Prestress Losses using Transformed Section Properties

6 posts / 0 new
Last post
gavinbv47
Prestress Losses using Transformed Section Properties

I'm currently trying to perform a comparison between stresses using transformed section properties and gross section properties. What I have found so far is that transformed section properties provide less Service stresses (as one would expect), but also lower effective prestress stresses, the difference between those two resulted in higher stress demand for transformed sections compared to gross section properties.

I took a closer look at the prestress losses detailed report and noticed that using transformed sections resulted in bigger prestress losses compared to gross sections. I then went through AASHTO BDS 5.9.3.2a equations and just have a gut feeling that you should have less prestress losses if you use transformed section properties. On top of that, C5.9.3.2a and C5.9.3.3 seem to suggest that elastic deformations (instantaneous effects) need not be accounted for if using transformed section properties. Am I missing something or interpreting the code incorrectly?

Note: With lower effective prestress and higher section modulus, the prestress contribution to the girder stress is even less. It seems like we are doubly penalized?

PGSuper Version: 4.1.3.0 (x64)

Thanks,
Gavin

Rick Brice
Prestress Losses using Transformed Section Properties

I don't think you are missing anything. When using transformed section properties, instantaneous effects due to elastic deformations are implicit in the computations, so you don't explicitly include them as you would with gross property analysis.

Your findings with respect to the relative magnitude of time dependent losses and effective prestress force are consistent with PCI BDM Examples 9.1a and 9.1b. These examples analyze the same bridge, with Ex 9.1a using transformed properties and Ex 9.1b using gross properties. The time dependent losses with the transformed properties analysis is greater than the gross section analysis (43.9ksi vs 42.7ksi) and the effective prestress is slightly less with the transformed properties (165.8ksi vs 167.8ksi). Keep in mind that the loss equation in the LRFD are based on transformed section analysis as described in NCHRP Report 496. Explicit elastic gains were added by AASHTO is to compensate for the fact that the loss equation and method was changed to gross section analysis. The losses computed with transformed and gross properties (with appropriate elastic gains) are essentially the same, given the accuracy with which we can compute losses.

As you point out, transformed properties provide an advantage when computing service stresses. A comparison of transformed and gross section analysis is provided in "Evaluation of common design policies for precast, prestressed concrete I-girder bridges", Brice, R., Seguirant, S., and Khaleghi, B., PCI Journal, Fall 2013, pp 68-80. http://www.precastconcretejournal.org/i/297255-fal-2013/69?m4=

gavinbv47
Rick - Thank you for the

Rick - Thank you for the response! The references provided are a big help.

gavinbv47
Apologies for re-opening the

Apologies for re-opening the topic.

As you have stated in your response, as well as in PCI BDM 9.1a.6.6 and AASHTO BDS C5.9.3.3, losses due to instantaneous effects are implicitly accounted for when using transformed sections, thus they should not be explicitly included when calculating concrete stresses. I further investigated stresses due to prestressing between gross and transformed section properties by hand vs. results from PGSuper. It seems that PGSuper is explicitly includes losses due to instantaneous effects when it is calculating girder stresses due to prestressing when using transformed sections?

Attached in the updated Original Post are the PGSuper files and some hand calcs that I ran through. Ag,trans, Sbg,trans, and ep,trans values were taken from PGSuper Details Report.

Again, apologies if I'm missing something basic here.

Thanks,
Gavin

Rick Brice
stress calculations

Gavin,
Thanks for getting back on this. There appears to be a bust in the software for girder stresses computed with transformed section analysis when live load is included.

The stress due to prestressing is

f = (Aps)(fpj - dfpLT)(1/A + e/S) where dfpLT is the long term time dependent effects and A and S are transformed non-composite properties.

Alternatively,
f = (Aps)(fpj - dfpLT - dfpE)(1/A + e/S) where dfpE are the elastic effects and A and S are net non-composite properties.

As you pointed out, PGSuper is incorrectly including the elastic effects.

I'll get this fixed up ASAP.

Rick Brice
stress calculations

BridgeLink 4.1.5 has been posted. It fixes this issue.

Log in or register to post comments