designing for torsional load
designing for torsional load
(OP)
I'm a chemist functioning as a Process/Materials engineer and have a problem that is well beyond my mechanical engineering skill set. We have a continuous production process where stranded aluminum or copper conductor is insulated. The conductor naturally twists over sheaves and a twisting motion is imposed during one section. I need to design a method to connect two ends of the conductor with a device that can handle the torsional load imposed in the process and be equal to or less than the diameter of the conductor. I don't know how to calculate the load imposed by the process or determine what load the device can handle. I know the conductor will rotate 8 times over 600 ft. worst case assuming it is fixed at each end. It isn't totally fixed at each end but if I can design as if it is I have a safety factor built in. Can the problem be simplified for example if the device is fixed relative to the longitudinal axis of each end of the conductor, does the device simply have to be the same or stronger than any section of the conductor? Sean





RE: designing for torsional load
That sounds impossible.
Consider also that if you take a cable that wants to twist, and restrain both ends while whatever causes the twisting is still going on in between you will wind up with a pretzel.
RE: designing for torsional load
RE: designing for torsional load
Imagine the cable moving continuously into fixed point A (capstan) through fixed point B (capstan). At a point close to B a twisting motion is applied to the cable. The fixed points don't allow the cable to twist but do pull the cable through so only a fixed amount of twist is applied to any point in the cable (in reality there is some twist allowed through the fixed points). What I need is something the same or smaller in diameter that connects two pieces of cable end to end and will twist with the cable but not more than the cable (fixed radially) and not break under the torsional load. Make sense?
RE: designing for torsional load
RE: designing for torsional load
RE: designing for torsional load
What do you use to splice the same size cables together easily enough now?
RE: designing for torsional load
RE: designing for torsional load
RE: designing for torsional load
RE: designing for torsional load
RE: designing for torsional load
RE: designing for torsional load
Read back through this post. With each of your posts we get a new, not previously disclosed requirement.
Take some time to read some of the threads here to get a better feel for what makes the different between a well asked question and a poorly asked question.
RE: designing for torsional load
So, how do I calculate the torsional load on the system and how can I relate that load back to standard referenced materials properties so a design solution can be found? Once this is established I will be happy to share the overall problem and ask for potential design solutions. I have been looking for off the shelf solutions to this of and on for about 5 years now. If there is an off the shelf solution I will be surprised but extremely grateful and ecstatic.
RE: designing for torsional load
Answer: Yes.
"So, how do I calculate the torsional load on the system and how can I relate that load back to standard referenced materials properties so a design solution can be found?"
Answer: Start off with a university course in statics, then dynamics. Follow this up with a materials course. You may want to include calculus. When you have the background, you will be able to calculate the torsional load on the system.
RE: designing for torsional load
RE: designing for torsional load
If you'd had a materials class or paid attention to it, you would already know "how to relate the normally referenced tensile yield strength to the torsional or shear yield strength" for materials.
RE: designing for torsional load
For the record, I know the rough and dirty estimation is the shear yield is 50% of the tensile yield. I was looking for something a little more exact or data that says 50% is worst case.
RE: designing for torsional load
So I'd suggest a test of some reasonable length of cable. Obtain torque v. twist.
From there, I suspect you could get away with using equations for solid shafts. Knowing torque, angle of twist and length you can calculate the polar modulus of inertia of the "equivalent solid shaft".
Then design your connector with an equal or greater polar moment of inertia.
RE: designing for torsional load
A capstan to me means three wraps of the rope or cable around the capstan so it can be used to pull the load or lift the sail. And, I’ll bet that’s not what you mean, so start there, and explain. What you see as a capstan, we can’t even imagine, and we can’t see it from here. What applies and how is it applied, on both the cable pulling tension and the torsion/twisting action, that’s 2. Is this a cable pulling operation from floor to floor, that’s 3. What kind of cable is this, fiber optic, computer cable, how’s it made, is it a twisted/layed up cable which twists a bit by the nature of its construction, when you pull it over the capstan, that’s 6, 7, 8 and 9? There aren’t many simple answers to complex problems. And, if you think this is a simple problem with all the variable you have already laid out, and the limitations you have put on it; that just shows that you don’t understand your own problem very well, and sure aren’t describing it very well. You can weld some of these, but they have to be taken apart when they reach the upper fl., that just doesn’t wash. Can you lose or gain wire length in the separation? The connector must be smaller than the wire dia., that’s a tall order and a dream. Are some wires in the cable splices and others not, that’s 10? I’d wrap the cable in electrician’s tape, and unwrap it at the upper fl. Would something like the old Chinese finger lock/puzzle work? There are cable pullers like that on the market. We don’t even know what size the cable and individual wires are, or what size duct you are fishing this through.
