This experimental work was used to develop the steel angles as external dissipaters and the design of the dissipaters evolved throughout the testing. The first three specimens did not have the desired behaviour that was achieved in the last two specimens. The main parameters that will be discussed here are the achieved forces, ductility and cyclic behaviour. The achieved force for a target displacement is the main parameter for the design of the dissipater. The ductility is important because the dissipaters work mainly in
Table 3.2: Tensile tests for steel used in specimens.
fy εy σmax εmax εsu Es
(MPa) (%) (MPa) (%) (%) (GPa) L150x150x15 309.9 0.15 433.8 21.46 30.70 201.1 L200x200x16 311.8 0.15 436.6 21.94 33.22 214.9
a plastic range, needing ductility to achieve the design displacement. The achieved forces and displacement are summarised in Table 3.3. The cyclic behaviour is evaluated by the shape of the hysteretic diagram in each cycle, which gives the information on the stability of the dissipater along the cycles, the energy dissipated and the response in tension and compression.
Table 3.3: Forces for one angle and maximum sustained displacement. F(kN) - Maximum force for different displacements dmax(mm)
5 mm 10 mm 15 mm 20 mm 25 mm 30 mm A1 104 116 127 138 148 - >25 A2 123 143 156 175 188 192 30 A3 116 138 154 176 193 181 25 A4 100 122 136 149 135 - 20 A5 103 125 138 152 165 188 30 3.2.5.1 Specimen A1
The deformed shape of specimen A1 is shown in Figure 3.15. From this picture it should be highlighted that the contact between the angles and the base starts in the middle of the washer plates and the washer plates were bent by flexure. This means that the washer plates and the bolts were not able to control the position of the plastic hinges position, that formed below the washer plates instead of next to them. Another consequence of this fact can be observed in Figure 3.16, that shows the lack of contact between the base and the edge of the angles. The lack of contact that occurred in the compression phase and in the beginning of the next tension phase, was due to the residual deformation left by the prying action responsible for the plastic hinge near the bolts of the horizontal leg.
The hysteretic response diagram is shown in Figure 3.17. The displacements were measured between the vertical legs and the base using transducers D1 and D4 (Figure 3.10a). The displacements shown are the average of transducers D1 and D4. The maximum achieved force for a given displacement can be obtained from the hysteretic diagram or from Table 3.3. The cyclic behaviour was not the desired one, because in the reload part of the cycles, when the specimen goes from compression to tension, the stiffness decreases significantly, compromising the cyclic response. The maximum achieved displacement was 25 mm with no evidence of failure.
The problem referred to in the cyclic behaviour was associated with three possible causes: the prying action forces that resulted in the lifting of the edges of the angles, which can be observed in Figure 3.16; an eventual yielding of the bolts in the horizontal legs submitted to bending and tensile forces and the longitudinal ovalization of the holes
in the horizontal legs which allows the angles to slip. The washer plates suffered plastic deformations and were not able to impose the plastic hinges position in the angles or to control the prying action. To mitigate these problems, the specimen tested after this had three main differences: thicker washer plates; more resistant bolts and standard washers in all bolts and threaded rods. The use of thicker washer plates and the more resistant bolts intended to control the position of the plastic hinges and to keep the edges from lifting. The standard washers should allow a more uniform distribution of the stresses applied by the head of the bolts to the washer plates. All these changes were made to improve the behaviour of the washer plates and bolts on the horizontal legs of the angles and mitigate the first two possible causes of the cyclic behaviour problem outlined above.
Figure 3.15: Picture of deformed specimen A1.
200 150 100 50 0 0 -200 -150 -100 -50 Fo rc e (k N ) Displacement (mm) 5 10 15 20 25 30
Figure 3.17: Hysteretic response diagram of specimen A1 cyclic test.
3.2.5.2 Specimen A2
The deformed position of specimen A2 test is shown in Figure 3.18. In this specimen the washer plates had rigid body behaviour but the angles lifted from the base in the middle of the washer plates, meaning that the plastic hinges in the angles horizontal legs still formed below the washer plates. Figure 3.19 shows the residual deformation of these angles. This figure shows that the edges of the angles still lifted up. In this test it was also possible to notice some slippage of the angles below the washer plates.
The hysteretic response diagram is shown in Figure 3.20. The displacements were measured between the vertical legs of the angles and the base using transducers D1 and D4 (Figure 3.10a). The displacements shown are the average of transducers D1 and D4. The cyclic behaviour was not as desired, because when the specimen went from compres- sion to tension the stiffness decreased significantly, compromising the cyclic behaviour. This means that despite some improvements, the main problem detected in specimen A1 remained. The maximum achieved displacement was 30 mm, with imminent rupture. Ob- serving the response diagram it is possible to see an horizontal plateau near the maximum force and a crack is shown in Figure 3.21 in the plastic hinge below the washer plate zone, meaning that this specimen was near failure. The horizontal leg holes were oval, allowing slippage of the angles and this could influence the variations of stiffness during the test. These holes were changed to circular in the next tested specimen.
Figure 3.18: Picture of deformed specimen A2.
Figure 3.19: Plastic residual deformation of specimen A2. 200 150 100 50 0 0 -200 -150 -100 -50 Fo rc e (k N ) Displacement (mm) 5 10 15 20 25 30
Figure 3.21: Picture of a crack in specimen A2.
