HSLA madness & stiffness controversey

[TheTick]

We are building a support frame from tubular steel. Boss is not satisfied with the amount of deflection (no plastic deformation) and wants to address deflection by switching to HSLA.

Everything in my experience tells me there is no difference in stiffness between HSLA and mild steel when operating below yield stress. Have I been deluding myself? I've checked MatWeb, and Young's modulus varies less than ±2% for all the grades I checked.

 

[GregLocock]

You are correct. There are very fundamental reasons which I have forgotten as to why Young's Modulus is what it is in steels, and so long as the main constituent is iron, it doesn't change much.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Tmoose]

http://www.advancepipeliner.com/Resources/Others/Beams/Beam_Deflection_Formulae.pdf

It is not an intuitive concept, but Deflection is all about geometry and E.

2X Thicker walls might produce a roughly similar reduction in deflection.

It is shaky out on this limb without any pictures, but it Sounds like there is a decent chance some individual members may be asked to withstand bending, so there is nearly as good a chance that at least 10X stiffness boost might result from a little carefully placed triangulation.

 

[dhengr]

Tick:
Look at the stress-strain curves for various construction steels in your Strength of Materials or Theory of Elasticity text books. The first portion, straight sloped portion, of the curves represent E, the slope of this line is E the modulus of elasticity which is essentially the same for all construction steels, 29x10^6 psi. The lower grade structural steels show a distinct ‘yield point;’ the HSLA steels and quenched and tempered steels don’t have a well defined yield point, so we set their ‘yield strength’ at a .2% strain offset, sloped parallel to the slope E. These steels perform essentially the same up to these yield points or strengths and are elastic when unloaded. Beyond these yield strengths, various steels exhibit slightly different characteristics in the plastic range and on up to ultimate strength. You said you are working in the elastic range, thus the same structural configuration will deflect the same whatever strength steel you use. The way to change your deflections is to change the various member moments of inertial, or shorten beam span lengths, or column lengths so as to reduce side sway. Maybe change connection details to stiffen rotations in the joints, fixed ends vs. pinned ends, etc. Maybe brace the structure in some way.

Your use of the word “stiffness” relating to Young’s Modulus or the modulus of elasticity is probably not the best choice of words. But, you do want to stiffen your structure and that’s done by changing the moment of inertia of your member cross sections, I (inches^4), and lengths (ft. or inches) of members which show up in bending moment, stress and deflection equations often ^2 (squared), ^3 (cubed) or ^4 powers, to great effect. Using the same sized members, but different strength steel, gives you the same stresses and deflections, but the higher strength steels are just being used at much less than full strength. Put another way HSLA steel is great if you can tolerate the higher stresses, deflections and vibrations, because they lead to lighter, less stiff members.

 

[TheTick]

I've fought this battle so many times. It's to flexy. What about HSLA? Boss thinks it's stiffer because it's stronger. Ugh!

The other perennial boondoggle is to just fatten up the gage thickness when something is too weak. That almost always puts material in all the wrong places. A little more web goes a long way.

These are all things I learned before my junior year. I've been re-teaching them ever since.

 

[drawoh]

TheTick,

I have been in variations of this discussion too. Common sense dictates that high strength steel should have a higher elastic modulus than low strength steel. When you design a structure for stiffness, you wind up way, way over the limits at which something can break. You do not need high strength materials.

I take it your boss is not an engineer?

JHG

 

[TheTick]

Worse, he's half an engineer. It might be easier if he was less than that.

 

[ivymike]

one of the interesting little quirks of stiffness calculations that I came upon years ago while working on a valvetrain vibration problem is that you can't change the ratio of stiffness/mass for a steel compression member (pushrod, in this case) by changing the cross-section, if you're looking at a relatively constant cross-section member [(FL/EA)/M = const]. Anything you do to bump up A has the same effect on M. Not necessarily applicable to your situation, but gave me a chuckle when i first wrote it down so long ago.

 

[Wrenchbender]

I found this to be a common mistake, especially with engineers who have been out of school for a long time. You have to refresh their memory while not being condescending and making them feel like a dope, which they will once they recall the phenomenology. BTW, the fundamental reason for E modulus (slope on the elastic curve) has to do with equilibrium spacing between the atoms.

A good reference for optimizing parameters like strength, stiffness, weight, cost, etc is Ashby - 'Materials Selection in mechanical Design'. I would also be cautious about buckling and unintentional eccentric loads.

