# Unsolved Engineering Problem: Runaway Anchor Drops

• anorlunda
In summary, it seems that there are several common problems with runaway anchor drops, and that the design problem remains unsolved.
anorlunda said:
By the way, this video illustrates the chain runaway scenario on a micro scale. It doesn't help much in our discussion other than to illustrate how the distributed masses and momenta play roles. Note that inertia of the ascending portion of the chain is also significant.
Is there a significant length of chain hanging off the backside of the capstan? I assume its piled in the bottom but hanging vertically between there and the capstan.

russ_watters said:
I don't understand what you are saying here: why would the anchor brake need to stop the ship? ...and why as soon as it touches bottom?
Somebody hasn't been studying their history. The last stand of the USS Missouri in 2012 teaches us how to deploy an anchor in an emergency.

Last edited:
russ_watters
anorlunda said:
As Russ Waters said, a "sea anchor." I don't think you can make one rugged enough to survive. Runaway can start both before and after the anchor hits bottom.
I guess the question is how large, and durable does it need to be to limit the decent velocity to a safe range. I imagine one of the sea anchors fixed to the anchor with a small buoy on top to keep it upright.

I get the following equation of motion, for what seems like would be worst case scenario...complete primary brake failure.

$$\lambda \left( x+ l_o\right) \ddot x = \lambda g \left( x-l_o \right) - \beta { \dot x}^2$$

Where:

##\lambda## is the chains mass per unit length
## l_o## is the length of chain suspended on the back of the capstan
## \beta## is the total drag coefficient
## \downarrow^+ x## is the length of hanging anchor

If I had the figures, I could try to solve for the projected area of the sea anchor, and force on it required to limit its velocity to some threshold.

I know you said they can start after the anchor has landed too, but how often is that the case when it is a controlled decent? If its uncontrolled, and the anchor hits bottom while the chain is carrying a lot of momentum, then the "sea anchor" doesn't help, but if the sea anchor always provides some level of controlled decent is it likely to happen?

EDIT: This model isn't quite correct either. I think there is supposed to be a linear ##\dot x## term in there as well. I remember having a conversation in PF before about the solution being somewhere between Work/Energy and Impulse/Momentum for the rising chain (on the back of the capstan) which is accelerating the links in the pile from rest...I didn't account for that

EDIT: Maybe its ok after all. That was a slightly different problem where the length of the chain rising was increasing. This problem the length of the chain rising is constant.

Last edited:
erobz said:
Is there a significant length of chain hanging off the backside of the capstan? I assume its piled in the bottom but hanging vertically between there and the capstan.
Chain is stored as low as possible to keep the ship's center of gravity low. So if sea level is 0, chain is stored at 0 to -15 feet. The hawse pipe where chain leaves the ship is almost at the top deck, +50-70 feet, anchoring depth of water on the continental shelf is typically 100 feet, so yes, the ascending part is a significant fraction.

Re-watch the chain fountain video carefully. At 17 seconds, the man let's go of the chain with his fingers, at about the same level as the bottom of the glass beaker. But the chain has a high velocity at the instant he let's go. The fountain starts from there. That tells me that it doesn't matter how long the distance is to the floor, but rather that velocity is the key parameter. It would be interesting to repeat that experiment with the glass beaker on a table, to see if the fountain continued if the final resting place of the chain is on the table top.

Regarding the sea anchor. First, I agree with @russ_watters, the runaway can start after the anchor hits bottom. Indeed, in some of the cases on video, the chain first stops then moves again before runaway. I interpret the stop as the time when the anchor hits bottom. Therefore, to anchor in 100 feet of water, you need another sea anchor every 100 feet of chain.

Sea anchor durability: I'm most worried about dragging the chain across the bottom. That tends to rip off anything attached to the chain.

russ_watters
This video is better. The capstan wheel is visible. The chain stops several times. The men on the brake are seen turning the wheel to loosen the brake for minutes. Finally, runaway begins. Here are significant events in the video.

There is so much dust in the air, that the men can not be seen trying to tighten the brake after runaway begins. However, I think they may need to turn that wheel the other direction for 3 to 4 minutes before the brake starts braking again. But the chain is lost in less than 1 minute.

I also found some data from the USS Nimitz. I'm going to try to estimate chain velocity from the videos, and then calculate braking power needed to prevent a runaway.

erobz and hutchphd
Anchors Away!
(Sorry couldn't resist the pun)

anorlunda
anorlunda said:
This video is better. The capstan wheel is visible. The chain stops several times. The men on the brake are seen turning the wheel to loosen the brake for minutes. Finally, runaway begins. Here are significant events in the video.

