, it is possible that unobservable failures occurred after the first load tests of wide beams 1 and 2. Figure 3.6 and Figure 3.7 show data which suggest this suspicion. Figure 3.8 shows the notch numbering scheme referred to in the preceding graphs and following discussion.
Figure 3.6 Load Test WB1.1 Load vs. Deflection Response
Figure 3.7 Load Test WB2.1 Load vs. Slip Response
In Figure 3.6 notch #1 of wide beam 1 undergoes a large relative slip between the wood and concrete layers, roughly four times the slip measured at the other five notches. If it is assumed that notches #1 and #6, closest to the supports, have identical forces in the notch, then Figure 3.6 suggests an ineffective or failed notch connection contributed to the behavior. However, Figure 3.6 does not plot notch force vs. slip, thus the presence of an ineffective notch is only speculation. As support for this hypothesis, there are a number of potential causes for the ineffective notch. One probable cause is that the dowel connector for notch #1 for load test WB1.1 was either not torqued sufficiently or was over-torqued, resulting in a large gap between the concrete layer and the wood layer at the notch. Figure 3.7 displays the load vs. relative interlayer slip results for WB2.1. The data suggests that wide beam 2 had one very inefficient notched connection as evident from the large relative slip (compared to the other notches) occurring at notch #2 in wide beam 2. Comparing Figures 3.6, 3.7, (and subsequently Fig. 3.9) it is seen that the gradually decreasing efficiency of wide beams 1 and 2 is likely predominantly due to gradually decreasing effectiveness of the notches #1 and #2 of these respective beams. However the gradually increasing slip at the ineffective notches does not explain the sudden and dramatic drop in composite efficiency observed for wide beams 1 and 2 in Figure 3.6.
Figure 3.8 Wide Beam Specimen Notch Numbering
Figure 3.9 Load Test WB1.2 Load vs. Deflection Response
The large sudden drop in efficiency of wide beam 1 is likely from a connection failure. This is deduced from data in Figure 3.9 where notch #2 of wide beam 1 exhibits large interlayer slip behavior that was not displayed for the same notch in Figure 3.6. The onset of large slip at notch #2 from the beginning of load test 2 suggests that failure occurred sometime between Load Test WB1.1 and Load Test WB1.2. It is probable that the failure materialized, but was not recognized, during equipment re-calibration tests that were made in the time between load tests 1 and 2.
The sudden drop in efficiency of wide beam 2 appears to be due to an entirely different type of failure. The comparison of load vs. slip curves for load tests WB2.1 and WB2.2 do not suggest any failed notches. However, a comparison of the load vs. deflection plots for load tests WB2.1 and WB2.2 indicates that a possible flexural failure occurred in wide beam 2 sometime between the two different load tests. Specifically, no sudden drops in stiffness were observed during either load test WB2.1 or load test WB2.2 which suggests that no flexural failures occurred during the conduct of these load tests. Figures 3.10 and 3.11 show a large decrease in stiffness and a large increase in deflection at the 3/4 point location of the span of wide beam 2 in Load Test WB2.2 compared to load test WB2.1.
Figure 3.10 Load Test WB2.1 Load vs. Deflection Response
Figure 3.11 Load Test WB2.2 Load vs. Deflection Response
Unlike the results for wide beams 1 and 2, the low composite efficiency results for wide beams 3 and 4 do not appear to be the result of prior failures. Table 3.3 shows that these beams experienced only a gradual decrease in composite efficiency as the number of load test repetitions increased. Further, wide beams 3 and 4 were found to have low composite efficiencies from the outset of their testing program. The load vs. deflection and load vs. slip responses of wide beams 3 and 4 changed very little in going from load test 1 to load test 2 for each of the two beams. Inspection of the load vs. deflection responses shows that both wide beam 3 and wide beam 4 had predominantly linear, elastic load-deflection responses with small relative slip measurements at the notch locations. Figures 3.12 and 3.13 are representative examples of these observations.
Figure 3.12 Load Test WB4.1 Load vs. Deflection Response
Figure 3.13 Load Test WB4.1 Load vs. Relative Slip Response
Figure 3.14 depicts the typical failure observed at a notch for wide beams 3 and 4. The failures of wide beam specimens 3 and 4 occurred at the notch locations, similar to deep beam specimen 2. However the mode of notch failure for the wide beam specimens differed from that observed for deep beam specimen 2. In the wide beams the concrete layer cracked across the top of the notch, parallel to the layer interface. Failure at a single notch occurred suddenly with subsequent progressive failures at other notches, followed by a flexural type failure near mid-span. It is notable that notch type failures occurred in wide beam specimens 3 and 4, which exhibited very little composite efficiency (11-25%). In general notch type failures require sufficient composite efficiency to generate the shear stresses needed to result in failure at the notch. That is flexural type failures are consistent with the expected failure of a fully non-composite beam (low composite efficiency beam), and notch type failures are consistent with the expected failure of a fully composite beam (high composite efficiency beam).
Figure 3.14 Notch #4 Failure of Wide Beam Specimen #4
One likely contributor to the failure response observed in beam 4 is the particular rebar arrangement for the notches of that beam. Unlike the other three beams, the bottoms of the notches of beam 4 were not reinforced with transverse rebar. Beam 4 also did not have wire mesh placed in the bottom of its notches. Figure 3.15 shows the reinforcement details for beams 4 and 3 pictured first and second from the left-hand side of the picture. Beam 4 is pictured on the far left-hand side of Figure 3.15; the absence of transverse rebar at the notch locations is apparent by comparing the picture of beam 4 with the picture of beam 3 just its right. The unusual failure characteristics of beams 3 and 4 are one supporting argument that further studies of the reinforcement details for wood-concrete composite beams (and decks) are needed.
Figure 3.15 Rebar Placement for Beams 3 and 4
Another possible cause for the failure response of wide beams 3 and 4 may be due to the stresses generated by slipping and creeping of the beams for 2 1/2 years. It may be that the slip resistance lost during a long duration of static dead load is not completely recoverable via connector re-tightening. It should also be noted, as seen in Figure 3.14, that concrete in the notches of the wide beam specimens was not completely consolidated to fill the notch shown. The poor consolidation is surely detrimental to the performance of the notch, and this perhaps is the cause of the failure characteristics of wide beams 3 and 4. However, a creep analysis of wide beams 3 and 4 would be a pre-requisite to better explaining the link between their low composite efficiency behavior and notch type failure response.