MintJ. is right on the money above. So many of the questions we get here are so pie-in-sky that the OP’er. can’t begin to describe them sufficiently to illicit some meaningful discussion. We are all guessing, wasting our time, and getting nowhere; and if by chance someone guesses crazy enough to find a solution, they get a shout-out. You’ve got several smart people posting on this thread, and your just not getting through to them, and that’s not for their lack of effort. These problems, on this forum, are a real pain in the arss, because we have to ask so many questions before we tease another important fact out of you, and you didn’t even know it was important. You haven’t described what it is or how it works well enough so we can ask some meaningful questions, we’re all just fishin in the dark. These discussions would be so much more productive if we could see some properly proportioned sketches, with dimensions, loads, etc. and some photos of the system, so we knew what you where dealing with, what your idea of a capstan was. If we could be standing together looking at the problem, this give-n-take or exchange of ideas would flow much better. And, if the OP’er. gives out important info. in drib-n-drabs it just frustrates the hell out of a bunch of smart people trying to be helpful.
Shylanel.... this is not specifically directed only at you, it just happened to finally get posted on your thread. I am not trying to pile on here. This seems to be all too common an occurrence on these fora, the issue of ill defined questions or problems. You haven’t presented a simple problem, and you certainly have not presented it well. Again, WE CAN’T SEE IT FROM HERE.
Good Luck
RE: designing for torsional load
Reasons to believe you don't know what you are talking about:
I have been looking for off the shelf solutions to this of and on for about 5 years now.
I obviously didn't explain it well.
Why don't you ask me for the information needed for the calculation?
This is where you led me to think you are angry and stupid:
I did not start out asking for a design solution. If I had, the post would have been much different defining the various requirements of the design.
This is where you made sure:
spew insults sit on a pedestal keep your pompous arrogance to yourself
RE: designing for torsional load
I've already started figuring out design options. I'm trying to put some engineering data behind it instead of engineering judgment with experimentation.
The overall problem and process is a very lengthy explanation. Since everyone wants to know, I'll try to give a cliff notes version. We make medium and high voltage power cable from 2AWG to 2000 MCM, 5KV to 69KV using a Continuous Catenary Vulcanization (CCV) process. The overall line length is over 1800 ft spanning over 4 stories. The smallest cable is less than 0.5" and the largest > 3". The process is continuous but the conductor comes on reels (not continuous) so there has to be a method to join each end of the conductor without stopping the line. This is done with welds or crimps depending on the conductor type. When in continuous mode, the crimps or welds do not have to be disconnected. They simply have to be smaller than the conductor so they will pass through the extrusion tooling. When changing sizes of conductor, the extrusion tooling has to be changed requiring a stop in the process but you don't want to waste the strand between the extruder and payoff so the different conductors have to be joined at the payoff. When the new size gets to the extruder, the line is stopped and the two conductors are disconnected. The old conductor is pulled through the tooling. The tooling is changed. The new conductor is then put through the new tooling and rejoined with the old conductor for the line to be restarted. Knots can be used at the extruder but they are notoriously unreliable, take time and are difficult to make small enough as they have to go through rubber seals that can only open so much. The entire connection also has to be capable of sealing water at 150 psi from the strand. This is can be handled with what is called a water block weld when possible. When the insulation thickness is large relative to the conductor OD we apply a twister after the seals to rotate the cable as it exits the extruder to prevent the hot insulation from drooping. This keeps the dimensions in spec. There are multiple capstans in the process to pull the cable through the line. The primary capstan is a belt capstan which is basically a very large sheave with a rubber belt on it. A second belt is held in tension against the sheave. The cable passes between the belts and is pulled with friction. The other capstans are all caterpillar capstans which pull the cable through between two belts propelled with a caterpillar mechanism like a tank but these are horizontal. This is just the mechanical path of the process leaving out other auxiliary steps not germane to this problem. I also haven't touched the extrusion or curing process' which are the real challenge to making the cable. This cable is made by the mile by various companies all over the world every day. I can continue making the cable this way and everything will be fine. I'm trying to find a better and more efficient method that will reduce changeover time, increase reliability and reduce scrap.
