Evaluation of the test results

reason of the non-linearity was found to be a local failure in the lower chord at the structural joints next to the support, symmetrically on both sides of the truss (Figure 95.). Despite the local phenomenon no loss of load-bearing capacity occurred, but the truss showed to further load increase ever-decreasing stiffness. As the load was increased local buckling waves were observed in the upper chord member second next to the ridge joint, with forming pinches in the web and web-flange junctions (Figure 96.), similarly to the phenomena in the first test.

Figure 95.: Failure in the lower chord Figure 96.: Local buckling in the upper chord.

The final failure of the truss was caused by the out-of-plane flexural bucking of the upper chord members second next to the ridge joint on one side of the truss (Figure 97.). Although not connected to each-other, both chord members buckled in the same direction similarly to the failure in the third test. The load-bearing capacity of the fifth specimen was 47.4 kN/jack (Figure 98.).

The failed upper chord member is similar in its arrangement to a member with a C arrangement in the single C-section tests (see Chapter 2.2.5, group A), at the ridge joint the load is introduced in the upper chord at the flanges and web as well. This results a more uniform stress distribution along the cross-section compared to the failed member of the first test; the difference is similar to that between a C-section with SimpleC and one with a C arrangement.

Figure 97.: Failure of the upper chord. Figure 98.: Force-deflection diagram – Test 5.

deflections and acting internal forces calculated from results of the strain measurement.

Comparisons involving the design method are presented in Chapter 3.3.

The direct results of the tests: load-bearing capacity, failure mode and force-deflections diagrams are summarized and compared in Table 16 and Figure 99.

Table 16: Summary of the test results.

Test Initial stiffness [kN/mm]

Ultimate

load [kN] Failure mode

1 0.83 28.5 Interaction of flexural buckling and bending;

upper chord

2 0.86 35.5 Joint failure; ridge joint

3 - 36.4 Flexural buckling of built-up member; upper chord 4 1.04 37.4 Interaction of axial compression and bending;

brace member

5a 1.07 37.0 Joint failure; lower chord

5b 1.07 47.4 Interaction of flexural buckling and bending; upper chord

According to the measured initial stiffness two groups can be distinguished among the test specimen: the initial stiffness of first and second test specimens with a lower, and the fourth and fifth specimens with a higher value. The higher stiffness of the two latter is clearly a consequence of the stiffer upper chord. In the case of the third specimen the changing stiffness is caused by the slipping bolted connections, the slope of the curve is in this case approximately equal to that of the first two tests. A clear plateau cannot be observed on the force-deflection diagrams; partially because in some tests the loading process was cancelled to preserve the specimen for another test (Tests 1., 4.), in the other cases a rapid limit-point type failure was obtained, preceded only by ever-decreasing global stiffness.

0 10 20 30 40 50

0 10 20 30 40 50 60 70 80 90 100

Deflection [mm]

Load [kN/jack]

Test No. 1 Test No. 2 Test No. 3 Test No. 4 Test No. 5 ULS level

Figure 99.: Load-deflection diagrams of the tests.

The obtained load-bearing capacity is in the first test under the ULS load level. The premature failure of the truss is the consequence of the ridge joint configuration, resulting big out-of-plane eccentricity. In Tests 3, 4 and 5 the detailing of this joint was changed to result more favourable structural behaviour; in the last three tests no sign of failure was observed in this region, and the ULS load level was reached, in the last test the ULS level was exceeded by 50%.

The observed behaviour of the specimens follows the structural symmetry. This reflected in all tests in the symmetrical deflections and the failure modes obtained in Test 4, where brace columns in the same position on both sides of the truss failed, and in Test 5, where the failure in the lower chord was observed on both sides. The symmetric behaviour can be observed in the internal forces as well. Acting axial action and the biaxial bending were calculated in all cross-sections where at least three strain gauges were used; in the case of cross-sections with more than three gauges the forces are calculated by permutation the gauges used, the results are averaged. The calculation method is based on the basic assumptions of elasticity: rigid cross-sections, linear elastic material, hence it is not capable of following local buckling. In the evaluation the properties of the effective cross-section are used in case of compression members, for tension members the gross cross-section is used.

