Accuracy of portal method to analyse multi-story building for a real life project. 4

[Pretty Girl7]

People compare the the results of "portal method" of frame analysis with the modern software like "SAP2000" and "ETabs". They show the portal method is not accurate and has an error percentage of like "30%". It might be a bias to sell their software.

I wanna know, if I use "portal method" to analyse a multistory building with manual shear and moment calcualations (I know it takes a long time) and with Eurocode, will the building be safe?
What kind of structural analysis methods are accurate and can be used if I really don't want to use any software packages, to design safe buildings and structures?
Being economical is not my concern, but the buildings I design should be safe.

 

[Lomarandil]

Largely, the answer will vary depending on the complexity of your structure.

If the building is regular, the portal method does fairly well. I've designed a number of multi-story buildings "by hand" using it with some comparisons to FEA, and would estimate the accuracy around +/-15%. In my case, that was well within the precision of other parts of my design, so I'm confident the design is safe and not significantly un-economical.

If the building is not very regular, very architecturally driven, then I believe the claims of portal method errors around 30%.

 

[KootK]

The portal method relies on assumptions regarding the locations of inflection points. As such, the accuracy of the method is dependent on the accuracy of the assumed inflection point locations. To a large degree, one can usually assess that by inspection. If the complexity of the structure is such that you're not able to do that, it may not be the right method for you.

The Portal Method and Cantilever Method seem to be the approximate methods that have sort of survived the onslaught of computers. But there are others, including the Spurr Method and the Muto Method. One nice feature of the Muto method is that, at the least, it always satisfies equilibrium, if not elastic displacement compatibility.

If I were to attempt to design anything remotely complex by hand -- and I wouldn't -- I would probably go with the Moment Distribution Method which is 100% accurate given the right assumptions and execution. Prior to Moment Distribution, a lot of folks were using Virtual Work Methods for indeterminate structure analysis.

Employed strategically, virtual work can sometimes be more computationally efficient than Moment Distribution. It has the disadvantage of being kind of a "spot check" method however. With moment distribution, you have to do a whack of tedious -- yet simple -- accounting but, at the end of it, you've got a relatively complete picture of what's going on all over your structure. I suspect that it was this combination of simplicity and completeness of output that led moment distribution to largely supplant virtual work.

 

[Pretty Girl7]

If the building is regular, the portal method does fairly well.

would estimate the accuracy around +/-15% ... if the building is not very regular, very architecturally driven, then I believe the claims of portal method errors around 30%

The 15%-30% error in portal frame means the applied safety factors of 1.35, 1.50 is almost gone. It's like the building is designed with no safety factors added. So, I don't understand the fact you're saying that "the portal frame does fairly well" etc. How come a building designed with no safety factors "does fairly well"? just asking since It doesn't make sense to me.

 

[Pretty Girl7]

Moment Distribution Method which is 100% accurate

Isn't the moment distribution used where lateral loads are not there? how about a case where wind acting from sideways on a wall of a multi-story building?

 

[KootK]

Moment distribution is a valid method for any loading condition.

 

[Pretty Girl7]

Thank you

 

[Lomarandil]

The 15%-30% error in portal frame means the applied safety factors of 1.35, 1.50 is almost gone. It's like the building is designed with no safety factors added. So, I don't understand the fact you're saying that "the portal frame does fairly well" etc. How come a building designed with no safety factors "does fairly well"? just asking since It doesn't make sense to me.

If the building were designed to 100% utilization, yes, the safety factor would be reduced.

In the real world (except PEMB manufacturers and aerospace), it is rare that a design pushed to 100% is the most economical and constructible solution. Even when certain elements are pushed to their limit, redundant load paths, ductility, etc make many structures safer than the sum of their individual parts. And then on top of that, designers tend to calibrate their "target utilization" to the precision of inputs and analysis used.

All that to say -- if my design calcs out at 80%, and my analysis method is +-15% accurate, then I haven't yet dipped into the safety factor.

 

[271828]

Several error percentages have been quoted. Errors compared to what standard? Elastic frame analysis isn't exactly "exact." LOL

 

[Settingsun]

I'm curious why computer analysis is outlawed?

 

[271828]

I'm curious why computer analysis is outlawed?

Me too.

I like the Portal Method, Moment Distribution, etc., but it's pretty hard to imagine a scenario in which I would use them for design. I'd find a free or cheap 2D analysis program before I'd do that.

In addition to the obvious issues of multiple load combinations and atypical loads, changes are killer for manual calcs. When the architect changes something, the manual calculations are a lot harder to change and book-keep.