Unsure girder member forces and strange girder deformation
Dear Technical Support,
Thank you for your help for my questions raised in my previous emails. I have a better understanding about the construction staging load cases and results.
I modified my data file and examined the results from the data file, and I found somethings unusual:
(1) Strange deformation: some elements have excessive bending deformation as shown in the attached image
The bending moments at Node 202 created by the prestressing forces at the Node 202 alone is +18762 kNm, which should not produce a curvature change as that large (CS7 stage 9, Step 1, first step). The forces at CS7 Stage 9 Step 1 should not neither produce this strange deformation. (Elements from Node 202 - 212 are newly added, each is 0.5 m long)
The girder section at Node 202 is as follows:
Ax = 9.51 m^2, Iy = 16.61 m^4, neutral centroid Gc is 1.454 m to the top deck and 1.894 m to the bottom of the bottom deck.
The deformation is strange:
CS7 Stage 9 Step 1: the deflection at Node 203 is 22.8 mm while it's 53.9 mm at Node 202, the 31 mm change is too large for the length of 0.5 m element.
(2) The axial force created by the 4 tendons at Node 202 after loss at CS7 Stage 9, step 1) is 15620 kN, while the 19 stands/duct, 4 ducts with 0.8 fpu jacking force is 15832 kN. This means that the loss at the tendon anchorage end is very small ( 1 - 15620/15832) *100% = 1.34%.
This seems incorrect. I configured the tendon loss as normal (mu = 0.18, K = 0.002) with an anchorage recess of 12 mm at the end.
(3) I configured the creep, shrinkage and compression strength development as CEB-FIP2010 for time dependent concrete parameters, which should be close to Canadian code. The concrete strength is 50 MPa for the girders. But how do I find the Young's Modulus for the concrete at a known CS stage?
(4) How can I instruct the program to write stiffness matrix (diagonals) at CS7 Stage Step 1 and the right hand force vector into the .OUT file?
This is the way I could find why the deformation is so wrong.
The data file is attached in this ticket.
Thank you very much to help me to resolve my doubts.
Dear DK Lee,
I think that the first issue (Issue 1) in my ticket is due to the fact that the lasted added segment is drawn to the initial position of the segment, which is wrong with the segmental cantilever construction.
The new segment is have zero displacement and zero stress relative to the previous configuration of the structure when it is in the curing state in the formwork, and the new segment will have a special shape to match the final profile designed by highway alignment.
If this is the case, I could process myself for the movement. I hope that this deflection mismatch will not create any additional force in the structure. Please confirm that my guess is right.
Dear DK Lee,
I checked my input data again for tendons and I found that the anchorage recess of 12 mm is assigned to the beginning of the cantilever tendons in Pier 1, while the jacking force is applied to the end of all cantilever tendons in Pier 1. When I modified the recess at the correct end instead of at the beginning, I get the correct girder axial forces in the girder.
I have a suggestion for the tendon property assignment in Midas, which should follow the same logic as in CSiBridge: one tendon is assumed to have uniform section area and jacked at one or two ends with the equal jacking forces; the anchorage recess at both ends is equal. The current tendon jacking forces have two values for the prestressing forces at the start and the end, which may never happen to have different values, and subsequently there are no reasons to have different values for anchorage recess at the end and the start. The 3-span continuous post-tensioning concrete bridges normally have two piers with mirror arrangement for almost everything including tendons. The same types of tendons with the same parameters have to have two set of properties values for jacking forces and anchorage recess in this case in current Midas system.
So the issues (1) and (2) I raised in my ticket may not be a problem. The Issue (3) is important to know how Midas calculate the Young's modulus of the concrete varying in time. Issue (4) may not be available in Midas.
Anyway, I would like to hear from you about these 4 issues I raised.
Thank you very much.
Dear DK Lee,
Am I right about the guess that: the initial position of the newly added segment is drawn to a location relative to the initial position plus the displacement loaded in the current stage?
The initial position of the newly added segment should be the one obtained in the previous stage at the last step with natural extension to the tip with the same 3D coordinates and zero rotation for the current segment. This initial position for the newly added segment is far away from the initial position set by the coordinates of the joints. This is why I have a 200 mm difference for the first and the send joints in the same segment.
It seems that this large displacement change does not produce any force effects in the new segment.
Please confirm the above conclusions.
Thank you very much.
Dear DK Lee,
I am looking for the Young's modulus of the concrete during construction staging and found that the values are out of my imagination.
I have two concrete materials CSA C50 (for box girder) and CSA C35 (for concrete piers) and defined the time dependent parameters according to CEB-FIP 2010. At Stage CS8 in the last stage, I found that the Young's modulus is as follows:
Ec = 4.72 GPa with age = 32 days for C50 configured with average strength of 60 MPa.
Ec = 44.439 GPa with age = 123 days for C35 configured with average strength of 42 MPa.
Both values should not be correct. The modulus of elasticity of the concrete should be approximately equal to:
Ec = 38 GPa for C50 concrete with age = 30 days, 45GPa with age = 60 days
32 GPa for C35 concrete with age = 30 days, 38 GPa with age = 60 days
I did a lot tests in China for Young's modulus of concrete and have some experience. The values given by Midas are not correct.
Please look at the attached files in this new post for the results.
For the modulus of elasticity, I will reply separately.
Dear DK Lee, thanks for this important information about the default and real displacement.
For default displacement, the value for the new segment at the time of addition should have an initial value to the one at the tip of the previous step, without rotation. The rotation is complicated and the new segment never follows the rotation, especially for cast-in-place segmental construction; it's not followed even in precast segmental construction because the new segment has a special shape to get the required profile at finish.
Thank you very much for your explanation. I need to correct the results of real deflection obtained by Midas to remove the contribution from the rotation.
Dear DK Lee,
Thanks for your help.