Stress Analysis Results

[bhowe]

Hi All,

I have a stress analysis result that doesn't make sense to me. I was wondering if someone could take a look and let me know their opinion.

I have a flat plate, that is loaded with 100psi equally throughout the surface. I have an array of bolt holes that are 120mm x 109mm spacing. When I do the simulation and look at deflection, I see more movement on the ends. I expected the movement to be consistant eveywhere?

Is this an Inventor issue, or are my expectations in error?

On a side bar, is there a better way to constrain holes than on their edges? I would think that would give errors on the stress levels due to the point (line) contact.

[JDMather]

Attach the ipt file here.

How is it constrained in the real world?

Fastner with clamping face of some definable area?
Conact with another component?
More information is needed.

[bhowe]

I have attached the IPT file. I essentially have two identical plates bolted together. Yes, in real life there would be the are under the bolt heads as the constraining area ( not the line around the clearance hole. The mating half will have channels cut into it with 100 psi water. I tried doing a split face around the bolt to simulate the bolt head. I got mesh errors. I tried extruding a round area 0.001in high to represent the bolt heads, again mesh errors. I tried to resize the mesh, still got errors.

[cwhetten]

Concerning the greater displacement at the ends, this is reasonable.  I will try my best to explain it, but I may fail.  This picture may help:

If there was more material on the other side of the last set of bolts, it would act like a lever under the pressure, with the line of bolts as the fulcrum.  This would tend to cause the material to the left to not deflect as much.  And indeed, this is the situation at each of the intermediate areas.  But because the edge bolts don't have this "lever" material on the other side, the material to the left deflects more than it does at any of the intermediate points.

I hope I explained that in a clear way.

[bhowe]

Is there a more realistice way to place constraints at the holes? I currently have them at the hole edge. I believe this is giving me a much higher local stress than will really be the case. As I mentioned above, I tried SPLIT and a small Extrusion to try to spread the load over the size of a screw head. I had mesh errors when I tried them.

[rhinterhoeller]

Would the following more realistically represent the real world:

1/ For one bolt:

  a/ create a circular split on the opposite face of one part

  b/ apply a force equal to the pre-load from the bolt head & washer

  c/ create a similar split on the other part

  d/ apply the equal and opposite pre-load from the nut & washer

2/ Repeat for each bolt

You will now have created a clamping pressure on the mating surfaces.  If you've done the design right the pressure in the channels won't permit the parts to separate.

[bhowe]

I tried using a split to represent the bolt head. I ended up with mesh errors? I'll try it again.

[m.granata]

bhowe,

I have only used ANSYS and Algor (now Autodesk's) FEA packages and am not familiar with what you are using.  From the geometry your deformation results don't look out of line.  From an Engineering standpoint, and if this plate will be subject to 100 psi loads, I would try different geometries.  For example, first off increase the number of bolts, visually, your bolt spacing looks too large.  Try reducing the bolts spacing to almost half of what it currently is.  Second, increase the material thickness.  If that is not feasable add flat bar stifeners on edge or a 1.5" to 2" wide backing plate around the perimeter to thicken the material where the bolts pass through.

I see you are concerned about localized high deformations but don't forget this plate needs to seal with a 100 psi load on it.  With FEA, localized yielding at discontinuities become visible and are sometimes nothing to be concerned with.

[JDMather]

---

@bhowe wrote:

I tried using a split to represent the bolt head. I ended up with mesh errors? I'll try it again.

---

Not sure why you would experience meshing error.

Also - what is the Actual deflection (I get less than the thickness of a sheet of paper).

[bhowe]

Originally, I posted, because I didn't think the results were correct. I didn't expect more deflection at the ends than the middle, given the uniform constraints? The second issue was the high stress at the points of constraint. I constrained the plate at the edge of the bolt holes. I expected high stress by this method. I was looking for a more realistic way to constrain, but didn't have any luck using split or extrusion. Yes, the deflection is low, but with o-ring seals, movement of 0.005" can cause a leak.

[JDMather]

---

@bhowe wrote:

 but with o-ring seals, ...

---

What does the mating part look like?  Can you attach it here?
It seems to me like an entire area of your part should be split away because it is sealed from the pressure.

[bhowe]

Well, this time I was able to add the split to each hole, and constrain to an area (under the screw head) instead of a line. The stress went way down ( as expected). Takes time to do a split for each holes. Is this the metod of choice that everyone uses to simulate bolt heads?

[bhowe]

No, the entire surface could see pressure if there is movement. The other plate is the same except for water channels. Everything is looking good now.

Thanks

[rhinterhoeller]

I think it's obvious that it's my preference.

Do the results look reasonable to you?

I didn't look carefully at your initial image but would it have been possible to model half of the part and use symmetry to reduce your workload?

Richard

[bhowe]

---

@rhinterhoeller wrote:

I think it's obvious that it's my preference.

Do the results look reasonable to you?

I didn't look carefully at your initial image but would it have been possible to model half of the part and use symmetry to reduce your workload?

Richard

---

The results look resonable now. Happy with the results!

[JDMather]

---

@bhowe wrote:

constrain to an area (under the screw head).

---

The screw head will not contact that face. The screw head will hold a mating part to that part.  What does the actual contact area look like? 

Why would the mating part move? 

[bhowe]

---

@Anonymous wrote:

---

@bhowe wrote:

constrain to an area (under the screw head).

---

The screw head will not contact that face. The screw head will hold a mating part to that part.  What does the actual contact area look like? 

Why would the mating part move? 

---

JD

The actual condition is a tapped hole thru the part on the part I had trouble with. The mating part would have screw heads. The movement is when the parts seperate, water will fill the void, creating additional area for the pressure to react with.  

[m.granata]

bhowe,

Just joking around, no insults here.  How many hours do you have to date in the FEA analysis of this flat plate?  Or is this a graduate thesis?  Is your boss breathing down your neck yet?  Aside from FEA I still believe the design needs modifications.

[bhowe]

---

@m.granata wrote:

bhowe,

Just joking around, no insults here.  How many hours do you have to date in the FEA analysis of this flat plate?  Or is this a graduate thesis?  Is your boss breathing down your neck yet?  Aside from FEA I still believe the design needs modifications.

---

Not sure where you were going with this? I haven't spent much time on the FEA. The design is not done. Merely testing the bolt spacing on two 1/2inch plates with 100 psi coolant flow. I was looking to see if the deflection between holes would cause a leak at the oring seal. It looks like it should work, now I'll finish up the design. The rest of the thread is educational, I don't use the FEA module often. I was curious about how most are constraining the bolt interface to get a realistic localized stress. Was I asking too many questions?

[m.granata]

No you are not asking too many questions.  With FEA, sometimes you have to simplifiy geometries from the actual part.  As a Mechanical Engineer and just looking at the attachment image of your part your bolt spacings look too large and the bolt hole size seems too small.  If your FEA model load and boundry conditions are as close to actual conditions as you can get, I am wondering if your FEA Model is fine but the design needs some changes.  Often that is the case.  I analysed a pressure vessel to comply with ASME BPVC using my clients existing design.  After many, many, hours with FEA I finally had to inform them the vessel head type currently used will not comply with ASME BPVC.  They hum-hoe'd around befored deciding to change it.

After you finally get a design with reasonable FEA reults I suggest digging through Roark's Formula For Stress and Strain to find similar plate loading and boundry condition, do the hand calculations, then compare the results.  They won't be exact, but close enough to verify wether or not your FEA model is correct.  Both in undergraduate and graduate school it was drilled into me you need another method to verify FEA results otherwise it may just be garbage-in-garbage-out.

Good Luck