Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations TugboatEng on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
Properties of Combination Sections 8
- Thread starter vintage70
- Start date
Trainguy,
You may be relieved to know that I am the one and only genuine (perhaps even dinkum) Austim.
I have happily embraced computer technology as it has been developed, and encouraged others to do so, but to use it intelligently and for appropriate purposes. No, my slide rule slides no longer, but my pocket (programmable) electronic calculator still works full time. [Perhaps you were never required to prepare alternative bridge girder designs for tender within an unreasonably brief tender period, before in-flight movies had been invented ?]
My previous lengthy post was directly aimed at Vintage70's initial question - "Is there a way to calculate S1 & S2 & I ?". Rightly or wrongly, I concluded that Vintage70 had either forgotten how to do it, or had never been taught the basic principles. That would not be a good start for setting up a spreadsheet.
1. Re my hand calcs for section properties:
I still stand by my view that rushing to a computer to do the calculation needed by vintage70 is the wrong way to go. In fact, for any beam calculation that I might do, the three properties sought by Vintage70 would only be part of the problem. What about Iyy, J, W (required to determine lateral/torsional buckling properties), and the plastic moduli about each principal axis? ACAD does not appear to offer any help with J and W or the plastic moduli, so I would have to pull out the pencil and paper anyway. And in one quick exercise I have my submission documentation finished.
I have never been one to avoid a challenge, but I cannot (quite) match your nominated 3 minutes. (But have you actually timed yourself)?
Under 'exam conditions' (starting with blank sheet of paper, section table booklet closed and calculator switched on), my best time so far for calculating the elastic moduli to the top and bottom of an Ibeam-plus-angles section is 4min, 34.9 seconds (including time to start and stop the digital stop watch on my laptop). Allowing for improved neatness of presentation for serious work, say about 5 minutes.
But that has already got me further than ACAD can do, since I have written down the two section moduli, which you cannot read from ACAD, and I have got a calculation sheet (including a sketch of the section) to incorporate into my record file.
I never use ACAD for drawing (only for reading drawings that others send to me), so any attempt of mine to confirm your 3 minutes estimate would be pointless. But I seriously question whether you could, within your suggested 3 minutes, find and load the relevant drawing, find a suitable cross-section, verify that it has been drawn accurately, select it, convert it to a region, find its CG, move it so that the relocated CG is at (0,0), record the CG location and M of I, and then calculate and record the two section moduli that Vintage70 wanted.
At the very best, the ACAD process will derive four elastic properties of the gross cross-section (A,ybar,Ix,Iy). Within current Australian practice, that is really not much use as a general procedure, and could mislead inexperienced engineers into some dangerous design decisions, particularly if they were to use it for large plate girders.
The Australian steel design code abandoned ASD methods for all structural work 12 years ago. As a result, the elastic moduli on their own are virtually useless to us. To determine the bending capacity of a section, we have to be able to calculate the plastic modulus and take account of possible reductions in effective section due to the slenderness ratios of the various plate elements involved in the section. Then, of course to calculate the bending capacity of a beam, we have to allow for possible lateral/torsional buckling, which requires us to know the two torsional parameters J and W.
[The best that I can get from ACAD is an upper bound to the design values for A, Ixx, Iyy, which can be reduced significantly by plate slenderness effects.]
Essentially, these factors force us to use standard sections for which all tabulated properties are available, or alternatively to know what we are doing, and calculate for ourselves the design properties for all special sections. In much of my own work (heavily loaded plate girders etc) the latter option is unavoidable.
2. re my non-linear suggestion (which was not entirely tongue-in-cheek. Some day I will try it for myself).
Again, current Australian practice comes into the picture. Since 1990, our LFRD design code has specified quite clearly that if we only use first-order (linear) analyses, then we have to multiply all of our output moments by moment amplification factors which vary according to the critical axial buckling load applied axial load in each member.
Alternatively we may use the output from a non-linear analysis without further modification. (Provided that we keep above a buckling factor of 5, the non-linearity does not have to include the effect of axial load on the member stiffness). The 2nd-order alternative is much 'neater' and simpler in practice. It rarely adds much to the analysis time, and avoids a very messy process once we have our analytical results. So, whenever I can, I adopt 'non-linear' as a standard process.
