My Anti-roll Bag Experiment

[Marco Flamingo]

I thought I would post this as a separate thread instead of occasionally commenting in other stabilization threads.

As with anchors, one’s choice of stabilization is based on the boat’s size, usage, location, etc., as well as the crew’s tolerance and the captain’s wallet. One of the things that made me think about putting in a stabilization tank is I may have the perfect place to install.

A prior owner had modified the seating on the flying bridge. Originally, the boat had the fairly common setup of a single helm box-seat starboard and a double seat port with a walkway between. Both boxes had access from the top. The modification had been to connect them (eliminating the walk-through) in order to install a long built-in cooler box in the port seat. It is kind of clunky to have to now climb over that area, but the PO must have thought that having a cold case of beer on the flying bridge was worth it. You can see the modified seat area (painted brown) in the pictures below. The top opening on the port side seat is the the ice box extending into the middle section. So I already have a 24” x 16” x 102” area for an anti-roll tank.

Experimenting with the bag came about because I was curious to see what, if anything, was the effect of having water moving around on the FB. I had read a few studies regarding “free surface” water tanks and it is generally the case that free surface water waves travel side-to-side faster than the roll period of the boat. The basic idea is, say when the boat rolls to port, the water in the tank will also roll to that side, but it needs to “arrive” a fraction of a second after the boat begins to recover. That extra weight, now having been transferred to the port side just when the port side is lifting, is the whole ballgame.

This is an old concept. Lord Kelvin presented a paper over a century ago on the effectiveness of a free-surface water tank on a vessel’s deck. At the time, a new sailing vessel was often “rolled” at anchor to determine various weaknesses and remedy them best as possible. 20 men would run synchronously back and forth on the deck to induce a roll and the roll period, angle, attenuation, etc., would be noted. Very scientific, right? Of course, the vessel had already been designed and built (where the issues could really be addressed), but some changes to ballast and rigging might be possible. The addition of a free surface water-filled tank on deck was surprisingly effective stabilization. (I have done the same testing at the dock with my anti-roll bag.)

Over the years, changes to the anti-roll tank’s placement, shape, internal baffles, etc. have been tried to increase the effectiveness. There are also tanks which are not “free surface” and have constrictions, pressurized air, water pumps, etc. to increase effectiveness. Of course, the more complex the system, usually called “active” systems, the more likely the system can fail. There is a famous “wreck dive” on a ship that had an active stabilization tank malfunction on its maiden voyage. A glitch caused the water in the tank to shift to the point that the ship simply rolled over and sank. This incident probably helped perpetuate the "hair on fire" reaction some have to the idea of shifting water for stabilization. A 700# dinghy, two kayaks, a freezer, and 4 obese Americans on the FB is okay. 500# of water? No way!

 

[Marco Flamingo]

My idea was to experiment with a strong bias towards the KISS principle. I’m not looking for a system that would need to respond effectively to 30-degree rolls or stabilization in 10-foot seas. I’m just looking to calm my boat’s “excessive stability,” caused in part by its large beam to length ratio (10.5’ x 29’ at the waterline). It has a deeper than average single keel (46”) but even still the boat is “snappy.”

Instead of tearing out the cooler in the seat box and building a tank, I decided to try an anti-roll “bag” first and maybe even transfer the bag inside the seat box. That might save me from having to modify the boxes to ensure that they are water-tight. By not having it “free surface,” I thought that I might be able to slow down the wave a little and see just what effect having a liquid counterbalance weight on the FB would do.

My FB is about 8.5 feet across (and 10.5 feet above the waterline). I bought 3 yards of vinyl coated polyester (54” wide) and heat welded it into a bag, adding a fill port. That produced a bag that was about 24” across and 9’ long. I wanted the extra length because I figured that in a big roll, a lot of water would go into the end of the bag and I didn’t want to constrict that motion. As you can see from the pictures below, it may constrict some of the water despite the extra room at the end.

Wave speed is affected by several variables. One of the variables is the pressure above the liquid surface. By not having a “free surface” and using the tension of the bag, I think the wave is slowed a little. We are talking about a fraction of a second. My boat’s roll period is about 3.4 seconds. Port side up to port side down is 1.7 seconds. If I can slow the water down to 1.8 seconds, mission accomplished. Filling the bag semi-tight (about 20 gallons) with no air in it seemed to do the trick. That makes the bag about 3 inches thick, so that amount of water would easily fit into the seat box.