A complete list and position of the cross-sections used in the analysis is presented in Figure 100. and Table 17. The symmetrical behaviour is shown in Table 18 and Figure 101. Cross-sections with the same letter and different number indicate that measurement was carried out on both C-sections (i.e. chord members). Cross-section B1 is not shown in Figure 100., as it is on the opposite (“right”) half of the truss, symmetric to cross-section A1. Although strain measurements were carried out in the lower chord as well, these are not used in the evaluation, as in these members only two strain gauges were used each cross-section. It is to be noted, that as in Test 4 and Test 5 the same specimen was used (with member 12 changed), hence the evaluation of the strains is based on the same gauge positions.

Figure 101. shows that the behaviour of the truss is symmetric and can be considered linear for loads smaller than 20 kN/jack, the limit where the global behaviour of the truss used in Test 1 becomes non-linear (Figure 99.). Note, that in this evaluation the elastic state of the material is not checked, hence internal actions for load levels higher than 20kN/jack are not valid, but included, to provide a complete overview.

Figure 100.: Notation of cross-sections with strain measurement. Blue numbers indicate measurements in Test 4 and 5, red numbers in Test 1.

Table 17: Position of cross-sections with strain measurement.

Cross-section Specimen Section Position

A1 1

A2 1 Member 2, 2250 mm from ridge

B1 1

C150/2.0

Member 2, 2250 mm from ridge, opposite side

C 1 Member 14, 540 mm from lower end

D 1 Member 15, 850 mm from upper end

E 1 C100/2.0

Member 16, 880 mm from upper end

F1 4, 5

F2 4, 5 Member 1, 820 mm from ridge

G2 4, 5

C150/2.5

Member 2, 2250 mm from ridge H 4, 5 C100/2.0 Member 13, 540 mm from upper end

Table 18: Internal actions calculated from measured strains, A1, A2, B2 cross-sections.

N [kN] My [kNcm] Mz [kNcm]

Load

[kN] A1 A2 B2 A1 A2 B2 A1 A2 B2

0.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 10.0 -20.07 -21.56 -22.55 4.86 6.27 2.31 -10.94 -12.82 -8.55 15.0 -30.85 -33.16 -34.73 3.69 6.00 5.47 -16.93 -19.75 -12.60 20.0 -40.84 -43.98 -46.77 4.21 7.53 7.99 -23.09 -26.81 -16.82 22.5 -42.82 -46.73 -53.25 35.15 39.08 7.30 -28.76 -33.56 -20.46 22.7 -43.33 -47.29 -54.01 35.30 39.29 7.66 -29.15 -34.01 -20.99 24.6 -54.06 -58.35 -60.46 -29.36 -21.65 14.41 -32.27 -35.24 -26.62

-70 -50 -30 -10 10 30 50

0 5 10 15 20 25

Load [kN/jack]

Axial compression [kN], Bending [kNcm]

A1, N A2, N B1, N A1, My A2, My B2, My A1, Mz A2, Mz B2, Mz

Figure 101.: Internal forces in function of the load – A1, A2, B2 cross-sections.

The structural symmetry results compression and bending actions in the two members of the upper chord (A1, A2) and in the members on the opposite side of the truss (A2, B2) with values in good agreement and tendencies. The peaks in Figure 101. also show that the failure in this test occurred on the side of the truss to that cross-sections A1 and A2 belong. It is to be noted, that the measured bending moments are small in value, the biggest absolute value is 0.35 kNm.

The comparisons show the coherence of the measurement data: deflections can be considered accurate, the calculated internal actions show satisfactory agreement when compared to each-other. Further analysis of the internal actions is to be found in Chapter 3.3.2.

3.3. Design method