Hi, I have a question for modelling long (~50m) tie rods,
Tie rods subject to settlement can sometimes lead to high bending in the tie rod, reducing axial capacity. For large marine structures (reclaimed land) - when modelling tie rods i have always adopted a Node-to-Node anchor approach. When settlement ducts are not specified, we check the vertical displacement of the soils after tie rods are installed (reset displacements) and read the settlement profile at tie rod level. We can prescribe the rod to this profile and insert appropriate articulation at the points of highest curvature.
This approach is extremely conservative, as it follows that the tie rods will deform in the same profile as the soils. In reality, there will be settlement around the tie rods, a bit like cheesewire cutting through a material, particularly when the ties are subject to ~4MN tension each.
It got me thinking whether we can use embedded beam rows to model tie rods, turned on the side so the vertical loading and settlement is experienced by the embedded beam row as lateral behaviour. Reviewing the PLAXIS Reference manual, Embedded beam rows, 6.6.3 P293 (top) it states:
The lateral displacements can be induced by a transverse force applied at the top of the pile or as a consequence of the transverse distributed load induced by the lateral displacement field of the surrounding soil.
This would suggest that i could use an embedded beam row to model the tie rods (articulation allowed by short node to node anchors at either end) and the soil would be allowed to flow through them - dependent upon the ultimate lateral resistance and lateral interface stiffness being correctly calculated for the soil.
Has this been done, and do you consider embedded beam rows suitable for such an application?
Many thanks
Scot
Dear Scot,
Nice overview! Maybe the embedded beam can be used as indeed they do allow "soil" to go around the structure (based on the "special" interface behaviour), but I am not fully sure about the connections to be considered.
Can you share a screenshot of what exactly you want to model?
Coming back to this, the main problem is the connectivity of the embedded beam.
As I understood you want to connect it on each edge but unfortunately the embedded beam has only one connection point (picked by the user)
So, in that case, I am not sure you can use it the way you want, which is why I wanted to see a model or a sketch or a screenshot of what exactly you want to consider.
Hi Stefanos,
Thanks for the reply!
I have connected the 'top' of the beam to the main quay wall using the hinged connection to model the articulation at this point. I have connected the 'bottom' of the embedded beam by a very short Node to Node anchor which is then also connected to the anchor wall.
Under this approach, i reset displacements just before activating the 'tie rod' so i can check the total settlement of the fill behind the wall. Naturally this imparts a significant load onto the connection point and first ~4 metres of the tie rod, as the tie cannot move exactly to the same profile as the soil. I have therefore been trying to calibrate / sensitivity check the lateral resistance parameters with those of a tie rod forced upwards through the soil.
Screenshot below (i have now kept only 1 embedded beam rather than chaining them together to model additional articulation points after 5.7.3 of the manual:
Would be happy to share a model if you think there is value in this approach. Unfortunately there is no real design methodology to follow on this - it is accepted that the settling soil must move through the tie rods, particularly the soils close to the quay wall, but i cant see that its ever been analysed in such a way.
Thanks for the response and interest.