Prestress Losses

[marc.rogue]

Hi I have come across different design software and most of them use PCI design method to estimate losses which is essentially picking a critical section point and estimating the losses there (critical points could be .5L,.4L or the transfer length 50db) But recently I came across a software that checks stresses at multiples points and uses different losses for every point not always the same. At a glance it seams reasonable since some losses are a function of distance (X) but I was wondering what other engineers think about this method?

 

[r13]

The important thing here is do the programs/methods yield consistent results. I am curious on that if more points taken are better for design? I remember the total loss is more of the concern, sure, I could be wrong.

 

[marc.rogue]

Hello there,

Actually what is happening is that the software calculates losses at every point it checks. for example calculating elastic shortening
Kcir*(P/s+Pe^2/I-Mg*e/I) this equation determines the stress in concrete surrounding the strand
1 eccentricity for a draped strand is different at all points
2 Mg also changes with X towards the middle of the span
This makes it seem possible that you could estimate losses at every point but how realistic is this really?
The final results seem to be accurate enough to work with

 

[r13]

I don't get the benefits you are trying to convey. But as long as the results on the main points are comparable/conforming, then it is good to go.

 

[marc.rogue]

Benefits are many but one of the biggest ones Ive noticed is that it greatly reduces end stresses which are usually a concern for design.

 

[rapt]

Considering that the loss effects are different at every point in the member, they should be calculated differently.

Friction,creep and elastic shortening are definitely different at every point, shrinkage would be different if the section shape changes.

Early PT software in the USA treated it as an average, so that became the norm there.

I have been designing PT for over 40years and have always considered the variation along the member, even before computers became involved.

 

[marc.rogue]

I see, Hey rapt do you have any experience in the effects of elevated shoring on prestress members?

 

[JoshPlumSE]

I'm not expert in PT, but when I was in school, we calculated losses along the entire span. Granted our class was entirely focused on bridge work / Caltrans projects. So, maybe there's a difference between what the bridge folk do and what the building guys do?

 

[rapt]

marc.rogue,

You would have to be a lot more specific with your question!

 

[marc.rogue]

ok so, some jobs call for an elevated shoring as part of the design. Meaning a literal post shore is used to push the joist up to a certain height to literally induce additional capacity into the joist. For effects of analysis should this be considered as just a point load acting upwards on the joist? Mind you it also affects final deflections. I believe I have a pretty good grasp on how it all works but it never hurts to double check

 

[Ingenuity]

So, maybe there's a difference between what the bridge folk do and what the building guys do?

For PT in buildings within North America there is. The majority of UNbonded PT building designers simply use 175 ksi as effective stress in the strand (after all losses) and adopt 26.8 kips per 1/2" dia strand (175 ksi * 0.153 in[sup]2[/sup]) and use that along the tendon length. Similar to rapt, I always consider the variation of the tendon force along the member.

 

[r13]

Ingenuity,

Although the basic concept for pre-stressing and post tensioning is the same, but seems there are some different concerns in practice. What is the use of variation in tendon force along the span, for instance? By the way, I've no training on PT. Thanks in advance.

 

[Ingenuity]

PRE-tensioning and POST-tensioning are the variants of "prestressing".

Friction losses, for example, in post-tensioned construction, accumulate due to intended tendon curvature (drape) and unintended curvature (wobble) and the result is that the tendon force is not constant along the member. When the tendon is too long (say more than 120 feet) then double-end stressing is used to reduce the effects of friction losses. The argument that unbonded PT designers use to justify the average tendon force of 26.8 kips per strand is that due to the greased and sheathed tendon the strand is able to slip sufficiently that friction losses equalize over the tendon length.

When checking service level stresses along the member at discrete (or critical sections) you typically use the prestress force applicable at that location to determine such resulting stressing. For multi-span members, with significant drape, and if only single-end stressed, the uplift forces (due to prestress) of spans closer to the dead-end anchorages will be reduced, and that can effect deflections.

 

[r13]

Thanks for the correction. I think friction is more of a concern in PT due to geometry, unlike pre-tensioning, usually deal with one piece of straight member, so it is easier to estimate the loss.

 

[rapt]

From what I understand many US designers also use long term losses calculated by a method from Zia et al which is a very approximate guess and does not vary along the member as it should, especially for creep.

In pretensioning there is a large immediate end loss due to its transmission length. The force drops to zero force about 6 * diameter from the end of the bonded length of the strand and does not reach full stressing force until about 60 diameters from the end. And this force is not increased under tensile strain. So it is very important to consider the loss in the calculations and in determining bottom tension forces developing into the support, as often with pretensioned decking units it will be close to zero contribution from the pretensioned strands depending on the end support detail.