standard way to apply the force of a human being standing on a plate?

[Walterke]

As the title says:
Is there any standard way to apply the force of a human being standing on a plate?

European standard EN280 says I should apply a point load, but a point load in FEA results in much too high stresses where the force is applied. How is this normally taken care of?

NX 7.5
Teamcenter 8

 

[rb1957]

i'd use a point load and look at element centroid stresses (that should be far enough away from the node.

else, you could apply the load over an area, say 2*50% footprint ?

 

[IRstuff]

A point loading is presumably the worst case. The smallest plausible area might be something like 2*20 sq. in. to allow for someone in heels.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[racookpe1978]

Couldn't be 2x20 inches square. (Not in heels anyway.)

Imagine a 200 lbs "lady" hitting the floor in heels. Wouldn't each shoe would have 3x4 sq surface at the toe, plus maybe a 3/8 x 3/8 area on the heel? Worse case, the heel comes down with an impact load of perhaps 60% body mass on the hell (er, heel) as a single point load over that 3/8 x 3/8 surface.

The rest would be on the tow of the other foot.

Then again, figure my weight of 155 lbs over a 12 inch long safety-toes work shoe. 155lbs/2 shoes/(12x3 per shoe) ... But that isn't a point load either.

 

[Walterke]

I don't think women (or man, for that matter) will be operating the work platform my floor comes in, in heels. The way I see you people discuss it, there seems to be no standard way to do this.

I'll just work something out then I guess. Thanks anyway.

NX 7.5
Teamcenter 8

 

[jhardy1]

The other thing to remember when you are working with plate / shell elements is that the element formulation is typically based on the mid-plane of the plate.

Ignoring for a moment the issue of what is the smallest physical "point load" in the real world, the concentrated load distributes as it propagates down from surface of the slab to the mid-plane. Using the old rule of thumb of a 45 degree angle of spread, by the time the "point" load reaches the mid-plane of a slab of thickness D, it will have an effective "footprint" approximately diameter D (or a square D x D).

In general, I spread "point loads" over an area comparable to thickness of the plate / slab to "dissipate" concentrated point load effects. If you need to understand stresses at a scale comparable to or smaller than the plate / slab thickness, you need to be considering other modelling techniques, such as solid elements, and a "real" footprint for the applied load.

http://julianh72.blogspot.com

 

[Walterke]

On a 2mm thick plate that wouldn't really help since my mesh is bigger then 2mm.


Basically, all I want to know is: will my sheet metal plate (supported by square tubes) be able to support 3 people standing on it. Or, how far do I have to put the tubes apart so that the structure is as light as possible, yet still strong enough.

When I apply my 3 people as point loads, the stresses are noticeably bigger then when I, for example, make 3 circles with diameter 50mm (~2inch) and apply my forces there. The size of the circles has a direct impact of the maximum stress, which is logical.
Only problem is that I can't find any reliable source regarding the size of said imprint.

I feel like I'm terribly over-thinking this, and there should be an easy way to fix my problem.

Hoping you guys can help.

NX 7.5
Teamcenter 8

 

[corus]

The standard clearly indicates a conservative method for deriving stresses by applying a point load. That should be the way you do it. If you use FEA then the stresses will be infinite as force over a zero area is. The standard will look at nominal stresses in the plate, and not those from disontinuities (such as restraint/change in shape/or point loads). Try plotting your stresses up to the point load. You should see a linear trend before they rapidly rise up to the load point. Use interpolation on the linear values to estimate the stresses at the load point, and so igrnoing the discontinuity stresse you get from someone wearing high heels.

 

[racookpe1978]

3x people can't stand right next to each other.

So your per-person "point" loads will be spread out over the area of each foot, then each of those will be about 1 foot apart.

Hint: 2 mm isn't thick enough.

 

[Walterke]

I put 3 loads 0,5m apart, as described in the European standard.
I'm currently looking to use 3,5mm alu treadplates, with a maximum span of about 40cm. with a distributed "foot-sized" load, this seems to be working fine. Point loads still give problems though.

NX 7.5
Teamcenter 8

 

[Twoballcane]

I’m not familiar with European standard EN280, however, if a point load is called out, this may mean that it is a force load and not a mass load being used. Not knowing how old this standard is, maybe this came out during the time when all Engineers did hand calculations. In hand calcs, you would take the force load to calculate stress in a beam for example or to find deflections and moments. So I would recommend doing the hand calcs first and then seeing what you get in FEA. How do you know the stresses are too high?

Tobalcane
"If you avoid failure, you also avoid success."
“Luck is where preparation meets opportunity”

 

[Walterke]

All I know is that there is a huge difference in result when I take a point load vs a load distributed over a small area. It could very well be that, when you do apply a point load (someone standing on a pin) it would cause the plate to deform. I'm just not sure if this is a good representation.

The standard is from 2009 but doesn't mention FEA. You're probably right about the force load. I was kind of thinking the same.