I want to use a crimping tool normally used for hydraulic hose that gives a stronger, smaller and faster crimp than what's normally used in this industry. This style is used for EHV cable. This crimper can most easily be installed and applied at the payoff for various reasons. To take full advantage of this crimper I want to develop a device that can connect two different size cables at the payoff, be disconnected at the extruder and then reconnected after the extruder during changeovers. There exist devices that allow for rotation. That is fine except when I want to apply the twister. If I can develop a device that is rigid and as strong as the conductor on either side then I can put the twister on immediately at startup and save several hundred feet of scrap. The working thought is to have crimps for each size conductor but with a common interface to each other or a third tool. I can then crimp the end pieces on each conductor, connect them, run the cable to the extruder, stop the line, disconnect the two ends, pull the old conductor through, change the tooling, pull the new conductor through, reconnect and start the line. Again whatever device is designed to connect the two crimps has to be smaller than the biggest conductor to make it through the seals.
In general when changing sizes it will be from a smaller conductor to a larger conductor. We obviously won't be going from 2 AWG to 2000 MCM either. We will jump several sizes at once though. I don't need one device that will work from 2AWG (~0.25" OD) and 2000 MCM (~1.6"OD). I expect to have 4-6 sizes covering various ranges of conductor to span the entire range. To know what size range of conductor each size device can handle I will also need to know what load they can handle. The crimps will be throw away. The connector between the two will be reusable though not indefinitely.
I hope this is enough detail for everyone to digest and understand what I'm asking. If you care to offer a design ideas in addition to the load calculation I will be grateful.
RE: designing for torsional load
I have to admit I didn't expect to find trolls of your nature on an engineering forum.
RE: designing for torsional load
You've stated that you can crimp.
You've stated that you can crimp steel tube.
You've stated that crimped steel tube connections work.
You've stated that you can throw crimps away.
So why won't a steel tube that's sized for the smaller cable on one end and expanded to the size of the larger cable on the other end be crimped on and be done with it.
I'm imagining the device that you seem to be imagining as obscenely expensive and unreliable - if it's even possible.
RE: designing for torsional load
The following picture is something that is used to pull these cables overhead in the field. They aren't very expensive and I would use them in a heartbeat if I wanted something capable of rotation. I'm trying to get past that. Even if they didn't rotate I would be concerned about uneven load on the pins during rotation.
These don't have to be dirt cheap either. They will pay for themselves at upwards of $500 apiece. I'm looking for around $100 or less though.
http://65.49.46.9/media/products/grips_swivels/swi...
RE: designing for torsional load
If it is enough $$ to warrant, just use two crimpers. Crimp on the above picture (tailpipe adaptor tube I presume, great visual) at the unroller, cut it off at the extruder, and crimp on a new one after the extruder. Only throw away 10 ft of product.
Forces, I think you have to test. You are dealing with torsional but it appears catenary tension forces also. Getting x degrees of torsion in L length depends on the cable tension as well as the torsional stiffness of the material. certanly solid acts different than stranded
RE: designing for torsional load
http://origin-m.te.com/;jsessionid=Dj4SGvwLDY3rmNQ...
RE: designing for torsional load
With the forces you are describing, I would not look at this as a reusable tool connected to the end, but as a crimped/swaged/welded device that is cut out later and disposable.
RE: designing for torsional load
TL/jG = torsion amount. T is the torque, L is the length, j is the polar moment of inertia, and G is the shear modulus of the material. Torsion is measure in radians as long as the units cancel out. For copper G is listed as 48GPa. j -> http://en.wikipedia.org/wiki/Second_moment_of_area and http://en.wikipedia.org/wiki/List_of_area_moments_... ; pi * R^4/2 for a circular section.
I do wonder how the control system handles it when a larger diameter transitions to a smaller one. The much stiffer larger diameter will try to backdrive the built-up torsion into the short span of smaller diameter or the change-over from copper to aluminum. I wonder why they don't coordinate the turn from one end to the other like a rotissorie. Not for five years, so I'm not expecting an answer, but I'm guessing it's because the torsion load is miniscule and can be ignored.
