FEA in design

bill_daniels

I have a very specific FEA problem that I need an app to analyze.  It involves a drum on a rope winch withstanding the constrictive force of a long,slippery rope wound on very tightly.  I know how to calculate the constrictive force which can reach over a million pounds total acting inward on the drum core which is the inverse to what a pressure vessel sees.  Most of this force is supported by the hoop strength of the drum core barrel.  I need to make sure my design is strong enough.

My question:  Will any of the FEA simulation apps handle this scenario?

fastwayjim

I would recommend Simscale, which has a direct link to Onshape. You can do linear, non-linear, steady state, transient, kinematics, assembly contact, etc.

From what I understand, you want to apply a pressure to the outer surface of a cylindrical drum? Can you provide a picture or share the model?

bill_daniels

I'm happy to share the link to the Part Studio.
https://cad.onshape.com/documents/dbc50e4c06a0011863e3b8ea/w/46eeef81efb2a937f3aeaadf/e/742a9e8c69b2b8db96fe8102
The construction will be TIG/MIG welded HSLA steel.
The winch drum is intended to hold as much as 10,000 feet of 3/16" diameter Amsteel Blue (UHMW-PE fiber rope) with as much as half that wound on at 2200 Lbs-F tension.  This exerts an enormous constrictive force on the drum.

bill_daniels

Tried SimForDesign.  It's probable I made many mistakes as this was my first try with the App but I couldn't get a result.  The mesh generation failed last evening then restarted this morning.  When the mesh appeared complete, the App said I didn't have enough time left on my free session to complete the calculation.  Moving on trying SimScale next.

bill_daniels

Watched all the videos for SimScale today then ran a successful FEA simulation.  I was more than a little surprised it worked as well as it did.  While my drum design was workable, the SimScale results clearly marked areas I need to change in the model to make it more structurally efficient.

fastwayjim

Great news, @bill_daniels! Glad to hear. I have the same handle on that platform (fastwayjim), so feel free to reach out to me if you have any questions! Onshape/Simscale is the perfect 1-2 punch for web-based engineering work.

bill_daniels

Actually, fastwayjim, I do have a question.  While the drum was pretty easy - at least for the constriction forces - the side flanges are harder in that the actual lateral force generated by the rope mass squeezing out sideways is a gradient starting high near the drum core (barrel) and diminishing toward the rim of the flanges.  The force distribution is typically a quarter ellipse - I can't yet see how to apply stress like that.

fastwayjim

Simscale doesn't have an (r, theta) coordinate system right now, so it won't be straight forward, but underneath the entry for a force are two options for having a variable force: table, and function.



Here is the scope of functions:



Perhaps a way to simplify it would be to make concentric bands around the flange, with the same thickness as the rope diameter, and applying deceasing values to the bands. The bands can be made easily in Onshape using the split feature, and parametrically tied to a #ropediam variable like this:



The force profile wouldn't be a smooth ellipse, but rather "stepped" if you know what I mean.

bill_daniels

Thanks a lot!  I'd be very happy with a stepped function.

fastwayjim

Here ya go, @bill_daniels...

I prepped the CAD in Onshape using a derived part, and direct modeling techniques, and taking advantage of symmetry. HERE is the model.

Then, in Simscale, I added a torque to the drum and concentric bands of decreasing pressure on the flange. I put in some random numbers. The mesh needs to be bumped up, but this is a solid working model I think. I hope it helps. Here is the link to the project: https://www.simscale.com/projects/fastwayjim/rope_drum/

bill_daniels

Thank you so much for the time you've put in on this.  I have changed the geometry significantly but this morning I'm having trouble getting it meshed.

bill_daniels

BTW, here's some actual numbers. 

The radial pressure on the core can be as much as 13,789,514 Pa with the 1st rope layer exerting as much as 25% of that laterally against the flanges.  The torque on the drum can be as much as 2,286 Nm applied to the drum through a "dog clutch" using the 8-hole pattern in the end flange.  The rope is Dyneema (UHMW-PE fiber rope) with a 4.7mm diameter.  When full the drum can hold as much as 3,000m of rope.  Typically, only half the total rope will be wound on at high tension and the remainder at only about 100N.  The maximum rope tension is 1,000 DaN limited by a weak-link in the rope.  The radial force on the drum is calculated using the formula:
F=2T/Dd where T is the rope tension, D is the drum diameter and d is the rope diameter.  The area of the drum surface is divided by the result of the formula times the number of wraps in each layer to get pressure.

If the rope is laid on the drum in parallel wraps, the lateral pressure can be much higher than that as later wraps will try to wedge down between earlier wraps shoving them sideways.  The work-around uses an "X-Wind" pay-on technique where each layer is a wide helix with 50mm of distance radially between each subsequent wrap.  This causes the wraps in each layer to cross those in the previous layer at about 3 degrees which prevents wedging and significantly reduces the lateral force on the flanges.