You may be relieved to know that I am the one and only genuine (perhaps even dinkum) Austim.
I have happily embraced computer technology as it has been developed, and encouraged others to do so, but to use it intelligently and for appropriate purposes. No, my slide rule slides no longer, but my pocket (programmable) electronic calculator still works full time. [Perhaps you were never required to prepare alternative bridge girder designs for tender within an unreasonably brief tender period, before in-flight movies had been invented ?]
My previous lengthy post was directly aimed at Vintage70's initial question - "Is there a way to calculate S1 & S2 & I ?". Rightly or wrongly, I concluded that Vintage70 had either forgotten how to do it, or had never been taught the basic principles. That would not be a good start for setting up a spreadsheet.
1. Re my hand calcs for section properties:
I still stand by my view that rushing to a computer to do the calculation needed by vintage70 is the wrong way to go. In fact, for any beam calculation that I might do, the three properties sought by Vintage70 would only be part of the problem. What about Iyy, J, W (required to determine lateral/torsional buckling properties), and the plastic moduli about each principal axis? ACAD does not appear to offer any help with J and W or the plastic moduli, so I would have to pull out the pencil and paper anyway. And in one quick exercise I have my submission documentation finished.
I have never been one to avoid a challenge, but I cannot (quite) match your nominated 3 minutes. (But have you actually timed yourself)?
Under 'exam conditions' (starting with blank sheet of paper, section table booklet closed and calculator switched on), my best time so far for calculating the elastic moduli to the top and bottom of an Ibeam-plus-angles section is 4min, 34.9 seconds (including time to start and stop the digital stop watch on my laptop). Allowing for improved neatness of presentation for serious work, say about 5 minutes.
But that has already got me further than ACAD can do, since I have written down the two section moduli, which you cannot read from ACAD, and I have got a calculation sheet (including a sketch of the section) to incorporate into my record file.
I never use ACAD for drawing (only for reading drawings that others send to me), so any attempt of mine to confirm your 3 minutes estimate would be pointless. But I seriously question whether you could, within your suggested 3 minutes, find and load the relevant drawing, find a suitable cross-section, verify that it has been drawn accurately, select it, convert it to a region, find its CG, move it so that the relocated CG is at (0,0), record the CG location and M of I, and then calculate and record the two section moduli that Vintage70 wanted.
At the very best, the ACAD process will derive four elastic properties of the gross cross-section (A,ybar,Ix,Iy). Within current Australian practice, that is really not much use as a general procedure, and could mislead inexperienced engineers into some dangerous design decisions, particularly if they were to use it for large plate girders.
The Australian steel design code abandoned ASD methods for all structural work 12 years ago. As a result, the elastic moduli on their own are virtually useless to us. To determine the bending capacity of a section, we have to be able to calculate the plastic modulus and take account of possible reductions in effective section due to the slenderness ratios of the various plate elements involved in the section. Then, of course to calculate the bending capacity of a beam, we have to allow for possible lateral/torsional buckling, which requires us to know the two torsional parameters J and W.
[The best that I can get from ACAD is an upper bound to the design values for A, Ixx, Iyy, which can be reduced significantly by plate slenderness effects.]
Essentially, these factors force us to use standard sections for which all tabulated properties are available, or alternatively to know what we are doing, and calculate for ourselves the design properties for all special sections. In much of my own work (heavily loaded plate girders etc) the latter option is unavoidable.
2. re my non-linear suggestion (which was not entirely tongue-in-cheek. Some day I will try it for myself).
Again, current Australian practice comes into the picture. Since 1990, our LFRD design code has specified quite clearly that if we only use first-order (linear) analyses, then we have to multiply all of our output moments by moment amplification factors which vary according to the critical axial buckling load applied axial load in each member.
Alternatively we may use the output from a non-linear analysis without further modification. (Provided that we keep above a buckling factor of 5, the non-linearity does not have to include the effect of axial load on the member stiffness). The 2nd-order alternative is much 'neater' and simpler in practice. It rarely adds much to the analysis time, and avoids a very messy process once we have our analytical results. So, whenever I can, I adopt 'non-linear' as a standard process.
