# Shrinkage in multispan concrete bridge

Question:

Hello,
I am designing a four span bridge made up of prestressed beams with a concrete top slab. I would like to clarify a few things about how the model is working.
1) Shrinkage and creep primary - these are imaginary forces which produce an equivalent elastic length change/curvature to the time dependent effects, is that right? Hence we do not consider them at ULS.
2) For shrinkage, Midas takes the notional size h as the only geometric input for a beam element. This is defined as a property of the material, which is a little counter intuitive. If we use a composite section for construction stage analysis, we can enter another h for each part of the section - does this override the notional size in the material definition?
3) The shrinkage secondary My in the model is enormous for the precast beams. long term construction stage - 2760 kNm. This is a similar magnitude to the dead load bending which intuitively seems completely wrong. Similarly the reactions at the base of the piers, which are up to 1460 kN. Please could you advise if it is correct? And if not, have we defined a model property incorrectly, or would this force actually be relieved by shrinkage cracking in the real structure? Or by the casting sequence of the top slab allowing for autogenous shrinkage before the next pour.
4) Similarly to thermal, the stiffness of restraint must have a big effect on the size of the shrinkage forces. Do we need to account for this flexibility somehow - eg the bearings used on the structure will not be rigid against uplift.

Hello,

Question1: Shrinkage and creep primary - these are imaginary forces which produce an equivalent elastic length change/curvature to the time dependent effects, is that right? Hence we do not consider them at ULS.
Answer: Yes, Creep and shrinkage are imaginary forces that produce equivalent elastic length change to time dependent effects. So, we need not to consider them at ULS combinations.

Question2: For shrinkage, Midas takes the notional size h as the only geometric input for a beam element. This is defined as a property of the material, which is a little counter intuitive. If we use a composite section for construction stage analysis, we can enter another h for each part of the section - does this override the notional size in the material definition?
Answer: As shrinkage strain depend on the notional size of section hence, it is defined in shrinkage tab but generally, notional size value is generally updated using the properties>Time dependent properties>Change Property tab. Using this tab, software will auto-calculate the notional size of the selected section and shrinkage strains are calculated using this updated notional size.
Moreover, the value defined in composite section for construction stage analysis will override the notional size defined earlier.

Question 3:The shrinkage secondary My in the model is enormous for the precast beams. long term construction stage - 2760 kNm. This is a similar magnitude to the dead load bending which intuitively seems completely wrong. Similarly the reactions at the base of the piers, which are up to 1460 kN. Please could you advise if it is correct? And if not, have we defined a model property incorrectly, or would this force actually be relieved by shrinkage cracking in the real structure? Or by the casting sequence of the top slab allowing for autogenous shrinkage before the next pour.
Answer: I reviewed the model. The time dependent properties were provided to dummy deck elements which is not correct, as deck is In-situ 40/50 material which is already linked to time dependent properties. Dummy are just for load transfer between 2 girders and thus should not be linked with time dependent properties.
Similarly, kindly remove the time dependent property on substructure and end diaphragm. Giving time dependent properties to substructure will not create massive effect as there is no restraint in longitudinal direction of substructure.  For a each diaphragm element, the rigid elastic link are used at bottom. Shrinkage will try end diaphragm to shrink but rigid link will not allow thus causing the unusual forces in end diaphragm.

Lastly, update the value of h manually  in composite section of construction stage for girder part. It's value is 10m which is very high. It is generally 2A/u is 0.2-0.3, where A is area and u is exposed surface perimeter. Midas don't calculate the h value for general section

When all these values are updated, the shrinkage change significantly.

Question4 : Similarly to thermal, the stiffness of restraint must have a big effect on the size of the shrinkage forces. Do we need to account for this flexibility somehow - eg the bearings used on the structure will not be rigid against uplift.
Answer: No, Normal rigid bearing should be used on the structure against uplift also because when the structure is lifted like due to shrinkage force, then at that time, structure is already in compression due to dead load. Thus, uplifting will cause the decrease in compression caused by dead load but if uplifting force is neglected, then it will produce wrong results.

Hi,

I have a further question. Why should shrinkage not be considered on the transverse members? concrete shrinkage is a 3D effect so will cause strains in both directions of the slab plane. Do the dummy members not need to consider this? Similarly the diaphragm will be affected. As one side of the deck is released at the plinth members to simulate guided/free bearings, there should be no unusual forces against the rigid links?
Question: Why should shrinkage not be considered on the transverse members? concrete shrinkage is a 3D effect so will cause strains in both directions of the slab plane. Do the dummy members not need to consider this?
Answer: This is general engineering practice that we calculate shrinkage in longitudinal direction only.
Suppose there is box girder,for calculating the shrinkage we require notional size which we can easily based on cross-section but when transverse shrinkage is observed, the notional size is not standardized across the width of girder. Therefore, we neglect the shrinkage in transverse direction.

Question 2: Similarly the diaphragm will be affected. As one side of the deck is released at the plinth members to simulate guided/free bearings, there should be no unusual forces against the rigid links?
Answer: For diaphragm we can consider the shrinkage.
Let's see the for diaphragm element 11693 and 11592. Elastic link number 95,96 and 61 is rigid elastic link. When shrinkage occurs in 11693 and 11592 elements, the rigid elastic link 61 being rigid in nature won't allow it to shrink and hence huge internal forces will be developed in diaphragm due to this restraint which is not true.

Hence, if you need to provide time dependent properties, then rigid link can be used where master node is centre of diaphragm and other becoming slave node rather than rigidly connecting each two nodes (which is making each diaphragm element rigid.)

Hope this helps. Kindly let me know if further assistance is required.

Creation date: 5/31/2021 4:16 PM      Updated: 5/31/2021 4:16 PM