Here are a couple of pictures of the bag in action. This was taken while crossing the Straight of Juan de Fuca. We were experiencing max 10-degree rolls in a beam sea (one 11 degree). I filled the tank using my potable water. When I rolled to starboard, the port side of the bag was almost empty. When rolling to port, the water would arrive just as the port started to lift. We no longer built synchronous rolls to 10 degrees. 6 was the max and the rolls subsided faster. A noticeably better ride. With no air in the bag, it could not be heard from the lower helm.

The anti-roll bag has some shortcomings. First, my heat-welded seams aren’t perfect. I have a slow leak that is really irritating. Were I to do it again, I would probably use HH66, a nasty adhesive that I have used in the past on vinyl coated polyester. Because my bag is such simple construction, I could cut off both ends and the lengthwise seam and redo everything with a slightly smaller bag. But I’m thinking my next experiment might be gutting my seat box and making a real anti-roll tank. Because I won’t have the bag effecting the wave speed, I’ll have to come up with a way to slow down the motion of the water. That’s going to be my next experiment.

I’ve read quite a few studies on how to best slow down the wave speed in a tank. Baffles and complex plumbing seem to be common. One study had diagrams of 6 different shaped tanks with variable baffles. It seems to me that one of the problems is simply the liquid that is always used. Water (sometimes seawater). If one used maple syrup, the wave would be slower (but there would be other issues). In looking at what liquid could have the right viscosity, I came across “thick water.” It is water to which food-grade calcium chloride is added, making it goopy to the point of way to viscous.

But I think there may be a water to calcium chloride ratio that would slow the wave down without baffles, etc. As a plus, calcium chloride doesn’t increase the liquid volume, meaning that adding 2 pounds of calcium chloride to a gallon of water (8.34 pounds) results in the gallon of solution weighing 10.34 pounds. Thus, slowing the motion but increasing the liquid’s effectiveness. We shall see.

 

[Kit_L]

Marco,

That's a very clever 'I can do this at home' experiment. I have seen an anti-roll tank that had baffles, but only to half the depth of the intended water. The idea was that the baffles would slow the water movement sufficiently to partially cancel the roll. Unfortunately I did not get to hear of the results, but the idea looked interesting.

Your bag is some ways is better, I feel. And the calcium chloride idea is excellent, too. I think increasing the viscosity to only double what water is normally will be all that's needed to be able to see if this makes a difference. I am sure it will. I am following this with interest.

 

[Delta Riverat]

I'll be watching this one.

My boat has a large "under flybridge" compartment with an access door like yours. I could easily get a long tubular bag in there and completely out of the way. Not only that but there is a cold water sink and a drain already plumbed up there that could easily be accessed for fill, add and dump.

Possibilities...

 

[Barking Sands]

Man, thanks for the history and physics lesson. That was a fun read. Great idea on the CC. Sure seems that is should work and be able to be "tuned" to best performance. Great looking boat BTW!

 

[Benthic2]

This might not be a good permanent solution, but putting a rope around the middle of the tank would allow you an adjustable restriction ( like tightening a belt too tight ).

A more permanent option would be to increase the effectiveness by adding water ( or viscosity ) so that the effect is more than you want. Then run a hose from one end of the bag to the other with a valve in the middle. Open the valve and air will move from one end to the other, allowing for faster water movement. Close the valve and slow the movement.

 

[Marco Flamingo]

I also have a lot of storage space under the FB forward cowling. Unlike the area behind the seat boxes, the storage space extends an additional 5 or 6 inches on both sides, thus making a tank there a foot longer. This could increase the effectiveness both because of the water taking an additional fraction of a second and the increased leverage. The possibility of a curved tank under there is also interesting.

But there were downsides. My bag, which leaks, would have been a problem. Even if no longer used for storage (mainly seat cushions), I have all my upper helm electrical in there and access is already really tight (I figure I will only be limber enough to climb in there for a few more years). And that forward area is actually cantilevered out over the lower windows. I don't have much concern with #300 pounds spread out on the upper deck, but on the cowling overhang? Also, the bag really gets a workout. I would think that after a few years there might be wear issues.

Benthic: Your idea of valving either the water or the air above a tank has been used successfully. It is one of the simplest control systems in an "active" tank, so I looked into it. One of the benefits is that the bulk of the water can be withheld until just the right time. The problem has generally been how to control when just the right time is. It can be calculated, but then things tend to get complex. The common thought today is "computers to the rescue". That's basically what lead to the sinking of the MV Zenobia.

