Pulley FEA
#26
Super Member
Thread Starter
Join Date: May 2002
Location: SF South Bay
Posts: 844
Likes: 0
Received 1 Like
on
1 Post
2002 C230 Coupe
Originally posted by Buellwinkle
In the all alloy pulley you left out one huge detail, the thickness of the hub is considerably smaller because of the steel sleeve which is not supporting any weight so you have to take that into consideration. For example if the hub walls are .20" thick on the factory pulley then the alloy pulley may only be .12" thick.
In the all alloy pulley you left out one huge detail, the thickness of the hub is considerably smaller because of the steel sleeve which is not supporting any weight so you have to take that into consideration. For example if the hub walls are .20" thick on the factory pulley then the alloy pulley may only be .12" thick.
I'm making something more fun now
#27
Super Member
Thread Starter
Join Date: May 2002
Location: SF South Bay
Posts: 844
Likes: 0
Received 1 Like
on
1 Post
2002 C230 Coupe
More realistic simulation
I really wanted to install the pulley (can't resist its attraction...) so I went ahead and did a more realistic analysis:
Material properties were taken at 200F (~90C operating temperature) for stainless steel hub and set screws and T6061-T6 aluminum alloy.
Refine the mesh of the model (see pics below) and move the location of the set screw to the center of the model (the previous model has 2 half screws on each end). This results in a full screw in the middle and minimize boundary effects. And, because of the refinement, I was able to make the screw more "cylindrical".
Do not include any belt loads, vibration loads, etc.. Only analyze the stresses due to bolt force. I looked up some reference, using an assembly lubricant Molykote of u = 0.08, the bolt will develope 43 lb/ft-lb of axial force. So, for torque = 220 ft-lb, the resulting force is 9460 lb.
The hub flange surface area is 2.4 in^2, which is about the area of the washers. So the resulting pressure on the flange is about 4000 psi (huge isn't it?!!).
Material properties were taken at 200F (~90C operating temperature) for stainless steel hub and set screws and T6061-T6 aluminum alloy.
Refine the mesh of the model (see pics below) and move the location of the set screw to the center of the model (the previous model has 2 half screws on each end). This results in a full screw in the middle and minimize boundary effects. And, because of the refinement, I was able to make the screw more "cylindrical".
Do not include any belt loads, vibration loads, etc.. Only analyze the stresses due to bolt force. I looked up some reference, using an assembly lubricant Molykote of u = 0.08, the bolt will develope 43 lb/ft-lb of axial force. So, for torque = 220 ft-lb, the resulting force is 9460 lb.
The hub flange surface area is 2.4 in^2, which is about the area of the washers. So the resulting pressure on the flange is about 4000 psi (huge isn't it?!!).
#30
Super Member
Thread Starter
Join Date: May 2002
Location: SF South Bay
Posts: 844
Likes: 0
Received 1 Like
on
1 Post
2002 C230 Coupe
Then I "moved" the set screw location 0.04" farther, which is 0.05" from the hub surface, and re-ran the analysis. Guess what, the von Mises stress dramatically reduced to 27 ksi!!! And the overall max is on the steel hub now. Stainless steel has yield strength of about 30 ksi. But the max stress on the aluminum is still higher than 20 ksi.
Keep in mind that the results were due to bolt force only, and there are other types of loads on the real pulley. I really wish the screws could have been several hundredth inches more from the hub...
There is always a fine line between analysis and reality. So maybe I'll just put it on for a couple of weeks, then take it off to see if there is any cracks. If one pulley could last several months before fracture, I think a couple of weeks should be OK
Keep in mind that the results were due to bolt force only, and there are other types of loads on the real pulley. I really wish the screws could have been several hundredth inches more from the hub...
There is always a fine line between analysis and reality. So maybe I'll just put it on for a couple of weeks, then take it off to see if there is any cracks. If one pulley could last several months before fracture, I think a couple of weeks should be OK
Last edited by 20FHK02; 02-22-2003 at 03:00 AM.
#31
Originally posted by speedfrk
Buell, how do they weld an aluminum ring on to a steel pulley? Do they weld a steel ring on first and then press the aluminum ring on after?
Buell, how do they weld an aluminum ring on to a steel pulley? Do they weld a steel ring on first and then press the aluminum ring on after?
