Previous in Forum: HVAC Question Next in Forum: Bubble & Pump Impeller
Active Contributor

Join Date: Jun 2012
Location: Buenos Aires - Argentina - Mechanical Engineer
Posts: 15

# High Stress Concentration in Fillet Radius

10/25/2012 9:05 PM

Hi guys! I have made a FEM analysis of a roller and the results show below. The forces acting are flexion and torsion. The diameter max. of the roller is 330 mm and 5000 mm of total length. I do not like high stress values in the fillet zone. I tried refine mesh but the values increase even more. Also I tried increase fillet radius but there is not change. Is something wrong? Or only generates a hardening for localized plastic deformation and i can ignored this? I would like to know what you think Thanks

Sebastian.

http://imageshack.us/photo/my-images/11/82404028.jpg/

http://imageshack.us/photo/my-images/692/47940956.jpg/

http://imageshack.us/photo/my-images/547/94588059.jpg/

__________________
"Persistence prevails, like a stream that is temporarily blocked by boulders and then collects force enough to overflow onward."
Interested in this topic? By joining CR4 you can "subscribe" to

### Comments rated to be Good Answers:

These comments received enough positive ratings to make them "good answers".

### Comments rated to be "almost" Good Answers:

Check out these comments that don't yet have enough votes to be "official" good answers and, if you agree with them, rate them!
Guru

Join Date: Jun 2010
Posts: 1296
#1

### Re: High stress concentration in fillet radius

10/25/2012 9:39 PM

The axis of the roller is the x-axis and the end of the roller is in the y-z plane, correct? What is it rolling on - a flat surface, another roller, full width, partial width? What causes the flexural and torsional forces? I'm no FEM expert, but it seems to me that if you are trying to find the max. stress in what I think is the fillet, you need a lot more elements in the fillet itself.

Guru

Join Date: Aug 2009
Location: Glen Mills, PA.
Posts: 2385
#2

### Re: High stress concentration in fillet radius

10/25/2012 9:53 PM

I can't tell if the model includes a shoulder. Do you have a drawing of the roller and the loading? A rounded shoulder would definitely need a finer mesh there although I cannot see why your model shows the high stress' there.

__________________
In a time of universal deceit, telling the truth is a revolutionary act. George Orwell
Active Contributor

Join Date: Jun 2012
Location: Buenos Aires - Argentina - Mechanical Engineer
Posts: 15
#3

### Re: High stress concentration in fillet radius

10/25/2012 10:35 PM

Sorry but i don´t not understand way shows the same image three times. Again I put the links.

I think that the mesh is suficiently good to obtain a reasonable result....I will proof refining more the mesh.

__________________
"Persistence prevails, like a stream that is temporarily blocked by boulders and then collects force enough to overflow onward."
Guru

Join Date: Aug 2009
Location: Glen Mills, PA.
Posts: 2385
#4
In reply to #3

### Re: High stress concentration in fillet radius

10/26/2012 7:39 AM

The mesh looks okay there. What is the loading that gives this result?

__________________
In a time of universal deceit, telling the truth is a revolutionary act. George Orwell
Guru

Join Date: Oct 2008
Location: City of Roses.
Posts: 2056
#5

### Re: High Stress Concentration in Fillet Radius

10/26/2012 2:08 PM

I have found that when doing FEA analysis's, it's extremely important to fully understand your constraints and loads prior to the analysis. One poorly placed load/constraint can give you an answer completely different than the real world stress/strain.

Double check all your loads and constraints, have another engineer take a look at how you have set up your model for analysis. A second set of eyes are absolutely invaluable.

If you are absolutely 100% sure that all your loads and constraints are exactly as they exist in the real world, and your mesh is even in the ballpark, your results should be within reason.

Any time you get results far from what you would logically expect, back way up and look at it with a fresh set of eyes, and examine loads/constraints/geometry carefully.

The software can only provide you with values based on the input values...

Garbage in = Garbage out.

Best Regards,

-R

__________________
Don't believe everything you read on the Internet!
Score 1 for Good Answer
3
Power-User

Join Date: Nov 2011
Posts: 473
#6

### Re: High Stress Concentration in Fillet Radius

10/26/2012 11:24 PM

Sebastien, I think the problem is your boundary conditions as RVZ pointed out. Can you magnify the visual by 1000X and post a picture of the deflection?. I'd like to see how the material is behaving in your model. I've seen this before and it's always a boundary condition issue.

I don't know enough about your roller application but... if it's held in bearings and torque is applied in addition to radial loads, your FEM should look like this,

The small diameter has evenly distributed stress because its held in a bearing and allowed to twist to distribute those torsional stresses. It's the face of that diameter that is restrained in rotation. Secondly, the cylindrical support on that diameter is allowed to rotate parallel to the roller axis as the roller bends due to the radial load (ie it's restrained in translation along x,y,z, but allowed to pivot around x,y,z). This is what would naturally occur if the roller was supported by spherical type self aligning bearings. This means the bending is distributed along the entire length of the roller including the small diameters. The bending does not start abruptly at a sharp edge where your radius begins. Like this.

NOW...your model looks like this,

The small diameter is completely restrained as if it was inside an infinitely rigid hole and welded in place so it cannot move or twist in any direction. That's why there's no stress on it (nice blue color!). This means the torsional stress builds up at the junction between the radius and where that theoretical hole stops (sharp edge!!). Secondly, the small diameter can no longer follow the bending of the roller. Naturally, the stresses will peak at the junction between the radius and the theoretical infinitely rigid hole as shown below.