Even the simple baffle systems that I have seen appear to have a common shortcoming. They slow the liquid down at the expense of momentum. The end baffles shown on the MV Swan Song (second picture) would appear to slow a small fraction of the water down as it drained from one end and slow some fraction again as it poured into the other end, but possibly losing some of the effectiveness of a single glug (wave) arriving at just the right moment (a Goldilocks tank). Don't really know, but if baffles, valves, pumps, electronics, etc. can be eliminated, I'm there, even at the expense of increased stability.

As with all stabilization systems, "effectiveness" can be somewhat vague. The anti-roll tank studies actually have some of the most numbers and graphs of any of the stabilization research I've read (until my eyes glaze over). Most of the studies contain mathematical equations that would confound Einstein. A few end with graphs actually showing roll and period attenuation. Just the facts, Ma'am.

One of the take-aways is that most passive anti-roll tanks tend to have a roll angle (or, oddly, two or three) at which they are most effective. Other systems might also have this, but the Anti anti-roll tank folks have made a lot of this. The system might be super effective stifling a 9 degree roll, not so much with 12 degree (until it decays to 9), but then again at 3 degree and less. The stability graph is generally complex and a smooth graph line seems to be preferrable to some. The other problem some dwell on is that there will certainly be a roll angle at which point the liquid in the anti-roll tank can adversely effect stability. A >40 degree roll maybe? I've never done that in a powerboat and only one memorable time in a sailboat. I don't plan on getting into a situation where that is repeated. And if I do, it is the fault of my trip planning, not my stabilization system.

 

[PhilPB]

Great information, now you have me thinking!

 

[ksanders]

You reduced your roll by 40% by adding your anti roll bag.

I for one am impressed!

Going from a 10 to a 6 degree roll is huge!!!!

 

[Marco Flamingo]

Just came across this as an off-the-shelf anti-roll bag (the 12" x 10' bag). It is longer than what I need and I would be concerned about how much knocking around it can take, especially if placed under the cowling where I have some fairly rough edges. It could only be filled about 30% to allow for all the liquid to get into one end during a big roll. That's still about 20 gallons, about what I have tried. Also, as I found out, even my "flat" bag travels fore and aft. I'm not sure how one would control a round bag. You can see from my avatar approximately how far back my mast is from the FB seating. I found that the bag had migrated back around the mast several times in semi-rough conditions.

 

[Mako]

@Marco, what a great thread, thanks for authoring it.

I was going to hire Dr Don Bass to design my flume tank for a new build several years ago (the project was cancelled though), and will certainly like to use one in my next boat.

 

[Delta Riverat]

I noticed they offer "couplers" for connecting the tubes. These could offer much improved abrasion protection by wrapping the tube full length.

 

[Marco Flamingo]

I noticed they offer "couplers" for connecting the tubes. These could offer much improved abrasion protection by wrapping the tube full length.

And one could glue vinyl-coated polyester patches with a D-rings to the coupler in order to control the bag by tying it off. I would still be concerned about anything that still allowed the bag to bend/twist and effect the wave. The best way to control the bag would be a purpose built "bed." Of course that would mean I would be 80% on my way to building a tank (a bed with a lid). Gutting my seat boxes is still looking like the way to go.

If my seat boxes weren't already connected, connecting them with a regular step would probably work. No need for the full height between the boxes.
An under-seat tank would likely have about 4 inches of water in it. That wave would pass back and forth under an 8" step between the seats with only the original seat boxes needing more volume (height) for handling big rolls.

 

[Benthic2]

If my math is correct a 4 inch diameter PVC pipe that is 10 feet long would have an interior volume of 10 cubic feet, or 75 gallons. If 20 gallons of movable water gets your the desired damping, the PVC pipe would only be about a quarter full, so it would have plenty of free space for movement. Its cheap and durable, easy to work with and easy to make waterproof. Plus I think a rigid tank would transfer more energy from the water to the boat than a soft sided container, but that is total speculation. There are PVC pipe adapters so you could put 1 or 2 sections of PVC with a 2 or 3 inch diameter as a baffle.

As for liquid have you considered diesel fuel ? Slightly thicker than water and plentiful and if your project goes bust you can always use 25 gallons of diesel. Dish soap or shampoo might be even better, but that might get costly. Used motor oil would be cheap.