NX 7.5
Teamcenter 8

 

[bkal]

I might be missing the point here, but I would imagine that doing the FE analysis might be over the top for what you need. In general structural design, you are interested in moments and shear forces, and for plates or beams you could get the values from tables or first principles, and design the member according to the relevant design code. For these types of analyses, the contact surface area is not that important, e.g. moment in a simply supported, centraly point loaded beam is Force*span/4, and you are generally not concerned about local stresses.

Again, your particulr problem might be different, so I apologise if i missed the point.

 

[Walterke]

See added sketch (maybe I should've started with that)
Dimensions in mm
The plate you see is supported on the bent flanged (the 25mm ones).

Assuming the 740mm length can't be changed, how wide(current length 511mm) can I make the sheet so that it allows a person to stand in the middle of the sheet (and/or 2 people standing 500mm apart) without deforming the plate.

As you see this isn't exactly a beam problem and I can't seem to find a simple way to solve this manually. Hence the FEA.

Now when I apply a point load in the center of the plate of 1kN (80kg+25%), the local stresses go up to over 200N/mm², where the treadplate only allows 50N/mm² (yield strength=80N/mm²+1,65 safety factor)

Even if I reduce the span of 511mm to 300mm it still gives too big numbers, and, while I may be wrong here, I find it hard to believe that a 3,5mm alu plate of those dimensions can't support this kind of force.

So here goes my question: How am I supposed to apply my load of "1 person" so that it would give me a result I can work with?

NX 7.5
Teamcenter 8

 

[Twoballcane]

I am not allowed to open files, but if this is a plate problem and you can assume fixed or simply supported sides or attachments, you can use Roark’s Formulas to solve. For FEA, can you create a region that has the outline (area) of two shoes and then apply the force to the region. In essence, your just creating pressure (P=F/A) onto the plate. Also, (and this is why you should have a structural do the job or somebody who knows how FEA works) the high local stresses you’re seeing may be singularity functions which are inconclusive due to the way the model is meshed in that location. Your plate may be fine and the high stresses you’re seeing are just a few bad nodes. What you should be looking at is how over all the plate is responding to the load not just at a few places where stress measurements can be wrong. Another piece of advice is that I think your application may be too macro for the software you’re using. I suggest breaking up the plate to easier beam analysis that you can solve by hand and the FEA can have better results.

Tobalcane
"If you avoid failure, you also avoid success."
“Luck is where preparation meets opportunity”

 

[bkal]

I had a quick look in Pilkey, Formulas for Stress, Strain and Structural Matrices. If you distribute unfacatored load from one person (800N) over entire area of the plate (0.511m x 0.7m) you would get a peak bending moment of the order of 80Nm/m and bending stress of the order of 40MPa. As you know, having the load applied over a reduced area would increase this. However, deformation might be a problem; if it was steel, the small deformation theory displacement at the centre would be of the order of >2.5mm, i.e. you need a large deformation theory.

You can find more accurate formulas in Pilkey, Tables 18.

Please take above numerical results with a big pinch of salt.

 

[Walterke]

I used roark's to calculate an impression of 160x300mm² in the center of the plate. The results I got from NX were similar.

The only question that hasn't been answered yet is if the 160x300mm² is a good definition for a human imprint.

NX 7.5
Teamcenter 8

 

[corus]

From Roark's Forulae for Stress & Strain: Take a rectangular plate, simply supported on all sides, with a loaded rectangular section in the centre. and then let the loaded area tend to zero for the worst possible case. In my copy of Roark the values for a1/b & b1/b are left blank for zero values, but these values can be obtained by extrapolation of other values or from William Griffel's book, and are 2.6, 2.207, and 2.414 for a/b=1,1.4 and 2 respectively.

 

[Walterke]

Thx corus, the problem is that, when I lower the impression surface any further then the above mentioned numbers, the plate is no longer strong enough (after safety factors).

Extrapolating to a 0 value for a1/b and b1/b would equal applying a point pressure, but I don't know if that would be necessary, as it severely increases the required strength of the plate. (2,75 times higher). That would mean I'd need a 7mm alu plate, which I find very hard to believe.

NX 7.5
Teamcenter 8

 

[corus]

From your original question 'is there a standard way to apply the force...', the asnwer is yes, by a point load (see EN280). If you're unable to make a conservative approximation to your design calculation so that you can use simple plate or beam theory, or your hand calculated stresses exceed the design limits, then you're into the realms of 'Design By Analysis' using finite element methods.

In this case you can ignore the EN280 design standard methods of analysis and make what you consider are reasonable (but conservative) assumptions to reflect the real world situation in your model. Limit your FEA surface stresses to yield, and the mean stress through the thickness to 60% yield, and use some safety factor so that you are confident the design is safe even in the worst situation. In such a case then, I'd assume you have 3 people on tip-toe standing on one leg. The worst case could either be when they stand in the mioddle or at one edge (where buckling might occur).