#32
Super Member
Thread Starter
Join Date: May 2002
Location: SF South Bay
Posts: 844
Likes: 0
Received 1 Like
on
1 Post
2002 C230 Coupe
REVISED FINITE ELEMENT ANALYSIS OF ASP PULLEY
Okay guys: I know I'm going into too technical here... but since I've done the analysis so why not share with the few who are interested, right?! But if you don't know what I'm talking about below, just jump to the Conclusion section.
Background
After intensive (sort of...) measurements, I found that the gap between the set screw OD and the hub surface is more like 0.025", intead of 0.01" that I previously stated. And the torque-force conversion factor was too conservative because previous analysis showed that even the stress on the steel was relatively too high for safety concern (because MB will not design a component such close to the margin).
So I re-ran the analysis with the more realistic gap size, and a u = 0.09 (i.e., 38 lb per ft-lb of torque) and the new flange pressure is now 3500 psi. And the flange thickness is assumed to be 3/16".
Results and Discussion
New results show that, isolating the aluminum body from the hub, the von Mises stress on the aluminum is now 29 ksi (see Fig. 1 & 2).
Although this still exceeds the yield, when evaluating the component stresses - radial (Fig. 3), hoop (Fig. 4), and axial (Fig .5), a major portion was contributed by axial stress, which is not realistic because the hub is not welded onto the pulley body. Therefore, the actual axial stress should be much lower.
Radial and hoop stresses are the main contributors for cracking if they are tensile, but results show that they are all compressive.
In addition, I also did a linearized stress calculation through the width of the gap. Results show that the linearized max radial stress is 20 ksi, and 13 ksi for hoop (axial is invalid). These are more comfortable numbers!
Note that the torque-bolt force conversion used here is actually for nuclear reactor vessels using 2.5-inch diameter bolts, so I think, a BIG think, the actual force on the crank bolt should be lower. But due to uncertainties and the fact that other types of forces are not included, so this conservative force is okay to be used.
Conclusion
The results from this revised analysis make me more comfortable now, although I still wish that the gap could have been a bit larger.
I know a theory is junk if it does not reflect reality. So I've decided to install the pulley and see what will happen!!
Background
After intensive (sort of...) measurements, I found that the gap between the set screw OD and the hub surface is more like 0.025", intead of 0.01" that I previously stated. And the torque-force conversion factor was too conservative because previous analysis showed that even the stress on the steel was relatively too high for safety concern (because MB will not design a component such close to the margin).
So I re-ran the analysis with the more realistic gap size, and a u = 0.09 (i.e., 38 lb per ft-lb of torque) and the new flange pressure is now 3500 psi. And the flange thickness is assumed to be 3/16".
Results and Discussion
New results show that, isolating the aluminum body from the hub, the von Mises stress on the aluminum is now 29 ksi (see Fig. 1 & 2).
Although this still exceeds the yield, when evaluating the component stresses - radial (Fig. 3), hoop (Fig. 4), and axial (Fig .5), a major portion was contributed by axial stress, which is not realistic because the hub is not welded onto the pulley body. Therefore, the actual axial stress should be much lower.
Radial and hoop stresses are the main contributors for cracking if they are tensile, but results show that they are all compressive.
In addition, I also did a linearized stress calculation through the width of the gap. Results show that the linearized max radial stress is 20 ksi, and 13 ksi for hoop (axial is invalid). These are more comfortable numbers!
Note that the torque-bolt force conversion used here is actually for nuclear reactor vessels using 2.5-inch diameter bolts, so I think, a BIG think, the actual force on the crank bolt should be lower. But due to uncertainties and the fact that other types of forces are not included, so this conservative force is okay to be used.
Conclusion
The results from this revised analysis make me more comfortable now, although I still wish that the gap could have been a bit larger.
I know a theory is junk if it does not reflect reality. So I've decided to install the pulley and see what will happen!!
Last edited by 20FHK02; 02-22-2003 at 07:14 PM.
#38
Super Member
Thread Starter
Join Date: May 2002
Location: SF South Bay
Posts: 844
Likes: 0
Received 1 Like
on
1 Post
2002 C230 Coupe
That's all!
Thank you for your time and attention.
Notes:
1. A real stock pulley model will not be made because I'm lazy to take the measurement and create a whole new model. But since it is all steel and the hub geometry is identical to the ASP one, so Case 4 is applicable to the stock pulley for comparison.