 

[hydtools]

Look at deflection equations. The only material property involved is E. If you do not significantly change E you do not significantly change deflection.

http://www.engineersedge.com/beam_bending/beam_bending9.htm

Ted

 

[dhengr]

Tick:
As I said above, because HSLA steels are stronger, a material property, we are inclined to design lighter structures, IF we can tolerate the deflections and vibrations. Or, we over design to fix the deflection problem, thus under utilizing the premium material, stresswise. So, as you know, your boss has it bass acwards. As for increasing the thickness, that should increase the member stiffness and lower the stresses, these are section property related. But when you use stock or std. shapes, you do tend to get a bunch of material in places you don’t really need it. As is possible, I try to use stock shapes and sections, but am never afraid to fab. a section to put the material where I need it to achieve what I really want. Just pay attention to the fab. cost vs. saving weight and achieving what you want. Many times using std. shaped leads to such awful details elsewhere that they screw up the whole design and aren’t worth it. Show your boss the bending stress equations and the deflection equations for a simple beam, for example, so he can see what, and how drastically, some of these different properties change the results. The same principles generally apply to more complex structures, although the equations get more complex, or finally the analysis must be done by computer.

Drawoh:
If common sense were really so common, why is its application so uncommon? But, in this case, that particular common sense is wrong, even though it might be an uneducated extension of what one might first assume or guess, bosses included. But, then we all know where bad assumptions lead us...

We are really dealing with two different things here: material properties such as E (modulus of elasticity), Fy & Fu yield strength and ultimate strength, G (shear modulus), etc.; and section properties such as I (moment of inertia), S (section modulus), r (radius of gyration), etc., which when combined with member length and structural configuration, etc. lead to structural stiffness.
As Ivymike suggests you really have to study the formulas which govern what we are doing here to see exactly how each of the properties effects the results, or what we are trying to achieve. And, in his example maybe a hollow rod would offer the same stiffness and column capacity, while reducing the weight and mass. Maybe a higher strength material with the same compressive capacity would lead to a lower mass. Maybe a lighter weight material with slightly lower strength would do the trick. You have to manipulate those various properties and quantities, through those equations, to your advantage.

 

[ivymike]

@dhengr: nope, hollow does not improve compression stiffness. it does help euler buckling cover factor, but you only need so much of that.

 

[ivymike]

(should maybe mention that it's obvious you can make a rod stiffer in compression by increasing the cross sectional area - but the ratio of stiffness to mass doesn't change, because both mass and stiffness are linearly proportional to area)

 

[Ron]

steel is steel is steel...not much magic about it. As Greg noted, if you don't change "E"...you don't change much.

Lots of other properties can vary, but the mechanical properties are going to fall within a reasonably close envelope that doesn't change enough to make it significant.

 

[TheTick]

The HSLA material has arrived to build another proto. Pointless waste of time and money.

 

[vpl]

Tick

Maybe after you build it and there isn't any "improvement" due to the HSLA, it would be time for a discussion with your boss.

Something that would allow him to save face while pointing out that there could be time and money savings in listening to your technical input ...

Patricia Lougheed

******

Please see FAQ731-376: Eng-Tips.com Forum Policies for tips on how to make the best use of the Eng-Tips Forums.

 

[tr1ntx]

'Maybe after you build it and there isn't any "improvement" due to the HSLA, it would be time for a discussion with your boss.'

It's been my experience with such people that it will be declared to be "better" now and even if the measured data are the same, the decision will be made "we'll go with this one anyway".

 

[dhengr]

Tick:
Don’t change anything but the HSLA steel spec., and build the new prototype exactly as you built the old one, and test the two of them side by side as Patricia suggests. You will have a very expensive experiment program to prove your point to the boss. Be careful of changing section sizes, span lengths, connection details, welding details, loading conditions, etc. because they can change the outcome, but that can usually be explained too. As I said earlier, the easiest example would be a simple beam or a simple tension member, because there is less chance of detail variability affecting the results. After the testing, show your boss this thread, along with your own example work-up of a simple beam or tension member with formulas and all. I’d be glad to clean up my previous posts so as not to hurt any feelings too much, if I said anything derogatory about bosses. After all, I are one, so I know how that feels.

 

[LSPSCAT]

Appreciate the discussion and good intentions here, however, this is Engineering and Materials Science 101.

Discussion should go no farther than showing the boss the table of Young's moduli and an excerpt from a chapter on Strength of Materials. Typically there is a sentence that states this exact situation.

There is no room for discussion or argument. I realize you need to be more politically savy in broaching the subject, but this should be a simple and definitive argument to make.

 

[TheTick]

Part of it was His Bossness was so insistent and certain that he has seen this before and it worked (his way). He had me off balance and doubting what I knew. I'll admit I needed a little reassurance after that.

He'll see for himself, one way or the other, in a week or two.