constant tension on a winch drum 1

[subsearobot]

I'd like to provide constant torque in one direction (let's say CCW) using a hydraulic motor. because of its small size and low speed, it will likely be a gear motor with a harmonic reducer on the output.

the load on the drum will move in and out, necessitating back-driving the winch at times.

the simplest solution that I have: pressure will be routed to the "A" port, "B" to tank. when the winch is winding in, the motor will turn CCW. but when the load moves away from the drum, the motor will backdrive (CW)(at lease in theory).

If I use a check valve on the pressure side, and a cross over relief, will I get approximately the same torque when the load is unwrapping the drum?

Is there a better way to accomplish this?

Keep in mind, I don't need precise torque control- and this is a mobile application, so size and complexity need to be minimized.
this is not an overboard winch where we need to deal with wave surge etc. I can, if necessary, ramp the speed and direction changes of the load.

cheers all

 

[kcj]

Pressure reducing/relieving valve supplying the motor.
It reduces pressure to supply oil, or vents out back out (relieving) if the motor backdrives.

A cross over relief would pass oil continuously from the pump.

 

[hydromech]

Constant tension on a hydraulic winch is notoriously difficult and the success is directly proportional to the level of financial investment...in other words, if you want precise control...it will cost.

One can be simplistic and do as kcj suggests, but that set up may not work across all scenarios such as cold weather and rapid load changes.

It may be a mobile application and complexity may not appear necessary, but if the drum fails to de-spool the rope on demand, things can quickly go wrong. Aside from that, continual operation will make the system hot and I suspect that the motor will not last very long if you keep stalling it out.

By all means go for the simple option, but just be aware that winch loads are hugely transient and simple hydraulic systems may struggle to 'catch' the load changes and you could run into trouble.

At the very least, you should consider adding an accumulator to smooth out any sudden pressure changes...and stand well back!!!

Adrian

 

[kcj]

Well put.

I took the assuptions at face value, that subsea really doesn't need precision control.
However, it may be a valid assumption, or subsea may not realize that the situation is more complex than appears 'on the surface' no pun intended.
Now that I look at poster name as subsearobot, might indicate the application is far more complex than stated.

If crude really is acceptable, with the valve hysterisis, pressure rises, and winch mechanical friction, then PRR is cheap and easy.
If force is after all important, then pressure transducers, or better yet load cells, and closed loop control make things complicated quickly. But may be necessary.

What are the costs and consequences if the first easy solutions don't work? That may make the design route obvious.

k

 

[PNachtwey]

I would consider an accumulator. The accumulator would tend to keep the pressure drop and therefore torque across the motor relatively constant. The pump would not need to run unless the accumulator pressure drops to low. This would only happen end reeling in a lot. The relief valve would only relieve when the pressure got too high. In between the accumulator is acting like a big s spring. This would reduce wear and tear and power when just holding position and responding to wave action.

If force is after all important, then pressure transducers, or better yet load cells, and closed loop control make things complicated quickly.

That is from the perspective of a hydraulic guy.
A control guy says no problem.


Peter Nachtwey
Delta Computer Systems

http://www.deltamotion.com

 

[hydromech]

Isn't it funny how what starts as a simple hydraulic systems soon rapidly moves towards PID control...

 

[kcj]

Q
That is from the perspective of a hydraulic guy.
A control guy says no problem. Q



Not a problem for the hydr guys either, but usually is a problem for the maintenance mechanics, and a big problem for the co$t accountant$.

But if he needs it, he needs it...

 

[Oldhydroman]

Good evening gents – I’m new to this forum (came across it when searching for something else) but I believe I can throw a little more light on the issues – apologies beforehand for a long post.

This seemingly simple project is actually quite problematic. Firstly you have to recognise that as long as the number of layers of rope on the winch drum doesn't change, that there are no other pulleys or sheaves between the drum and the thing to which the rope is attached, and that the drum has no friction …. then a "constant tension" in the rope equates to a “constant torque” on the drive to the drum. These aren’t realistic scenarios but just bear with me for a while. The control strategy now boils down to one of “constant torque”.

Secondly, you need to pay particular attention to the characteristics of the device which is converting hydraulic pressure to mechanical torque and vice-versa, i.e., the “hydraulic motor”. The wording is unfortunate in this instance because that component with “motor” engraved on its label will sometimes be working as a motor: “you push oil in and make the shaft turn against some mechanical resistance” and it will sometimes be working as a pump: “you turn the shaft and make it push oil out against some hydraulic resistance”. But syntax aside, let’s keep calling it the motor.

Imagine for a moment that the tension in the rope was 1500 newtons, the drum had a diameter of 400 millimetres, the rope was infinitesimally thin and we were on the first layer. That 1500 N force would be acting at a radius of 0.2 metres and the resultant torque on the drum would be 300 Nm.