- Thread starter
- #22
Can I stick in just this time? Since I am the cause of this great debate. My question was very brief. I thought I would ask the question and get a response quicker than digging out the books. I was right there. I did learn the basics a long time ago. I occasionally come across problems that I am more curious about than resposible for. Areas I don't really get involved in but hey I still like this stuff. I wanted the hand calculation. That's for me; the only way I will understand it. The software was brought up by someone else. But I will not get in the middle of this debate. That was not my question. Austim thanks very much.
Austim,
The quality of your posts is, as always, highly informative and entertaining. The quantity of some of the above is nothing short of breathtaking. That you have the time to do all those hand calcs, and write posts that long is quite impressive....![[bigsmile]](/data/assets/smilies/bigsmile.gif "[bigsmile] [bigsmile]")
OK, maybe 3 minutes was an exaggeration, but if it's 5 or 6, it is that value for any arrangement of angles, channels, symmetric, non-symmetric by my Acad method. You're right on about needing other properties for design (esp. in bridges and buildings fields). You should see the wierd section shapes I encounter in railcar design, some aluminum extrusions are really quite complex. Many of my design criteria refer to permanent deformation and yielding so an ASD method is somewhat common practice here.
Point well made about the ORIGINAL post, it IS an easy hand calc. I'd probably still use a spreadsheet approach (without ACAD) that I'd written myself. At least, the arithmetic is foolproof if the data is entered right. It takes just as much time to enter the data on your hand calculator than it does in Excel, but with the added benefit of a nice print-out.
I can't believe I just took 12 minutes to write this post! I could have done 4 section property calculations...![[bigcheeks]](/data/assets/smilies/bigcheeks.gif "[bigcheeks] [bigcheeks]")
The quality of your posts is, as always, highly informative and entertaining. The quantity of some of the above is nothing short of breathtaking. That you have the time to do all those hand calcs, and write posts that long is quite impressive....
OK, maybe 3 minutes was an exaggeration, but if it's 5 or 6, it is that value for any arrangement of angles, channels, symmetric, non-symmetric by my Acad method. You're right on about needing other properties for design (esp. in bridges and buildings fields). You should see the wierd section shapes I encounter in railcar design, some aluminum extrusions are really quite complex. Many of my design criteria refer to permanent deformation and yielding so an ASD method is somewhat common practice here.
Point well made about the ORIGINAL post, it IS an easy hand calc. I'd probably still use a spreadsheet approach (without ACAD) that I'd written myself. At least, the arithmetic is foolproof if the data is entered right. It takes just as much time to enter the data on your hand calculator than it does in Excel, but with the added benefit of a nice print-out.
I can't believe I just took 12 minutes to write this post! I could have done 4 section property calculations...
trainguy,
I will let you into my time secret (just between the two of us, mind you).
I am due to retire permanently before the end of this month, and I now have no outstanding work commitments![[sleeping]](/data/assets/smilies/sleeping.gif "[sleeping] [sleeping]")
.
But I am glad to see that between us all we seem to kept vintage70 happy.
I will let you into my time secret (just between the two of us, mind you).
I am due to retire permanently before the end of this month, and I now have no outstanding work commitments
.But I am glad to see that between us all we seem to kept vintage70 happy.
- Thread starter
- #26
BrianWaligura
Structural
There is an important aspect of this discussion that was left out. Assuming Vintage70's beam (with the combinaton section) is going to be subjected to strong axis bending that will produce a changing bending moment in the beam, the fastening (weld, bolts, etc.) that joins the components of his combinaton section must be adequate to resist the shear flow that accompanies that changing bending moment. If the fastening cannot resist the shear flow, the bending properties of the section about the strong axis (S=Section Modulus and I=Moment of Inertia) are not those that one would calculate from the traditional method as described by austim. My point is this: the fastening must be adequately strong in order for the section to act as a combination section.
Boy i sure like the last post after all the pontificating. Imagine that 'Built-up sections' calculating section properties as though all one, and the finally how to connect it all so that it is one. The problem is more than one of geometry and the connection issue is surely more challenging. GOOD STUFF BRIAN !, your in the game for sure !!!. Get out Mohrs Circle, Principal Stresses, contour lines, oh oh lets not talk about stress concentrations, oh geez welded connections. Maybe a program for that somewhere ? hahaha.
- Status
Similar threads
- Question
- Question
- Locked
- Question