Again, thanks for your help

bill_daniels

Finally got it meshed with the help of a SimScale tech - I had left a tiny crack in the geometry.  The max von Mises stress on the drum core is 1.629e+08 Pa which is well within the yield strength of ordinary steel.

So much for the core, now I have to figure out how to use OnShape Slice to segment the flanges.

KenWelch

Bill, you may want to consider SIMSOLID for this analysis. You can find it in the Onshape app store. It does not create a mesh and is much simpler/faster to use than SimScale. It has a very friendly Spot feature that creates loads directly on any shaped geometry region (including concentric washers), so there is no need to split CAD faces, etc. This type of model would run in seconds.  Please let me know if you want more information.  Regards, Ken

bill_daniels

I watched the videos and SIMSOLID looks like a great product but, for my limited use case, the pricing takes it off the table.  To that I'd add that I really buy into the OnShape cloud-based run anywhere on anything.  I'm not even tempted by the 30 day free trial.

brucebartlett

@Ken_Welch I did try SIMSOLID on the 30 trail and was very impressed by how it run and for the right project would consider purchasing, however I am also taken with Simscale because of it's cloud based structure, I originally went searching for Cloud CAD and found Onshape in preproduction after seeing Simscale. Installing SIMSOLID onto a window's based computer is a big downer for me having been completely committed to the Onshape based system. Is their any chance you will becoming a integrated or cloud linked app?

julian_hardy

bill_daniels said:

Finally got it meshed with the help of a SimScale tech - I had left a tiny crack in the geometry.  The max von Mises stress on the drum core is 1.629e+08 Pa which is well within the yield strength of ordinary steel.

So much for the core, now I have to figure out how to use OnShape Slice to segment the flanges.

Bill,
I'm not an expert on winch drums  specifically, but a critical failure mode for hollow cylinders under external compression is often thin shell buckling, which can happen with stresses MUCH lower than the yield strength of the material. Have you considered this failure mode? You might want to do a buckling check in addition to your elastic stress check.

KenWelch

Couple of comments. First, while I appreciate the vision of cloud everything, I think you need to consider productivity and design-analysis throughput.  The model in this thread would be completely modeled and analyzed in minutes with SIMSOLID. The geometry is analyzed directly without modification, no meshing is required and the solution includes validation using automatic adaptive techniques. The traditional FEA process, as described above, requires substantial geometry modification (symmetry, face splitting, etc) and takes days to complete with no mention of mesh validation or solution convergence. How do you know the solution is valid?

Second, cost is extremely affordable. SIMSOLID has a low monthly subscription plan, not unlike Onshape. Considering the time savings, it is the most cost effective analysis solution in the market.

Third, traditional FEA requires potentially different mesh strategies for each analysis. For example, I would not recommend the mesh as presented for buckling evaluation. SIMSOLID meshless solution is adaptive specific for each analysis performed.  

Forth, SIMSOLID analysis is associative to any model update. The analysis definition needs only to be assigned once. Just change the geometry and rerun. SIMSOLID design studies makes it fast and simple to evaluate design change. 

I really feel that time well spent is to evaluate the total analysis process and not only the compute platform. 

bill_daniels

julian_hardy said:

bill_daniels said:

Finally got it meshed with the help of a SimScale tech - I had left a tiny crack in the geometry.  The max von Mises stress on the drum core is 1.629e+08 Pa which is well within the yield strength of ordinary steel.

So much for the core, now I have to figure out how to use OnShape Slice to segment the flanges.

Bill,
I'm not an expert on winch drums  specifically, but a critical failure mode for hollow cylinders under external compression is often thin shell buckling, which can happen with stresses MUCH lower than the yield strength of the material. Have you considered this failure mode? You might want to do a buckling check in addition to your elastic stress check.

Your insight is dead on.  Winch drums have a long history of buckling under the pressure of up to 3000m of rope wound on at up to 1000 DaN of tension.  That can add up to 4.5 meganewtons acting inwards on the drum core so structural failure is a real risk - that's why FEA is critical.  If that isn't a big enough headache, the rotational inertia of the drum must be as low as possible to enable precise winch control limiting the mass one can use in the design.  Glider winch drums require more engineering than some might think.   

It's especially problematic when using Dyneema rope since it has a very low coefficient of friction so it tends to cinch up tight on the drum and a very high coefficient of thermal expansion so if it is left wound tightly on the drum it will shrink even tighter on cold nights.  Good operating practice is to store the winch with the rope loosely wound onto the drum.