Now imagine that there was no friction in the drum bearings and we had a direct drive to a perfect hydraulic motor (and no gearbox). Suppose we wanted to haul in the rope with a motor differential pressure of just 100 bar, then the [theoretical] motor size would be worked out like this…
A perfect 63 cc/rev motor (actually 20 x pi cc/rev) would deliver 1 Nm per bar. We want 300 Nm from 100 bar so our motor needs to deliver "3 Nm per bar" which means it needs to be 3 times bigger than the one delivering just "1 Nm per bar". Answer = 3 x 63 = 189 cc/rev.

In practice, the actual output torque is less because of various pressure drops and friction losses in a real motor - so we can describe the motor as having a “hydro-mechanical efficiency” of, say, 92% ... which means we would only get 276 Nm for our 100 bar.

When the same “motor” is running as a pump because the rope is pulling the drum round, we should expect that a [theoretical] torque of 300 Nm would be needed to turn the shaft if the pressure difference across the ports was just 100 bar, but the hydro-mechanical losses in the [imperfect] device mean that we have to put in even more torque. It is fair to assume that the efficiency when working as a pump will be the same as the efficiency when working as a motor so our drive torque of 300 Nm will only generate a pressure of 92 bar (92% of 100 bar). If we had to get 100 bar out of the “pump” before the shaft could turn then the input torque would have to be higher, i.e., ~ 326 Nm

This is the result: if we could design a hydraulic circuit that held the difference between the motor ports at exactly 100 bar (regardless of how this was achieved) then the maximum tension when the winch was hauling in the rope would be 1380 newtons but we would need that tension to rise to 1630 newtons before we could pull the rope back off the drum. So it’s not really “constant tension” at all. Incidentally the ratio of these two numbers is the square of the efficiency: 0.92² = 0.846, 1380/1630 = 0.846.

Now let’s consider the effect of some of the things we previously ignored: There is friction in the drum bearings which means that some of the drive effort we put in when hauling isn’t converted into rope tension (tension is lower than expected) and the friction also means that, when rendering, some of the rope tension is lost to the friction in the bearings so you have to pull even harder to get the motor shaft to turn (tension is higher than expected). In fact, anything which decreases the efficiency of the winch when hauling will work the other way when rendering – increasing the gap between the two tension figures.

What can you do about it? If you want to take the “constant pressure” solution then you must maximise the efficiency of the winch:
1) Make sure the drum bearings are low friction.
2) Use a good quality radial piston motor if you can because these have a high hydro-mechanical efficiency and a very good starting efficiency (if the rope is changing direction then you are continually ‘starting’ the motor)
3) Don’t use a gearbox, increase the size of the motor instead and run it at low speed (radial piston motors are low speed devices)
4) Restrict the number of pulleys and sheaves between the drum and the load (or, if you must have them, make sure they have good quality bearings)
5) Make sure the drum brake doesn’t drag at all.
6) Be generous with the hydraulic pipework sizes between the pressure controlling valve(s) and the motor port and similarly on the outlet/suction side of the motor and the compensator.
7) Use a pressure control valve (or valves) with very little difference in pressure between the reducing function and the relieving function.
8) Size the valve(s) generously to minimise the pressure underride (apparent loss of pressure when hauling at full speed) and the pressure override (apparent increase in pressure when rendering at full speed).

If you want to allow the electrical boys some input then don’t bother trying to get them to hold the pressure constant (using some form of pressure transducer and closed loop control of an electro-hydraulic pressure modulating valve) because, as explained above, perfect control of pressure does not result in perfect control of tension. So get them to fit some sort of load cell and measure rope tension directly – then use the closed loop control system to modulate the pressure in order to keep the tension constant. (Note that this scheme will also compensate for changes in the number of layers on the drum ... but for the most precise control avoid changing layers while the system is active.)

Other issues to bear in mind with your hydraulic circuit:
A) The hydro-mechanical efficiency is a function of motor speed. The figure is lower at very low speeds (when poor hydro-dynamic lubrication regimes apply) and also lower at high speeds (when churning losses, flow related pressure losses and mechanical friction losses predominate). The efficiency is best at speeds about the middle of the motor’s speed range – so picking a low speed motor for a low speed application is a good choice.
B) Some motors need a boosted inlet pressure to allow them to work as pumps. The value of this pressure depends on the speed and any pressure on the inlet port (when pumping) will affect the available pressure differential – there may be an advantage in connecting the drain lines of the pressure control valves to the low pressure port of the motor rather than back to “tank”. If you can’t get a figure from the manufacturer then measure the inlet pressure needed to run the [bare shaft] motor at the maximum speed you want in service and use half of this figure as the boost pressure.
C) Most hydraulic winches will have some form of counterbalance valve on the motor port(s) – you need to find a way of piloting this valve open or circumventing its operation so that a constant tension function can be brought into play.

Sorry for such a long and detailed contribution – hope you made it to the end.