2. For an alloy pulley with steel sleeve, the stress will be worse than Case 3 because of a thinner alloy hub thickness. And the thin edge of the steel sleeve will create high stresses on the contact surface of the alloy side.
Notes:
1. A real stock pulley model will not be made because I'm lazy to take the measurement and create a whole new model. But since it is all steel and the hub geometry is identical to the ASP one, so Case 4 is applicable to the stock pulley for comparison.
2. For an alloy pulley with steel sleeve, the stress will be worse than Case 3 because of a thinner alloy hub thickness. And the thin edge of the steel sleeve will create high stresses on the contact surface of the alloy side.
Last edited by 20FHK02; 02-22-2003 at 07:03 PM.
#39
Senior Member
Join Date: Mar 2002
Location: Atlanta, GA
Posts: 461
Likes: 0
Received 0 Likes
on
0 Posts
C230 coupe 6sp
Originally posted by DCX Engineer
The new outer ring is made of steel. It is pressed on, welded and rebalanced. 10 and 15% larger rings are available.
The new outer ring is made of steel. It is pressed on, welded and rebalanced. 10 and 15% larger rings are available.
#40
MBWorld Fanatic!
Way over my head. What's the outcome of the alloy pulley with steel sleeve and would the thinner sleeve on the Kleemann pulley make much of a difference or should those people worry?
#41
Senior Member
Join Date: Dec 2002
Location: New Castle, IN
Posts: 355
Likes: 0
Received 0 Likes
on
0 Posts
C230Kompressor
Re: That's it!
Originally posted by 20FHK02
Thank you for your time and attention.
Thank you for your time and attention.
No, thank you! I only understand about 1/2 of what you're talking about, but it's 100% more than I knew before. My ASP pulley is on it's way to me now, hopefully to be installed by next weekend.
#42
Super Member
Thread Starter
Join Date: May 2002
Location: SF South Bay
Posts: 844
Likes: 0
Received 1 Like
on
1 Post
2002 C230 Coupe
Originally posted by Buellwinkle
Way over my head. What's the outcome of the alloy pulley with steel sleeve and would the thinner sleeve on the Kleemann pulley make much of a difference or should those people worry?
Way over my head. What's the outcome of the alloy pulley with steel sleeve and would the thinner sleeve on the Kleemann pulley make much of a difference or should those people worry?
#44
MBWorld Fanatic!
Originally posted by 20FHK02
If you really want to know, I can do a detailed analysis on this using contact elements so that the aluminum and steel are no longer treated as welded together. But I need the dimensions of the thing and the outcome of the analysis might get into some people's proprietary territory...
If you really want to know, I can do a detailed analysis on this using contact elements so that the aluminum and steel are no longer treated as welded together. But I need the dimensions of the thing and the outcome of the analysis might get into some people's proprietary territory...
#45
Super Member
Thread Starter
Join Date: May 2002
Location: SF South Bay
Posts: 844
Likes: 0
Received 1 Like
on
1 Post
2002 C230 Coupe
Originally posted by Buellwinkle
Use Kleemann's dimension's of .07" for the steel sleeve, the rest is T6061 alluminum.
Use Kleemann's dimension's of .07" for the steel sleeve, the rest is T6061 alluminum.
#46
MBWorld Fanatic!
Originally posted by 20FHK02
Not wanting to offend anyone nor be involved in any legal issues, I will email the results to you.
Not wanting to offend anyone nor be involved in any legal issues, I will email the results to you.
#48
Super Member
Thread Starter
Join Date: May 2002
Location: SF South Bay
Posts: 844
Likes: 0
Received 1 Like
on
1 Post
2002 C230 Coupe
I found interesting results after separating the 2 materials and take the pulley to 3000 rpm
Does anybody happens to know the guesstimate force from the belt tensioners?? (Like typically they are around xx lb, etc.)
Does anybody happens to know the guesstimate force from the belt tensioners?? (Like typically they are around xx lb, etc.)
#50
MBWorld Fanatic!
As far as legalities 20FHK02 , why not just state foir the record this is purely for fun, and the numbers don't neccesarily represent accurate values, since you're basing your analysis on theoretical values....
I kinda gathered that anyway, we're just dealing with theory here.
I kinda gathered that anyway, we're just dealing with theory here.