View attachment 313228

There is so much dust in the air, that the men can not be seen trying to tighten the brake after runaway begins. However, I think they may need to turn that wheel the other direction for 3 to 4 minutes before the brake starts braking again. But the chain is lost in less than 1 minute.

I also found some data from the USS Nimitz. I'm going to try to estimate chain velocity from the videos, and then calculate braking power needed to prevent a runaway.

View attachment 313231

I think I’ll use this data and update the model to include the buoyant force and the mass of the anchor, neglect drag for now and try to calculate the speed for some initial conditions. With any luck I’ll be in the ball park of what you find in the video.

anorlunda
Brake fade is another possible cause of chain runaway. The chain pull at the brake multiplied by the chain speed is a LOT of power, and thus heat generated in the brake. That heat is generated at the contacting surfaces between the brake pads and brake drum. The brake drum is almost certainly a steel alloy, possibly cast iron. I don't know what those brake pads are made from, but they are very likely a non-metallic material.

The surface temperature at the braking surface is the result of the heat generated minus the heat removed by conducting through the brake pad and drum. If the brake pad surface gets hot enough, it can vaporize. This could be the source of the (dust) (smoke) seen during the runaways. If that is what is happening, there is gas generated at the braking surface. That gas is a lubricant which causes the friction coefficient to decrease dramatically.

I once experienced brake fade. I was driving my 1961 Rambler American like a typical overconfident teenager. I came over a hill at 80 MPH, and saw a stop sign at the T intersection at the bottom. Standing on the brakes got the speed down to 25 MPH, at which point braking action ended. I was able to get around the corner by putting two wheels in the ditch to pull it around while somehow not hitting the stop sign. When brakes stop braking due to brake fade, they stop almost instantly. Brake fade is not a gradual process.

The last time I was in a private jet, the pilot landed a little fast on a short (for that airplane) runway. The brakes were smoking after he stopped. He claimed that it was a normal landing. I did not believe him, partly because that was the only time I saw those brakes smoking.

Lnewqban and anorlunda
I guess the main direction will be the missing redundancy and maybe an elevator style speed limiter

Last edited:
Lnewqban
May be a silly question, but would it make any difference at all if some kind of pattern is stamped on to each link of the chain that would maximize turbulence and hence drag? Analogous to engineering the surfaces of golf balls, etc ?

I haven't looked at all of the videos. but the several that I did see have 1 thing in common:
Operator error. When the chain stops due (presumably) to the anchor reaching bottom, they continue to back off the brake. When motion resumes, they are backed-off so far that they don't have a prayer of catching up. There isn't anything wrong with the equipment - it just isn't designed to recover from a situation that never should have occurred.

This seems like an idiot-proofing exercise.

A couple of things that I'd look at:

The braking mechanism shouldn't be able to back off so far.
A hydraulic speed limiter might be simple enough to prevent the 'out-of-control' situation from ever happening. Drive a pump from the capstan and friction-limit the flow / waste the energy into a water-oil HeatX. Properly executed, this arrangement might also be used as a way to measure 'deployed weight' - could be useful for determining the (apparently) confusing transition from lowering the anchor to just adding scope.

Or not.

erobz
..
Dullard said:
When the chain stops due (presumably) to the anchor reaching bottom, they continue to back off the brake. When motion resumes, they are backed-off so far that they don't have a prayer of catching up.
Agree. But it the accident is repeated so often, year after year, on different ships with different operators, that we must look at system design flaws.
Dullard said:
The braking mechanism shouldn't be able to back off so far.
I agree. In fact, I suspect that a single 360 degree turn of that wheel controls the full range from full brake to no brake.
Dullard said:
A hydraulic speed limiter might be simple enough to prevent the 'out-of-control' situation from ever happening.
Agree again. In fact that's what I'm going to explore next. I'm thinking something like a torque converter. The key to success is the shape of the brake force versus speed curve.

Dullard said:
I can imagine simple, reliable solutions
The thing is mounted on the bow of ships and got permanent battering of water, salt and every harm seas can offer.
I think it's a requirement that maintenance could be done by excess amount of grease only and field repair is by hammer and wrench

anorlunda and jrmichler
Here is what I get for the chain velocity in the absence of drag, either from the capstan or fluid drag. The velocity is almost linear in time.

