Deflections and creep

7 posts / 0 new
Last post
gharrell
Deflections and creep

Please confirm the creep coefficient for Dcreep3 is being computed consistent with the formulae reported beneath the tables in the Details Report. I can replicate values for Dcreep1 and Dcreep2 for both minimum and maximum timing, but I have to compute Y(2000,300) and Y(2000,600) instead of Y(2000,40) and Y(2000,120), respectively, when computing Dcreep3 using the Dbarrier and Doverlay values reported. Thank you.

Rick Brice
Please post your PGSuper

Please post your PGSuper project file and I'll take a look at it.
Also, please specify which girder you are analyzing so we can compare apples-to-apples.
Rick

gharrell
File attached sa requested

Girder B is the girder for which I was making the comparison. Thank you.

Gregory Harrell, PE
HNTB Corporation

Rick Brice
Dcreep3 is computed

Dcreep3 is computed consistent with the formula given in the Details Report.
The formula is:
Dcreep3 = [Y(2000,1) - Y(40,1)](Dgirder + DpsLs) + [Y(2000,30) - Y(10,30)](Ddiaphragm + Duser1) + [Y(2000,40)](Dbarrier + Doverlay + Duser2)
Using your input file and a Details Report for Girder B, PGSuper reports the following values:
Y(2000,1)  = 1.19771
Y(40,1)    = 0.660135
Y(2000,30) = 0.660505
Y(10,30)   = 0.180925
Y(2000,40) = 0.638459
Dgirder    = -3.239 in
DpsLs      =  5.606 in
Ddiaphragm =  0 in
Duser1     =  0 in
Dbarrier   = -0.302 in
Doverlay   = -0.379 in
Duser2     =  0 in

Dcreep3 = (1.19771-0.660135)(-3.239+5.605) + (0.66050-0.180925)(0) + (0.63849)(-0.302-0.379) = 0.837 in
Dcreep3 reported by PGSuper is 0.837 in.
You can change the creep timing parameters in the project criteria library.
 

gharrell
Strength and Time Development factors for creep

Rick -

I agree the value for Dcreep3 follows the formula. The value of the creep coefficients is what has me puzzled. All coefficients computed based on Tinit equals 1 are spot on. When Tinit equals 30 or 40, we differ. By changing fci to fc in my calculation of ktd and kf, they agree with PGSuper for Y(2000,30), Y(10,30), and Y(2000,40). My interpretation of these factors, as explained in NCHRP Report 496, is that they are based on the initial concrete strength at transfer, not at the time the individual loads are applied. Therefore, my sense is that changing the strength term in these factors from fci to fc for coeffiecients with initial times other than 1 actually recharacterizes the concrete. Please comment on this interpretation and/or provide reference to support how PGSuper currently treats the computation of these factors. Again, this is my interpretation, and I am certainly open to opinions/thoughts to the contrary.

Thanks,
Greg

Gregory Harrell, PE
HNTB Corporation

Rick Brice
Greg, I've been reviewing

Greg,
I've been reviewing NCHRP 496 as well as recent PCI Journal articles on the topic all afternoon. I cannot find anything that would justify using f'c in the ktd or kf parameters.
I was using f'c when the loads where applied at intermediate stages. In these cases the concrete will creep due to these loads based on the stiffness of the concrete at the time the load was applied. While I think this is sound reasoning, it does not appear that the analytical model provided in the LRFD was created based on this idea. The model in the LRFD is clearly based on the concrete strength at release. The age of the concrete is modeled with the age adjustment factor X=0.7.
I did notice that 0.8f'c for the deck concrete should be used when compute kf when computing the deck shrinkage strain. Refer to page 27 of NCHRP 496. The example on page 30 demonstrates the calculation. The language in the LRFD isn't clear, but it you can tell what the intent is.
I have made the necessary changes to the software and it will be available in the next release.
Attached is a Details Report for Girder B using the updated software. I would appreciate it if you could review the output and compare it with your analysis. Thanks
Rick

gharrell
Agreeement

Rick -

The creep coefficients (and camber values) presented in the posted file are in agreement with what I'm calculating. Thank you for taking the time to look into this issue.

I agree that the right side of the page in AASHTO could stand to elaborate a bit on the assumptions and intentions, without making one dig up 8-year-old research. Maybe in an interim revision, as if they don't already come out frequently enough. :)

Thanks again,
Greg

Gregory Harrell, PE
HNTB Corporation

Log in or register to post comments