Also, If you do another trial run, instead of comparing your roll period with and without the water bag deployed, it might also be helpful to compare with the water bag deployed to with an equivalent static weight on the bridge. Putting 80 pounds 10 feet above your center of gravity should have some dampening effect even if it isn't moving. Whatever you are comparing you want to be in the same place when you measure. If you were on the bridge for the stabilized test and in the wheelhouse for the unstabilized test your body weight could throw off your results.

 

[HeadedToTexas]

I work with PVC pipe a lot and that came to mind for me too. Being rigid means the antiroll tube could be mounted in the underside of the flybridge cowl thus keeping it off the floor of that area and allowing wiring and other components to be routed around it.

A 40% reduction in roll from a section of 4" Sch. 40 PVC pipe is a lot of magic. Fascinating.

 

[Delta Riverat]

How about ABS? Much less expensive and easier to work too.

 

[ksanders]

If my math is correct a 4 inch diameter PVC pipe that is 10 feet long would have an interior volume of 10 cubic feet, or 75 gallons. If 20 gallons of movable water gets your the desired damping, the PVC pipe would only be about a quarter full, so it would have plenty of free space for movement. Its cheap and durable, easy to work with and easy to make waterproof. Plus I think a rigid tank would transfer more energy from the water to the boat than a soft sided container, but that is total speculation. There are PVC pipe adapters so you could put 1 or 2 sections of PVC with a 2 or 3 inch diameter as a baffle.

As for liquid have you considered diesel fuel ? Slightly thicker than water and plentiful and if your project goes bust you can always use 25 gallons of diesel. Dish soap or shampoo might be even better, but that might get costly. Used motor oil would be cheap.

Also, If you do another trial run, instead of comparing your roll period with and without the water bag deployed, it might also be helpful to compare with the water bag deployed to with an equivalent static weight on the bridge. Putting 80 pounds 10 feet above your center of gravity should have some dampening effect even if it isn't moving. Whatever you are comparing you want to be in the same place when you measure. If you were on the bridge for the stabilized test and in the wheelhouse for the unstabilized test your body weight could throw off your results.

I love the idea of pipe but...
a 4" pipe has a volume of 12.56 cubic inched per inch in length. so 10' would be 1507 cubic inches and a gallon of water is 231 cubic inches.

Change the pipe size to 12" and it's 58 gallons of water when full.

 

[Delta Riverat]

Now I'm thinking about a couple of plastic waste tanks connected with a 4" ABS pipe with a gate valve in the middle to "control flow"

 

[Benthic2]

I love the idea of pipe but...
a 4" pipe has a volume of 12.56 cubic inched per inch in length. so 10' would be 1507 cubic inches and a gallon of water is 231 cubic inches.

Change the pipe size to 12" and it's 58 gallons of water when full.

Thank you. I totally blew the math.....each 4" x 10' pipe only holds aprox 6.5 gallons. That's what I get for doing math in my head on a weekend!!"

 

[Benthic2]

Eureka !!! I forgot to label the black center piece as flexible rubber with hose clamps on each end to create a partial hinge.

 

[ksanders]

Now I'm thinking about a couple of plastic waste tanks connected with a 4" ABS pipe with a gate valve in the middle to "control flow"

Good idea...

The question for the day would be, is a 4" pipe sufficient to move enough water quick enough

 

[Benthic2]

Good idea...

The question for the day would be, is a 4" pipe sufficient to move enough water quick enough

I thought the idea was to slow the water down so it arrives after the roll has switched directions ?

 

[Delta Riverat]

Ideally the system would be "tuned" to the period of the boat.

But the "open tank" concept is pretty fast, as fast as it gets.

Only way is to experiment, cost is low but there is the risk of flooding the fly bridge.

 

[ksanders]

I thought the idea was to slow the water down so it arrives after the roll has switched directions ?

I think you are correct and I like the idea of two tanks connected together.

My only off ther cuff thought is wether a 4" pipe is enough volume to allow the water to move fast enough.

The idea, at least to me is to get a huge volume of water to suddenly exert a force just as the boat has changed directions in the roll.

My thoughts are that it's not just the weight of the water, it is the velocity of the water as it hits the end of the tank that produces the energy needed to counteract the energy of the roll

Is this correct?????

 

[psneeld]

Seems some rolling tanks are designed in a U shape that helps with the "tuned" flow of water in adition to any restrictions or flow restrictions.

 

[SteveK]

The devils advocate in me has to ask.
Are boats built to withstand lateral weight shift continuing over time. Is no one concerned about the weakening of the framework and a sudden collapse of the house?