I don't know if this is of interest anymore, but there it is.

Rive said:
The thing is mounted on the bow of ships and got permanent battering of water, salt and every harm seas can offer.
I think it's a requirement that maintenance could be done by excess amount of grease only and field repair is by hammer and wrench
This. But I would change it slightly to say BIG hammer and BIG wrench.

Rive
Here's my thinking. I would use a water brake as a velocity limiter, or as a runaway preventer.

https://en.wikipedia.org/wiki/Water_brake

And this video,
https://en.wikipedia.org/wiki/File:Tech-Talk_Animation_on_How_Water-Brakes_Work.webm

The torque is controlled by the flow of cooling water. I would use a flyball governor to monitor capstan wheel RPM, and use the governor to control flow of water to the brake. The goal is to provide a soft limit to capstan wheel RPM; and to do it mechanically with no aux power.

The harder question is the required rating. I'm not sure how to calculate that. So partially arbitrarily, I'll choose the case of decelerating 30 feet of chain from 25 to 5 m/s velocity within 2 seconds. Here are the calculations using @erobz 's numbers

 Initial V 25​ m/s Final V 5​ m/s Delta V 20​ m/s Period 2​ s Avg Acceleration 10​ m / s^2 len 30​ m weight 269​ kg/m total mass 8070​ kg Initial K.E. 2521875​ Joules Final K.E. 100875​ Joules Power 1210500​ joules/sec MW Power 1.2105​ MW HP Power 1622​ Hp

I tried to find a 1.5 MW water brake on alibaba.com, but I find nothing similar. Perhaps I'm using the wrong search terms.

@erobz, @jrmichler , I'm not a M.E. Do my calculations appear correct?
Rive said:
The thing is mounted on the bow of ships and got permanent battering of water, salt and every harm seas can offer.
That's true. Ocean sailors must deal with salt and corrosion all the time. We use bronze alloys or for many objects. For example, the windlass I used to raise anchor chain on my sailboat was 100% bronze, lubricated with special water resistant grease.

anorlunda said:
Here's my thinking. I would use a water brake as a velocity limiter, or as a runaway preventer.

https://en.wikipedia.org/wiki/Water_brake

And this video,
https://en.wikipedia.org/wiki/File:Tech-Talk_Animation_on_How_Water-Brakes_Work.webm

The torque is controlled by the flow of cooling water. I would use a flyball governor to monitor capstan wheel RPM, and use the governor to control flow of water to the brake. The goal is to provide a soft limit to capstan wheel RPM; and to do it mechanically with no aux power.

The harder question is the required rating. I'm not sure how to calculate that. So partially arbitrarily, I'll choose the case of decelerating 30 feet of chain from 25 to 5 m/s velocity within 2 seconds. Here are the calculations using @erobz 's numbers

 Initial V 25​ m/s Final V 5​ m/s Delta V 20​ m/s Period 2​ s Avg Acceleration 10​ m / s^2 len 30​ m weight 269​ kg/m total mass 8070​ kg Initial K.E. 2521875​ Joules Final K.E. 100875​ Joules Power 1210500​ joules/sec MW Power 1.2105​ MW HP Power 1622​ Hp

I tried to find a 1.5 MW water brake on alibaba.com, but I find nothing similar. Perhaps I'm using the wrong search terms.

@erobz, @jrmichler , I'm not a M.E. Do my calculations appear correct?
That's true. Ocean sailors must deal with salt and corrosion all the time. We use bronze alloys or for many objects. For example, the windlass I used to raise anchor chain on my sailboat was 100% bronze, lubricated with special water resistant grease.
I’ve got the illness that shall remain unnamed. It’s starting to wear on me. Last night my fever broke long enough I program that solution. So I’m a little more foggy than usual.

I think you need to consider the mass of the anchor as well? Also the final mass of the chain will be the initial hanging mass + the anchor+ the additional 30 m of chain at 5 m/s. You also have to change the kinetic energy of the mass that is traveling vertically inside the ship.

erobz said:
I think you need to consider the mass of the anchor as well? Also the final mass of the chain will be the initial hanging mass + the anchor+ the additional 30 m of chain at 5 m/s. You also have to change the kinetic energy of the mass that is traveling vertically inside the ship.
I'm assuming that the runaway begins after the anchor touches bottom. Also assume that as soon as chain speed exceeds ship speed, the chain becomes nearly vertical, and the length of chain in motion remains constant.. But yes, I should include the K.E. of the ascending chain also; that would need about 50% more power.