Don't feel less human, you are simply undereducated. You should find a licensed structural engineer to review whatever you come up with and some anger management support as well.
The strength requirements are primarily driven by how tight the cable is in the catenary. Those will exceed the torsion load by one or two orders of magnitude.
The question you should have asked - how much tension load is on 1800 feet of cable with x feet of droop that weighs y pounds per foot, assuming both ends of the catenary are at the same elevation? http://homepage.math.uiowa.edu/~stroyan/CTLC3rdEd/... explains how calculus is used to figure this out.
Only caring about torsion in this application is a good way to get people killed or do a lot of damage to property.
RE: designing for torsional load
http://www.engineeringtoolbox.com/torsion-shafts-d...
Cheers
Greg Locock
New here? Try reading these, they might help FAQ731-376: Eng-Tips.com Forum Policies http://eng-tips.com/market.cfm?
RE: designing for torsional load
Are you intentionally putting the twist in the conductor cable every 75', and is that a full revolution or something less? Why are you doing this and what torque are you applying? Or, does the pulling equipment (capstans right?) cause the twisting because of the lay and construction properties of the cable? This same twist is induced in the conductor under load (self wt., ice, wind), when it is hung and it just tends to tighten the lay or cable construction a bit, it is not a major stress issue on cable under normal conditions. I think I’m getting a vague idea what some of your issues are, contact me rwhaiatcomcastdotnet. You must have to stop things a foot before the extrusion tooling/die, make the disconnect, pull the lead cable through the die, change the die, pull/push the trailing cable through the die (how, with some intermediate device?), make up the cable end connection again, and then continue the process on the new reel of cable, right? And, you do waste those few feet of cable in the end. What are your pulling forces? You should be able to determine them or have some feel for them. Crimping or swagging should work as long as the crimp sleeve has an OD slightly smaller than the insulation dia. I would take MintJ’s crimping sleeve, cut it at the dia. transition, and make it a crimped end cap for each cable dia., with a cap pl. on the end. On one end cap I would have a single pin pl. and on the next end cap I would have a mating double pin pl. and you pull the pin/bolt and nut to make the disconnect, and this would provide some torsional strength and pulling strength to be determined. Gotta know more about what you have there to comment further.
RE: designing for torsional load
kcj - The scrap cost ranges from $1 to $30 per foot of cable. I can easily justify a second crimper however the only crimpers with enough tonnage are 'in process' crimpers in that the conductor must run through the crimper after being crimped. There is no place to put this after the extrusion head as there is a vulcanization tube pressed to the crosshead that is much larger in OD than the maximum ID of the crimper. Also there is no such thing as 10 ft of startup scrap with this process. After starting it takes on a great day 100 ft to get the vulcanization conditions and product dimensions in spec before it can be called good. On large KV products or if we can't get a good water seal, there is a minimum of 600 ft of scrap after start. On bad days it takes 2000 ft to get a good start. The purpose of this is to keep the startup closer to 100 ft.
I'm expecting the solution to be rigid but has to go over 10 6' OD sheaves prior to the extruder so it would have to be strong enough to handle that. There would be limits on the length. The goal is reusable and I've seen it done with rotation allowed but disposable is doable if necessary.
Mint - something like that scaled up could work. How would the two ends be connected in between?
Dave - As was described the line is 1800 ft over several floors. The catenary is only 600 ft of this length and is about a 100 ft drop. I'd have to look up the exact drop for this line. It travels through a 8" tube under temperature and pressure shaped in a 26° catenary drop. The control is handled between the capstans described in this case based on the position of the cable in the tube which is measured inductively. There are limitations when going between extreme sizes. In those cases the cable has to be dragged down the tube with extreme care to avoid breakage. The twister is not applied in these situations. The tension between the cables is essentially the weight of the cable which can range from 1 to maybe 8 lbs/ft. That is a measured value-not theoretical. The torsional load is significant as I've seen steel bolts get sheared in two with only the normal torsional load from traveling over sheaves and the strand. They sheared when the rotational motion designed failed. Adding torsional load is significant.
dheng - yes you pretty much have the idea except it is the tip of the extruder that is the problem. It is 10 to 30 mils larger in OD than the conductor depending on the size. There also isn't a crimper with enough tonnage that I can use after the extruder due to the vulcanization tube. The pulling force is essentially the weight of the cable between the capstans.
