Restrained prestressed slabs. 1

[Tomfh]

I have seen many structures where large prestressed slabs are restrained by walls or other rigid objects (e.g. basement pile walls, stiff columns, stair walls, lift cores etc). This is obviously not theoreticallly a very good thing, nor so great in practice, but yet I've seen it enough times in my limited career to know it must happen all the time and that many designers and contractors mustn't give it as much thought as they probably should. When I've seen it typically the slabs or support structures form their own "movement joints".

Aside from unwanted cracks what are the real world consequences? How is strength affected, how is deflection affected? What cases have you seen? What was done in response to such issues?

 

[asixth]

For post-tensioned slabs on basement walls I provide a slip joint on top of the wall. Generally a bond-breaker material and PVC sleeves over the wall reinforcement so when the slab is stressed movement is allowed and the prestressing force goes into the slab. I normally specify that the PVC sleeves get grouted up after the slab is fully stressed but some engineers nominate that they get grouted after 1 month and some wait 2 months. Similar details for cores when there are two or more cores that could restrain the slab, there are propriety products like lockable dowels (by Ancon) that can be grouted after the slab is stressed.

A big issue could be that the prestress doesn't get into the slab and that it ends up in the walls. The could result in higher deflections because you don't get to balance out as much of the load as you were hoping.

 

[KootK]

Wall restraint won't affect load balancing as that only depends on drape and tendon force, not concrete prestress. The loss of prestress will still affect deflections and strength, however, because less concrete prestress can mean more slab cracking and lower slab moment of inertia at service load levels.

Other than unsightly cracking, I mostly worry about the impact on the vertical elements. You can inadvertently impose an unintended shear demand on your walls that rivals the original design loads.

I've worked on a few buildings that were large in plan and had shafts at the far ends. The contractors poured half of the floor plate a few stories ahead of the other half. The strategy is similar to delay strips in mildly reinforced concrete. It doesn't solve all your issues though.

You may find this document useful: Link

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[slickdeals]

Not sure why you say wall restraint won't affect load balancing. Restraint would result in loss of precompression in a bay = less prestressing force = less load balancing.

 

[KootK]

Loss of precompression in a bay <> less cableprestressing force.

You still have the same cables, the same force in those cables, and the same drape. Therefore, the load balancing is the same.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[KootK]

Your cable force is probably higher due to restraint because intended losses due to elastic shortening and creep are also prevented to some degree.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[jike]

In the 40's and 50's restraint of movement in prestress (precast and cast in place) structures was not well understood. As a result, it would break the support or the anchorages or the welded connection (although the weld was usually much stronger than other components. There are now standards ways of preventing these failures. Some of these failures were pretty serious in that a loss of bearing support could result.

 

[rapt]

Kootk is correct in that the load balancing is not reduced. But that is not of much assistance in stopping cracking or reducing deflections.

Asixths solution is helpful in that it gets the prestress compression into the slab. Unfortunately, that is far less than the tension stresses that could be induced by restraint to shrinkage or temperature change shortening restraint. So while it will help to reduce the effects of cracking it will not stop the cracks and you will normally end up with some very wide cracks, as prestress tendons only provide good crack control until the concrete cracks. After it cracks, unbonded tendons are basically useless for crack control and bonded tendons are too far apart to offer good crack control (max 300 centres required compared to 1m - 1.5m spacing). The only real solution is to add extra untensioned reinforcement in both faces at a maximum of 300 and preferable at 200mm centres or less.

The best solution is to check the cracking under combined bending and direct tension loading and determine crack widths.

Even the AS3600 rule for restrained slabs should require the maximum spacing of bonded steel of 300mm, even where sufficient prestress has been added to solve it with prestress alone according to the formulae (hopefully this limitation will find its way into the next version of the code).

 

[Ingenuity]

And since restraint does reduce the P/A within the slab, your code-specific punching shear capacity for flat plates and slabs will be reduced accordingly - assuming you neglect membrane actions. You will still have the vertical component of prestress (usually very small) to assist in punching capacity.

 

[slickdeals]

Ok, just so that I can get corrected if my understanding is wrong.

What I meant in my post is that a slab with restraint will need more precompression compared to a slab without significant restraint. Load balancing is a function of the force and drape. If I had a slab with minimal restraint and balancing 90%, addition of restraining walls will reduce my load balancing as the force is sucked up by the walls and not delivered into the slabs.

 

[KootK]

This might just be a semantic thing Slick. Do you agree that balancing load is the upwards force exerted on the slab by the cables and nothing else? That is the definition that I'm used to.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[Ingenuity]

If I had a slab with minimal restraint and balancing 90%, addition of restraining walls will reduce my load balancing as the force is sucked up by the walls and not delivered into the slabs.

The axial pre-compression into the slab (the P/A) will get "sucked up by the walls" and reduce the P/A into the CONCRETE slab, BUT the FORCE on the PT tendon will not change (well it will due to losses, but let's neglect for this discussion) so the balanced loads are NOT effected by restraint. Think of it by isolating the tendon from the concrete - the prestressing force from the tendon imparts 'equivalent loads' into the concrete, both axial (P/A) and transverse (due to tendon curvature). The 'transverse load' (in your case the 90%) of the tendon/s within the span (the 'meat' that does the load balancing) is only dependent on the TENDON force and the ANGLE change (i.e. drape). The prestressing tendon FORCE does not change, so the magnitude of the load balancing does not change due to restraint. BUT, the restraint actions of slab/wall/columns etc will reduce the P/A that the slab.

 

[Ingenuity]

This might just be a semantic thing Slick

True.

 

[hokie66]

I think the introduction of load balancing into Tomfh's original question has obscured the good answers given. Read rapt's post. The problem of restraint shrinkage cracking is not just for PT slabs. It is for all slabs, just often worse in PT slabs due to lack of crack control reinforcement.