Calculation of prestress losses:Approach

[cecil123]

I am carying out a design of a post-tensioned floor - a flat slab using hand calculations first.
Now, I have assumed a slab depth, a reversed parabolic profile, and calculated the prestressing force required to balance the dead load.

I obtain the number of tendons in each span, ASSUMING A 20% LOSS OF PRESTRESS AT SERVICE.This is followed by computing the stresses at service,computing mild steel (if tensile stresses exceed permisable limits), computing ultimate moment capacities,comparing factored and ultimate moment capacities,provision of mild steel if required for ultimate state,checking for shear and COMPUTATION OF PRESTRESS LOSSES.

Now, my question is if prestress losses exceed 20% say 30% at service, then all calculations have to be repeated , right?In the second cycle now, we again calculate the losses and compare if they are within 30%, right?

So we continue this process until we get reasonable convergence between losses calculated and the losses intially taken,right?

Is this the procedure followed in the commercial software too?

What should be a reasonable convergence?(i.e. the difference between losses compted and losses used in calculations??)

 

[cecil123]

Can the steps be listed as below:

1)Assume slab depth
2)Assume prestress loss per tendon and get the force per tendon
3)Calculate the number of tendons based on percentage of load to balance
4)Then,calculate the losses
5)If losses exceed the assumed loss, again calcualte force per tendon, and again number of tendons
6)Again calculate losses, reperat with step 5 till assumed losses converge with actual.
7)Get the service stresses
8)Complete design with other steps.

Am i right? Then,

Suppose I have "n" tendons, should I calculate the loss for each tendon in that span?The losses would vary depending upon which end of that tendon is dead end and which end is stresing end.Right?
Please help, awaiting your response

 

[AUCE98]

Typical effective force for a 1/2" diameter grade 270 low-lax seven wire strand is 27 kips,which equates to approximatly 176.47 ksi. Some engineers prefer to assume 26.5 or somewhere in that range. These values would be valid for a single end stressed tendon 100 feet or less in length. For a tendon longer than that provide a double end stress on the tendon, or perform a friction elongation calculation. The friction loss calc will provide you with your immediate and long term stress losses.

 

[cecil123]

In post-tensioned construction, it is a common practice to bundle several unbonded mono-strand tendons together.
But,would not this result in poor consolidation of concrete around this bundle?

 

[christyleomin]

This is because, there is an increased potential for delamination at high and low profile points as more strands are bundled together.

 

[cecil123]

Dear Christyleomin,

Thanks for the reply.But, I do not understand when you say "there is an increased potential for delamination".
What do you mean by increased potential for delamination ?

 

[pappyirl]

Why are you assuming a prestress loss? Why are you assuming it is the number of tendons and not the profile which will balance the dead loads? Both are equally important in performing this role, you will have to iterate the two values to optimise the problem.

I think what christyleomin is trying to say is if there are a lot of tendons in a horizontal plane, there is little bond between the concrete either side of the tendons at the high and low points. This might cause the concrete to peel away from the tendons in service and "delaminate". Hence the need to bunch the tendons together.

There is a small loss depending on which end is live and which is a dead end. A greater loss occurs depending on which tendon is stressed first, elastic shortening of the slab on subsequent stressing of tendons increasing the loss of the first stressed tendon. It is generally assumed that the elastic loss is applied to all tendons, giving a worst case loss.

 

[dik]

I did a project about 40 years ago and the strands were about 350' long and I had stressed them from both ends because of the length. Strand was covered with poly sheathing and with the overstress, the friction loss was not measureable when stressed from the far end (humbling experience #47). With reasonable cable lengths and poly sheathing, friction loss is next to nil... Confirm that friction loss with poly is next to nil...

Dik