Sorry to hear you're not well. Get well soon.

erobz
anorlunda said:
I'm assuming that the runaway begins after the anchor touches bottom. Also assume that as soon as chain speed exceeds ship speed, the chain becomes nearly vertical, and the length of chain in motion remains constant.. But yes, I should include the K.E. of the ascending chain also; that would need about 50% more power.

Sorry to hear you're not well. Get well soon.
So the anchor is on the bottom, but the chain hangs close to 30 m before it touches water correct?

Here's how I would attack the problem of a anchoring system for a large ship:

1) Assume a worst case hanging mass. The anchor is just above the bottom of the deepest water in which the ship will be anchored.

2) Assume a maximum allowable speed of drop at that point. This value may be iterated.

3) Design a self contained passive (no electronics, operator adjustments, or electric water pumps) water brake to maintain that velocity at that load. Feel free to calculate the effect of water drag on the anchor and chain, but it's much easier to assume zero drag. That approach is most likely only slightly conservative because I think that water drag will be small compared to gravity force on the anchor and chain at any reasonable velocity.

4) Design a mechanical brake to stop that mass from that velocity in a reasonable distance. The mass and velocity are from above, the distance may be iterated. The brake should have enough braking torque to stop that mass at that speed, and enough thermal capacity to lower the full length of chain under mechanical brake control at the maximum depth at the maximum speed over ground. Design the brake operator for maximum of one or two turns from full off to full on. It should be operable by one person of average strength.

5) Design a hard stop on a spring and shock absorber arrangement to stop the chain in the case of a mechanical brake failure. The first design criteria is the free hanging mass of the chain at the maximum anchoring depth. The second design criteria is the worst case chain velocity, which may include the ship velocity over ground. The anchor will be on the bottom at this point. You may or may not want to consider the case of a powerless ship driven by wind and/or current and the anchor not slipping. The spring and shock absorber arrangement may activate if the brake is applied too hard, so it should be self resetting.

6) Design the system weak link. The existing weak link is the attachment of the chain to the ship. Remember that it's better to lose an anchor and chain than to pull the bow of the ship off.

This system should allow both fast drops and controlled drops, and allows for operator error on the brake. This system will allow manual control of drops with a failed water brake. It will allow anchoring with a mechanical brake that failed in the open (unbraked) position by letting the chain run all the way out, then winching back to the desired length.

erobz
hutchphd said:
Here's a nice very short compendium on magnetic brakes:
I considered that. These are incredible things - they can stop very, very, very heavy objects on a dime. However, I think in principle they suffer from the same problem as mechanical brakes - what do you do with the energy?

jrmichler said:
6) Design the system weak link. The existing weak link is the attachment of the chain to the ship. Remember that it's better to lose an anchor and chain than to pull the bow of the ship off.
A weakest chain link is supposed to be built into the chain near the end.
The chain can be manually emergency separated from the hull - ie weather deteriorates to the anchor drag situation

anorlunda said:
I'm assuming that the runaway begins after the anchor touches bottom. Also assume that as soon as chain speed exceeds ship speed, the chain becomes nearly vertical, and the length of chain in motion remains constant.. But yes, I should include the K.E. of the ascending chain also; that would need about 50% more power.
Ship speed should be zero when dropping anchor.

I considered that. These are incredible things - they can stop very, very, very heavy objects on a dime. However, I think in principle they suffer from the same problem as mechanical brakes - what do you do with the energy?
Clearly I have no "feel" for this size force. That being said three things mitigate for magnetic brake in this circumstance:
1. Water will not directly affect braking force by lubricating
2. Heat is delivered throughout disk...not just the surface
3. The force increases with speed (to saturation current I guess)
I believe the temperature dependence is not large (Curie temperatures may need consideration for permanent magnets). One worry is that all the new rare Earth magnets seem to oxidase very easilly and deeply. Salt air.
Surely there is an effective way to exract heat when an unlimited cold bath is meters away...directly submerge the entire brake somehow?

jrmichler said:
1) Assume a worst case hanging mass. The anchor is just above the bottom of the deepest water in which the ship will be anchored.

2) Assume a maximum allowable speed of drop at that point. This value may be iterated.
I like that. It's a rational way to rate the system based on a knowable worst case.