 

[ksanders]

The devils advocate in me has to ask.
Are boats built to withstand lateral weight shift continuing over time. Is no one concerned about the weakening of the framework and a sudden collapse of the house?

Good question!!!

Using the height of the flybridge as a lever for reducing roll imparts forces that the designer probably did not plan for.

In our boats with all with all of the windows that is a very reasonable concern

 

[psneeld]

The devils advocate in me has to ask.
Are boats built to withstand lateral weight shift continuing over time. Is no one concerned about the weakening of the framework and a sudden collapse of the house?

Probably not.....

 

[Marco Flamingo]

Also, If you do another trial run, instead of comparing your roll period with and without the water bag deployed, it might also be helpful to compare with the water bag deployed to with an equivalent static weight on the bridge. Putting 80 pounds 10 feet above your center of gravity should have some dampening effect even if it isn't moving. Whatever you are comparing you want to be in the same place when you measure. If you were on the bridge for the stabilized test and in the wheelhouse for the unstabilized test your body weight could throw off your results.

True. I'll bet no captain has ever thought: "It's getting rough. I'll go up on to the flying bridge so that my 200# dead weight will slow the roll period and make my passengers more comfortable." Or better yet: "Here comes a big wake! Quick, everybody up on the FB!" I'm sure the weight aloft would have an effect. I'm not sure whether enough to be measurable. That's why I wish my phone app inclinometer had a time/graph function. I would guess that roll period might be increased but maybe amplitude as well.

For me, the "weight aloft problem" regarding the strength of the deck is a non-issue. Yes, I had 200# of water going back and forth on my deck. I'm not quite 200# but it had never occurred to me that if I got down and rolled around on my deck it might collapse. Or that I should only have one person at a time on the FB. Or I should be afraid to stand on one foot because of the point load. My PO had the FB deck rebuilt and I have confidence that 3 or 400 pounds over 4 sf. would not be an issue. Were I dealing with 1,500# of water and a 300# tank, as was used on the Swan Song, I'd do some structural analysis.

The use of commercially available plastic pipe had occurred to me. Specifically, I looked at the 12" and 16" black corrugated pipe commonly used as a culvert for a driveway. It is fairly flexible and might even be jammed into a space. Elbows aren't available, but end caps could be made water-tight without too much work. The corrugation might have the effect of slowing the water. (One study I read had the tank lined with fibers that reminded me of the 3" slime that can grow on my hull. Turns out that, while it slows my boat, it didn't have sufficient effect in an anti-roll tank).

But the use of a pipe has some drawbacks. Because of it's shape, and the fact that it needs to be about 1/3 full in order to shift all the water end-to-end, one would get very little volume with a 12" pipe. Without doing the math, I would guess that 4-5" of water in a 12" pipe is less than half the amount of 4" of water in a flat-bottomed tank, even though a pipe essentially takes up as much space as a tank. It might work okay on small rolls, where the surface of the water ripples back and forth, but once you need the shifting of more weight for bigger rolls, the shape of the pipe leaves you with dribbles. The use of round bag doesn't have quite as much problem as it stays semi-flat when partially filled.

The 4" pipe between the two tanks as drawn above has the problem of both not enough flow and not allowing the free surface effect. Variations of it have been used on "active" tanks where a time-controlled reversing pump system is used (with it's associated costs and complications).

There are still some things that I haven't really seen in the studies. Baffle systems are generally shown in plan view, but I don't remember seeing any experiments that really show side views. Like convex/concave tank bottoms, subsurface floats (a kelp bed), flexible baffles attached to the bottom. Maybe a turbine, the mass of which has to be spun and reversed. Some of these might be called "semi-active" systems. Maybe rare earth magnets that momentarily held a baffle in place until overcome. Or a weighted pendulum that controlled the timing of a floodgate/throttle in the tank. Timing the release of water based on gravity rather than pumps, valves, and computer chips. Gravity never fails.

 

[HeadedToTexas]

A 40% reduction in roll from a section of 4" Sch. 40 PVC pipe is a lot of magic. Fascinating.

I love the idea of pipe but...
a 4" pipe has a volume of 12.56 cubic inched per inch in length. so 10' would be 1507 cubic inches and a gallon of water is 231 cubic inches.

Change the pipe size to 12" and it's 58 gallons of water when full.

Ah yes, there is the magic! Sufficient force generated by a 12" pipe almost full of water makes much more sense.

How about ABS? Much less expensive and easier to work too.

Good point. Light and cheap are often good. I work almost exclusively in PVC so that's what occurred to me.