DOL

 

[subsearobot]

I think calling this thing a constant tension drum is a little misleading. we are simply trying to maintain the wrap on the drum, and as such want to keep it taught. if tension varies by 50% depending on the direction of the load, that's fine. We are well within the strength rating of the 'cable'. Again, we are not trying to do heave-compensation on the cheap here. (Hydromech, you've worked on heave-compensated systems, correct?) I have a vague sense of the complexity (and power) required to create a successful system. this system is no where near as cool (at least hydraulically).

There is a capstan (with a brake) ahead of the drum (see block diagram below), which will create the required directional and tensile loading to the free end of the cable. An ROV will be attached to the free end; the ROV will be free-flying. it will nominally not apply any tension to the capstan. the capstan's tension to the ROV is mainly required in a dead-vehicle recovery scenario.

<ROV>~~~~~~[<capstan>~~<drum>~~<transformer, mux, etc>]~~~(30,000 ft fiber-optic / power tether)~~~<Launch/Recovery System (A-frame/ winch on the ship)>

Accumulators would be great, but for the necessity to operate at 13,000 psi (depth pressure is ~10,000psi + hydraulic system pressure).

Oldhydroman, thanks for the contribution! many good points here.

KCJ, while it is true the system (mechanics, power, control and hydraulics) is more complex than i'd outlined, the overarching hydraulic requirements here are very forgiving. NO PIDs here. just good ol' fluid power. The ROV however, that's a different beast.

thanks everyone!

 

[PNachtwey]

Then use an accumulator to maintain tension. An accumulator will reduce wear and tear on the hydraulic pump.


Peter Nachtwey
Delta Computer Systems

http://www.deltamotion.com

 

[Oldhydroman]

Hi Peter

As Subsearobot alluded to, it is very, very difficult to use an accumulator on a subsea hydraulic system because it isn't possible/practical to compensate the nitrogen pre-charge pressure for the massive changes in ambient pressure.

If you really must have an accumulator in your circuit then you need to know beforehand the exact depth to which you will dive - this is so you can set the absolute value of the pre-charge pressure accordingly. One consequence of this is that you can't then test the equipment on the deck just before you start the dive. The required nitrogen pressure might be extremely high and beyond the scope of the normal charging equipment. You will almost certainly require the use of an intensifier.

The accumulator has to be charged up on the surface (where the ambient pressure is just one bar absolute) so the accumulator has to be a version for very high pressures - this makes it extremely heavy which means a huge amount of extra buoyancy has to be added - this makes the whole TMS bulkier which badly affects it's dynamics and might even require a bigger launch and recovery arrangement - which needs more deck space which needs a bigger ship which needs more crew....the ramifications just go on and on (although I may have been a little tongue-in-cheek towards the end there).

Subsearobot; my opinion, for what it's worth, is that you're on the right lines with a pressure reducing/relieving valve to (1) drive the drum motor [to reel in the umbilical] and (2) allow the drum motor be back-driven by the action of the capstan [when the umbilical needs to be paid out]. I would also go for a spring applied (pressure-to-release) brake on the drum motor [you already have one on the capstan] to prevent the 'cable' on the drum from going slack when the power is off.

Just remember "worse things happen at sea". Oh sorry, I don't suppose that's quite as helpful as the other advice on this forum.

DOL

 

[hydtools]

Can the accumulater be a spring-loaded piston type vented to atmosphere. As ambient 'atmosphere' changes so compensates the accumulater to local pressure + spring load.

Ted

 

[Oldhydroman]

Hi Ted

This is basically what is used as a subsea reservoir (or 'compensator') but those devices only raise the oil pressure by about 0.5 bar above ambient (some go a little higher).

If you try to get the spring force to deliver pressures much above this then the spring starts to get huge (and it has to be made from something that doesn't corrode). The housing gets very big and heavy as well (in order to accommodate the compression loading of the spring and its compressed length).

If you are going to have sea water in one side of the housing then the construction needs to be something like 316Ti (1.4571). And because the thing has to be so big when you are then required to make it out of an 'exotic' alloy it becomes extremely expensive. I know ceramic coating can be considered but it's usually the weight and the space requirements that kill it off as a possibility.

DOL

 

[subsearobot]

Oldhydroman, thanks. I was not very clear as to why we don't typically use accumulators sub-sea. I'm sure your explanation was helpful to readers on the surface.

For another project, I looked into feasibility of a spring charged piston type accumulator. ouch- it was almost heavier and more expensive than the vehicle we were looking to support. And we were not going very deep for that one (1500m).

size and cost considerations are especially true for this present project, where we are going quite deep. would need a support ship just for the TMS!

"worse things happen at sea" couldn't be more true. when saltwater is involved, Mr. Murphy (of Murphy's law fame) goes into overdrive!

Good point regarding braking the drum on loss of pressure. we will do that.