But what you called a "hard stop" and a "weak link" are contradictory, and the hard stop seems to fill the same purpose of the water brake.
jrmichler said:
This system should allow both fast drops and controlled drops, and allows for operator error on the brake. This system will allow manual control of drops with a failed water brake. It will allow anchoring with a mechanical brake that failed in the open (unbraked) position by letting the chain run all the way out, then winching back to the desired length.
Music to an engineer's ears, fault tolerance and contingencies.
256bits said:
A weakest chain link is supposed to be built into the chain near the end.
Yes. The last link is called "the bitter end". Often it is not attached to the ship at all. No attachment and a weak link are approximately the same thing.
hutchphd said:
Surely there is an effective way to exract heat when an unlimited cold bath is meters away...directly submerge the entire brake somehow?
Yes, a water bath for cooling should be doable. I thought about both water brakes and magnetic brakes. Both may be viable.

But when I read that the resistance of the water brake can be controlled by a valve on the water feed, that tipped the scale for me. Then I could control that valve proportional to speed, and adjust the proportionality constant to change the "steepness" of the brake force versus speed curve.

With magnetic brakes, how does one make the braking force adjustable, and the slope of force versus speed adjustable?

256bits said:
Ship speed should be zero when dropping anchor.
No. That is discussed in #17.---

Great discussion everyone, I think we are zeroing in on a consensus solution. Too bad nobody wants to hire the PF community to design ship systems.

hutchphd
anorlunda said:
With magnetic brakes, how does one make the braking force adjustable, and the slope of force versus speed adjustable?
My stationary bicycle (only as a last resort) has magnetic drag. I think the adjustment is to move the magnet away. In any system I think it involves lowering the effective magnetic strength (and probably force it is quadratic in B). There are many simple ways to do this...I do not see it as a problem. The slope may be a little trickier but maybe it is as simple as radius of the disk? More thought required.

hutchphd said:
1. Water will not directly affect braking force by lubricating
2. Heat is delivered throughout disk...not just the surface
1. No, but salt water will interfere with the EM fields.
2. Heat is actually delivered by (and energy to) a dump resistor. The problem is what happens if the dump resistor overloads or otherwise fails.

1. No, but salt water will interfere with the EM fields.
2. Heat is actually delivered by (and energy to) a dump resistor. The problem is what happens if the dump resistor overloads or otherwise fails.
I am not sure we are considering the same system. I am thinking about permanent magnets inducing eddy currents in a rotating disk. The braking is from the intrinsic conductivity (resistivity) of the disk (or drum)
My thought was to have a reversable electric motor drive to raise and lower the anchor. In addition there would be a passive permanent magnet brake. This has the happy characteristic of being (passively) more effective at high speed and probably would need no active components. Sort of like dropping a magnet through a conducting tube. (maybe that is the appropriated geometry but it is not obvious how to make that work at scale). Perhaps separate windlass with a cooled disk and magnets. One would need to look at the numbers of course.

hutchphd said:
My stationary bicycle (only as a last resort) has magnetic drag. I think the adjustment is to move the magnet away. In any system I think it involves lowering the effective magnetic strength (and probably force it is quadratic in B). There are many simple ways to do this...I do not see it as a problem. The slope may be a little trickier but maybe it is as simple as radius of the disk? More thought required.
If you use a flyball governor to move the magnets, then you have the adjustment of the slope and intercept of the force versus speed curve.

anorlunda said:
If you use a flyball governor to move the magnets, then you have the adjustment of the slope and intercept of the force versus speed curve.
For the bicycle I think the fixed magnetic drag supplies a simulacrum of the air drag, although I simply think of it as a midievel rack so who cares. However I still enjoy the real two wheeled variety on early summer mornings...(not dead yet!)

Last edited:
I too thought of more sophisticated solutions, but at the end what I would add is really just an (oversized) elevator-style speed limiter. The kind which not only doing speed limiting, but making a full stop till further intervention. (With optional usage as secondary brake - so it could be initiated manually too, not just by the overspeed.)

Seems pretty simple to solve. Just use a eddycurrent brake. It will brake the chain more the higher speed it gets. So it is self regulating.

Replies
9
Views
2K
Replies
22
Views
3K
Replies
14
Views
30K
Replies
98
Views
10K
Replies
4
Views
2K
Replies
26
Views
6K
Replies
19
Views
4K
Replies
1
Views
2K
Replies
8
Views
3K
Replies
10
Views
1K