Base Stiffness and AS4100

[asixth]

How are others treating Base Stiffness when designing portal frames for drift. I know BS5950 Part 1 Clause 5.1.2.4 reads:

"In the analysis of all frames, the same base stiffness about the axis under consideration should be used for all calculations. In the absence of detailed knowledge of the foundation stiffness the following should be assumed.

a) Where the column is rigidly connected to a suitable foundation the stiffness of the base should be taken as equal to the stiffness of the column, except as in 5.7.3.1.

b) Where the column is nominally connected to the foundation, a base stiffness of 10% of the column stiffness may be assumed.

c) Where an actual pin or rocker is provided the base stiffness should be taken as zero."

I cannot find a similar clause written in AS4100. This opens the theory up to a few questions.

1. "In the absence of detailed knowledge of the foundation stiffness" implies that soil-structure interaction should be modelled wherever possible. But I have never been supplied with a geotechnical report that adequatley descibes what properties should be used to model soil-structure interaction.

2. A full-moment fixity should never be modelled at the support, but 4*E*I/L should be used as a rotational spring stiffness instead. This can greatly redistibute moments. For example, I have a 6 meter high 460UB column which forms part of a portal frame and experiences a base shear of 10kN. When no rotation is allowed at the base, the base moment transferred to the footing is 38.8kN-m and the total drift is 5.7mm. When modelling a rotational stiffness at the base of 39,500kN-m/rad, the base moment is reduced to 33.0kN-m and the drift increases to 10.0mm as footing rotates 0.00084 radians. This increases the drift by close to 200%.

3. I have also heard that under service winds (25yr return), pin connections can be assumed to have a rotational stiffness of 0.2*4*E*I/L provided the baseplate has a minimum of 4-anchor bolts and it can be proven that the foundation can resist this moment. For the example I described above, this reduces the base moment to 20.6kN-m and increases the drift to 15.9mm.

Basically, the question that I am asking is "what procedures are others following to calculate drift for moment frames".

Apologies for the length of the post. All input is welcome.

 

[rowingengineer]

asixth,
No need to apologise for giving all the relevant information, we struggle to often in this forum with people not supplying enough information.

I have a few answers for you, none of which match you questions;
1. You can certify to the BS code if you like, the certifiers (I use) will accept this, as I do it on almost every job.
2. I use some general rules of thumb to get my sizing right, and then I design my base plate/bolts and footing accordingly. Using the design manual attached to ensure my design is correct. This procedure is based on the component method; I prefer this for semi-rigid connections, makes logical sense to me.
3. Not getting enough information for foundation design is always a problem. However you must include the rotation for the pad/pier footing. The slab will help you here if connected to the footing; however this isn’t the case to often with large portal frames.


Arguing with an engineer is like wrestling with a pig in mud. After a while you realize that they like it

 

[OzEng80]

Asixth

I have had similar struggles and look forward to the responses to your questions. In the past I have ignored base stiffness completely and have only recently taken to applying a nominal base fixity as per information provided in another post (which I assume has prompted your questions).

It would be very interesting to see whether this load condition is being considered in base plate/HD bolt design in practice by the industry in general. I see a lot of 4 bolt connections with hold down bolts positioned close to the flanges and clearly sized as ‘pins’. I imagine bolt forces from the nominal base fixity and uplift in the serviceability limit state could exceed these designs.

Look forward to reading your document rowingengineer.

 

[apsix]

From the current BS5950-1:2000
5.1.3.3 Nominally pinned base
If a column is nominally pin-connected to a foundation that is designed assuming that the base moment is zero, the base should be assumed to be pinned when using elastic global analysis to calculate the other moments and forces in the frame under ultimate limit state loading.
The stiffness of the base may be assumed to be equal to the following proportion of the column stiffness:
a) 10 % when checking frame stability or determining in- plane effective lengths;
b) 20 % when calculating deflections under serviceability loads.

By this clause base stiffness should be ignored when analysing for strength (except when determining effective length), but 20% can be used when analysing for deflections.