3.2.5.3 Specimen A3
As in specimen A2, in specimen A3 the washer plates had rigid body behaviour (Figure 3.22) and the angles started to lift up from the base in the middle of the washer plates, meaning that the plastic hinges were still forming below the washer plates, as may also be seen in Figure 3.23. No slip was observed in the horizontal direction in this specimen because the holes in the horizontal legs were drilled circular instead the oval ones used in specimens A1 and A2.
In the hysteretic response diagram shown in Figure 3.24, the displacements were mea- sured between the vertical legs and the base using transducers D1 and D4. When the specimen went from compression to tension the stiffness decreased significantly, compro- mising the cyclic behaviour. This means that despite the improvements made, the main problem detected in specimens A1 and A2 remained. After this test it seemed clear that the problem could only be solved by restraining vertical displacements on the edge of the horizontal legs. The maximum achieved displacement can be considered 25 mm, since failure was already achieved for 30 mm displacement, as can be seen in the hysteretic response diagram. From Figure 3.25 it is also possible to see the rupture in the plastic hinge below the washer plate zone. Both angles presented a similar rupture.
After three tests with continuous detailing improvements, but without solving the main cyclic response problem, some changes in the design solution of the next specimen had to be made. The first change was the exclusion of the washer plates. The introduction of washer plates was explained in Section 3.2.1 but for the relative stiffness between the angles and the washer plates and bolts used, the washer plates were not able to fulfill their purpose. Without washer plates it is clear that the plastic hinges in the horizontal legs should be in the line of the bolt holes. To restrain the edges from lifting, the edges were welded to the base.
Figure 3.22: Picture of deformed specimen A3.
Figure 3.23: Angle edge lift on specimen A3. 200 150 100 50 0 0 -200 -150 -100 -50 Fo rc e (k N ) Displacement (mm) 5 10 15 20 25 30
Figure 3.25: Picture of cracking in specimen A3.
3.2.5.4 Specimen A4
The deformed shape of specimen A4 test is shown in Figure 3.26. The specimen without washer plates presented the plastic hinge in the line of the bolt holes, as expected, and the contact between the angles and the base started in the line of the bolts. From the residual deformation in the angles shown in Figure 3.27 it is clear that the edge of the angles did not lift, thus the welding fulfilled its purpose.
The hysteretic response diagram is shown in Figure 3.28. The cyclic behaviour was as desired once the stiffness problem recorded in the first three specimen was solved by the use of the welding. The maximum displacement was 20 mm. Observing the response diagram it is clear that for 25 mm displacement, failure had already been achieved. From Figure 3.29 it is also possible to see the rupture in the plastic hinge below the bolts of the horizontal leg, being the same in both angles. This specimen was the one that achieved the smallest maximum displacement, this fact is assigned to the welding, that restrains not only the vertical displacements, but also the horizontal displacements, increasing the tension in the horizontal legs. In the previous tests, the horizontal displacements were a function of friction, the tolerances of the bolt holes and the stiffness of the bolts, resulting in some small horizontal displacements.
Specimen A4 was considered a good solution and was applied in a hybrid rocking wall system that is described in Chapter 4. After this test it was decided that a solution without any welding should be considered, thus a second alignment of bolts was introduced in specimen A5 to substitute the welding. In construction it is recommended that only bolted solutions are used on site to facilitate the assembly and guarantee the quality of the execution.
Figure 3.26: Picture of deformed specimen A4.
Figure 3.27: Plastic deformation of specimen A4. 200 150 100 50 0 0 -200 -150 -100 -50 Fo rc e (k N ) Displacement (mm) 5 10 15 20 25 30
Figure 3.29: Picture of cracking in specimen A4.
3.2.5.5 Specimen A5
To avoid the lifting up of the edges of the horizontal legs without welding, it was decided to use two lines of bolts. For that reason, a larger angle (L200x200x16) had to be used in specimen A5.
The deformed shape of specimen A5 test is shown in Figures 3.30 and 3.31. The deformation pattern was similar to the one of specimen A4, with the plastic hinges of the horizontal legs located along line of the bolts. From the residual deformation in the angles shown in Figure 3.32 it is clear that the edge of the angles did not lift, thus the second line of bolts fulfilled its purpose.
The displacements of the hysteretic response diagram (Figure 3.34) were measured between the vertical legs and the base, using transducers D2 and D5 instead of D1 and D6, due to a problem in last cycles when the transducer D6 suffered an unexpected rotation (Figure 3.10b). These two transducers were located in the second line of bolts and the displacements measured are similar to transducers D1 and D6 until the moment when D6 rotated. The cyclic behaviour was the desired one as the stiffness problem observed in the first three specimens was solved by the use of the second line of bolts. The achieved displacement was 30 mm and observing the response diagram it is not possible to see any evidence of failure. However, in Figure 3.33 it is possible to see cracks in a plastic hinge located near the fillet of the angle, meaning that this specimen was close to failure.
The results of specimen A5 are considered as good and this solution was used in a hybrid rocking wall system that is described in Chapter 4. This solution was the final and recommended solution for the angles as energy dissipaters for cyclic action.
Figure 3.30: Picture of general deformation of specimen A5.
Figure 3.31: Picture of deformed specimen A5.
Figure 3.33: Picture of cracking in specimen A5. 200 150 100 50 0 0 -200 -150 -100 -50 Fo rc e (k N ) Displacement (mm) 5 10 15 20 25 30
Figure 3.34: Hysteretic response diagram of specimen A5 cyclic test.