RAPT - Net Column Stiffness - Long Term Deflections

I have been comparing some long term slab deflection results using both RAPT and SAFE using very simple examples, and have found the results vary quite a bit when I take the equivalent column stiffness approach in RAPT as opposed to the net column stiffness. Are there any recommendations or restrictions on using the net column stiffness in RAPT?

I am unclear on what the applications would be for using net column stiffness. I gather it would be in situations where you have no transverse torsional stiffness, like in a concrete portal frame (where the manual recommends you to use enhanced column stiffness)?

Incidentally, using the equivalent column stiffness approach for a simple three span 8x8m flat plate building, my long term deflections come out more than 50% more than in SAFE.


First, Read the manual. You should be using equivalent column stiffness for any normal building frame.

Second, Net column stiffness option was included simply to give you a way of comparing to 2d frames analysis software. It was never meant as a realistic design option.

I would have thought the information on Equivalent Column given below says it all!

The Equivalent Column Approach. [See T.2.1 Section Theory T2.1] Also takes into account the infinite stiffness of the slab / beam at the column interface. This is the default and should be used for most slabs and beams in building floor systems.

And why would you think SAFE’s answer might be more correct.

RAPT has been doing dong term deflection estimates since the mid 1980’s and its results have been very well respected for that time. Comparisons with deflections in existing buildings have been very good. SAFE has never been know for any reliability in its long term deflection estimates, which it introduced several years ago. Their early versions were very inaccurate from what I have been told. And as I understand it their results are very dependent on user input.

If your answer is close to 5 - 6 times short term uncracked deflection (assuming logical L/D ratios) then they are about correct.

Above is a snippet.

@rapt, Yes, reading the manual is always a good start. Hah