As it says in clause 5.1.3.1; " In the absence of detailed knowledge of the stiffness of the base, design may be based on the assumptions detailed in 5.1.3.2, 5.1.3.3 and 5.1.3.4."

RE; how can you certify to BS if they aren't referenced in the BCA?
I treat foreign standards as well researched text books, as long as it doesn't non-conservatively contradict AS codes they can be used for guidance.

 

[OzEng80]

I don't think there is anything specific in AS4100 prohibiting any of the assumptions mentioned. AS4100 states that analysis can be based on Rigid, Semi rigid or simple connections - my opinion is that the classification and assignment of values to these is based on engineering judgment (which can encompass other codes). Unless you are contradicting or designing for less than what is prescribed in AS4100 I don't think justification would have to put on the certifications.

 

[rowingengineer]

Yes you guys are right; my comment was flippant at best. If we were designing a base plate using semi-rigid design, then you probably wouldn't have to state this on your certificate as it would still comply with AS4100.

However, on some of my certificates I state other codes because I use them for significant proportion of the design. This is the case for wind turbine towers and footings (which have to be certified in accordance with the BCA), FRP structures and other structures. For these to comply I use the clause A0.8 and A0.9 of the BCA.

Arguing with an engineer is like wrestling with a pig in mud. After a while you realize that they like it

 

[asixth]

I was speaking with some colleagues in my office about this last week and the general consensus was that base fixity of columns should be taken into account when checking service drifts but there was not an exact answer put to it such as 0.2*4*E*I/L as proposed in British Standards.

One colleague in particular would model the structure twice, one with no rotational restraint at the base and one with full rotational restraint at the base, and they would average the service deflections from both models when checking service drifts. It just so happened that modeling a base rotational spring stiffness of 0.2*4*E*I/L gave a very comparable answer to this approach (Pinned=27.4mm drift; Fixed=5.7mm drift; 0.2*4*E*I/L=15.9mm drift).

As much as I would like AS4100 to give a specification on what base stiffness should be assumed, I also know that the responsibility is on the engineer to confirm whatever stiffness is assumed is a valid assumption. I am putting together a spreadsheet now for standard baseplate designs that can calculate how much moment the base fixing can transfer so I am going to continue with the assumption that 0.2*4*E*I/L is a valid engineering assumption.

It also creates a problem modelling the structure in SpaceGass because I can only model the structure for one set of restraints without copying the model. It would be nice if I could analyse the structure for separate restraint cases, e.g. model the structure with rotational base stiffness for 25yr wind events and no base rotational restraint for 500yr events. Probably an item I best put to the software developers.

RE

Was that a document you wrote?

 

[csd72]

I do not believe there is anywhere in the code that says you need the same model for both Ultimate Limit sate and for deflection.

We used to design the base as pinned for limit states and then add a serviceability restraint moment of 10kN.m at the bases for serviceability. Then you can by superposition calculate what restraint force you need to limit the deflection to a suitable level. We used to limit this restraint force to a maximum of 25 to 50kN.m to ensure it is within stiffness limits and then we chacked that the spread on the 4 bolts was sufficient to take it.

Believe me, I used to think that an engineers life would be easier with more code clauses and guidance but the reality is actually quite the opposite - there are more things to get wrong! By all means refer to reliable sources for the design method but be thankfull than people are not telling you exactly how to do your job.

 

[rowingengineer]

I believe you can thank
František Wald and his team from Czech Technical University, Faculty of Civil Engineering, Prague, Czech Republic for the publication.

They have been developing the Component method for steel column bases for a while now.

Arguing with an engineer is like wrestling with a pig in mud. After a while you realize that they like it

 

[apsix]

I don't think anyone is claiming that a base stiffness of 0.2*4*E*I/L is an exact answer.
It's a deemed to comply rule, presumably conservatively based on testing or refined calculation.

Of course we all know that deemed to comply rules aren't always conservative.
However, in it's current form it's been recommended by the SCI since 1991, and adopted by the BS in 2000; that's good enough for me unless shown otherwise.

 

[csd72]

There is no such thing as an exact answer!