There is a small natural twist between the two capstans. When the insulation thickness is the same or greater than the OD of the conductor, the insulation will droop out of the extruder. To counteract that, the cable is rotated in the direction of the lay to take advantage of gravity. This is why vertical CV lines are built for EHV cable. I can wait until the new cable is out of the pull out capstan before turning the twister on but that almost guarantees an extra 300 feet of scrap. That's why I want to design something that can handle the extra load. Generally the twister is set at 1/4 turn every 20 feet or 1 full turn every 75-80 feet. This is from experience and not something I've ever experimentally determined as the optimal twist. The twister is essentially a caterpillar capstan where one belt can be adjusted at an angle.
On the strand - I understand strand will behave differently than a solid rod. I could be totally off but I think in this case it will behave closer to a solid than strand most folks think of. This strand is alternate lay in all but the smallest. It is not bunch strand where all the strands are twisted at once. It is a layered strand where each successive layer is the opposite lay so in a 3 layer strand the inner layer will be left hand lay, the middle layer right hand and the outer layer left hand again. The larger cable can have upwards of 6 passes of alternating lays. The largest conductors are actually segmented cable. Each layer is also compressed significantly so friction should eliminate movement between the layers relative to each other. The end result in my mind is that each layer will counteract the natural tendency to shorten the lay experienced in a unilay or bunched strand. While more flexible than solid I would think it would behave more similar to a solid in terms of torsion. I could be wrong. I'm thinking that the torsional load of a solid will be greater than the strand so if the design can handle the load on a solid rod it can handle the load on the strand. If the values for solid are known I don't have to figure out how to measure the load on the strand.
General CCV line
The first picture is a vertical line for EHV cable. The second picture further down is a CCV line.
RE: designing for torsional load
Now for the ends to be held stationary under the calculated torque and I am assuming (which can make an A** out you and me) that both ends are riding over grooved sheaves you may want to consider grooved clamps over the top of the exposed wires within their sheaves. Friction would be the holding force and a sketch of this idea is in the attachment. Obviously the clamp is nearly a point load on the wire, however, you may be able to lengthen a little the point of contact without deforming the ends over the circumference of the groove. A small contact area from the clamp is probably all that is needed to maintain the calculated torque on the wire.
RE: designing for torsional load
RE: designing for torsional load
Conductivity or ampacity similat to the cable being processed is not required?
RE: designing for torsional load
RE: designing for torsional load
RE: designing for torsional load
I'm not fishing for a solution. All I ever wanted was how to do determine the torsional load and if my assumptions were correct. All the information needed for that answer was in the first post. I have the calculation I was looking for. Thank you to all that contributed. Any brainstorming ideas for a solution are welcome and will be considered but I am not nor was I ever here to solicit a solution. If someone figures out how to do it and wants to keep it to yourself I wish you luck. Feel free to contact me for testing and I may be able to suggest a few companies that may be interested in buying it from you...Greenlee and Sherman & Reilly come to mind. There's also Davis Standard, Troester and Maillefer who are the primary suppliers of CCV lines internationally. This isn't an industry known for paying a premium for technology that can be done or duplicated in house without risk of detection for patent infringement. Most in this industry have no idea how patents work and are in awe of anyone with a patent. Without plant inspection how can a company determine if the technology is being used? Nothing in the final product could be tested. You would end up paying a lot of money with limited potential for profit on an unenforceable patent.
RE: designing for torsional load
That's my theory, but maybe someone can point out if my approach is flawed.
RE: designing for torsional load
The tensile load in our process is essentially constant at 12.5 MPa and the torsional load (using the adjusted modulus) is always greater than the tensile ranging from 61% greater on small strand to nearly 10X greater on the large strand.
This can then be translated to determine minimum diameter steel is needed to work on a given conductor diameter. A better way to put it is given an OD of steel how much larger of an OD of conductor can it safely pull. The smaller sizes can handle up to an 80% increase in relative OD of the conductor. This decreases to about 40% for the larger conductors. This gives us plenty of range to work with.
This all passes the sniff test with experience. I'd feel better if all the data was known or measurable but as usual we aren't given those resources in this industry.