You also end up with a "crease effect" which does not represent reality. Below is image with deflection representation magnified 2000X to show you how the material is trying to fold in on itself and form a crease because of the way it is restrained on the small diameter inside the theoretical rigid hole.

I think this is what's causing your stress peaks. Is it something to worry about? Probably not. Take a section through the roller axis and see how far those high stresses penetrate from the surface towards the center. I'm betting they dissipate rapidly after a few millimetres...like this,

Just an opinion.

Good Answer (Score 3)
Active Contributor

Join Date: Jun 2012
Location: Buenos Aires - Argentina - Mechanical Engineer
Posts: 15
#8
In reply to #6

### Re: High Stress Concentration in Fillet Radius

10/26/2012 11:40 PM

Sorry Terraman!! We have crossed!Let me check your answer. It seems very interesting.

I will be back for you.

Thanks

__________________
"Persistence prevails, like a stream that is temporarily blocked by boulders and then collects force enough to overflow onward."
Active Contributor

Join Date: Jun 2012
Location: Buenos Aires - Argentina - Mechanical Engineer
Posts: 15
#9
In reply to #6

### Re: High Stress Concentration in Fillet Radius

10/27/2012 11:23 AM

Hi Terraman. Thanks you very much for so detailed response.You and RVZ are rigth. The problems was boundary condition. The new results show in below (in scale for major appreciaton). Note that maximum of V Mises stress is approx 190 MPa!!

Thanks and points for your good answers!!!

True scale:

__________________
"Persistence prevails, like a stream that is temporarily blocked by boulders and then collects force enough to overflow onward."
Power-User

Join Date: Nov 2011
Posts: 473
#10
In reply to #9

### Re: High Stress Concentration in Fillet Radius

10/27/2012 2:49 PM

I find that it always helps to magnify the deformation representations so that you can clearly visualize the mechanical behaviour of the material. This will help you determine if you've setup up your boundary conditions correctly.

Cheers.

Active Contributor

Join Date: Jun 2012
Location: Buenos Aires - Argentina - Mechanical Engineer
Posts: 15
#11
In reply to #10

### Re: High Stress Concentration in Fillet Radius

10/27/2012 3:13 PM

Thanks Terraman!!. Look at this:

Before

After

__________________
"Persistence prevails, like a stream that is temporarily blocked by boulders and then collects force enough to overflow onward."
Guru

Join Date: Feb 2011
Location: Atlanta, GA
Posts: 1053
#12
In reply to #6

### Re: High Stress Concentration in Fillet Radius

10/28/2012 8:20 PM

This is well beyond a "good" answer!

__________________
Think big. Drive small.
Power-User

Join Date: Nov 2011
Posts: 473
#13
In reply to #12

### Re: High Stress Concentration in Fillet Radius

10/29/2012 12:00 AM

Still only worth one GA!

There should be another option to rate as "really, really, really good answer". Maybe worth a few extra GAs?

Active Contributor

Join Date: Jun 2012
Location: Buenos Aires - Argentina - Mechanical Engineer
Posts: 15
#14
In reply to #12

### Re: High Stress Concentration in Fillet Radius

10/29/2012 8:58 PM

You are right K Fry, but my english language is limited....

My first attempt: TerraMan response was outstanding!!!!!

Thank you very much!!

Muchas gracias por sus respuestas!

Sebastián

__________________
"Persistence prevails, like a stream that is temporarily blocked by boulders and then collects force enough to overflow onward."
Active Contributor

Join Date: Jun 2012
Location: Buenos Aires - Argentina - Mechanical Engineer
Posts: 15
#7

### Re: High Stress Concentration in Fillet Radius

10/26/2012 11:35 PM

Dears

Thanks very much to all for responses. Really, I think that should not have problem with this. This is the upper roll that will be use in the pyramidal roll machine. The forces had been determinated from a analysis the a model relatively simple, but acceptable. The force value is 90 ton and is sufficient to cause plastic deformation in the sheet plate work. The torque necessary has been determinate as follows (please, correct me if this wrong):

conditions for not sliding

Fx => Fr

and Fr :U*N so

T=U*N**Radious upper roll=107,6 kN*m

U (contact Steel /Steel is 074) so

Radious upper roll is 165 mm and force is 90 ton.

Boundary conditions;

motor side: cilindrical support, only free tangential movement. The Torque is applied in this side.

no motor side: cylindrical support, all degree of freedom restricted.

The force is applied in a sector about all length

Gravity activate.

These high values of stress are given in a very Smallest portion.In the remainder stresses are low. I still Think that is not real.

__________________
"Persistence prevails, like a stream that is temporarily blocked by boulders and then collects force enough to overflow onward."
Interested in this topic? By joining CR4 you can "subscribe" to

### Comments rated to be Good Answers:

These comments received enough positive ratings to make them "good answers".

### Comments rated to be "almost" Good Answers:

Check out these comments that don't yet have enough votes to be "official" good answers and, if you agree with them, rate them!

### Users who posted comments:

bigg (1); K_Fry (1); passingtongreen (2); RVZ717 (1); Sebastian1981 (6); TerraMan (3)

 Previous in Forum: HVAC Question Next in Forum: Bubble & Pump Impeller