On the ROV hydrodynamics. Improving shape


Hi all:

As you may know, I've been worried about ROV hydrodynamics from my arrival to this site.From that beggining, as well, I'm impressed by the very good, nice and smart job that is being done around the project. Hence, please don't take what follows as an attack, but as a contribution from my side.

Always from the "Underwater Vehicle" and not from the Robotic point of view ...........

When a body moves through a fluid, it changes its state from steady to distorted. The more difference between both states, the more energy is obviously transferred from the body to the fluid.

In the case of a submersible of anykind, the best performance is wanted. Meaning that the less energy must be tranferred from the propulsive + hull system, to the fluid.

Have in mind, that all that wasted energy, in the form of turbulence, heat and noise, also comes from the batteries, shortening their life and the "sub" range.

First thing that jumps to the eye when having a look to the ROV, is its asymmetry with respect to an horizontal plane.

Even at the low speed of 1.08 m/s turbulences have a big effect.

Pressures are not evenly distributed. A body moving under such conditions will dive or go up. In the case of the ROV, it dives. The faster ahead, the bigger that diving condition.

We know it dives from two sources: Reports from users, and computation of pressures distributions.

The stagnation point of the ROV when going ahead, is located on a horizontal line on the E_Cylinder surface.

From this point, flow is broken into two main streams. One runs over the "smooth" upper chassis cover, while the other, has to turn around the cylinder, for finnally arriving to the complex internal ROV framing.

That structure is so complex from the fluid point of view, that stream twists again and again, leading to a low pressure area which "sucks" the ROV down.

Battery cylinders, which will be discussed later, help for this effect due to their plannar ends.

Now, we have two streams: A "clean" one running over the body and a "dirty" one which runs below and inside it.

The upper one, almost entirely passes by, not entering the propeller ducts. While the "dirty" one, is the only one feeding the propellers.

Propulsion has now to do three works.

1-) To attrack fluid that is not leading towards its rotation plane.

2-) To tidy flow up.

3-) To drive flow astern, allowing for propulsion.

Only number three is useful. At the same time, a big ammount of energy is extracted from propulsion for eddies formation inside the framing.

The pic below, froma CFD study, shows what is said above.

Note the blue area under the cylinder (low pressure-Coanda effect) and the heavy eddies area inside the ROV.

Well. Once problems stated, solutions can be found.

The question is: Holding main dimensions, price and simplicity, try to improve the "sub" performance.

Two main things have to be taken into account. Fluids don't "like" to be distorted.

Symmetry, where possible is a MUST.

Hence a new internal structure and a partially changed external chassis has been designed. It has been done by trying to approach the ideal shape, but being restricted by construction materials and building tools. It means, no curved surfaces and no complex parts.

Holding the size and the E_Cylinder with all its parts, allowed for a flow study and hence for finding the shape that better fits the fluid paths around the ROV.

At the same time, flow has been symmetrically directed to the propeller ducts in the more possible tidy state.

With both conditions, manouverability and propulsive performance would have to be improved.

Many attempts have been done (Surely you know that fluid dynamics is not an "exact" science), and after what seems to be a reasonable time spent, with a reasonable result, an improved ROV hull has been obtained.

The pic below, shows the new flow path.

Both convergent planes above an under the central girder are only two sheets from the same material alredy used for the ROV which have been sanded at the E_Cylinder edge.

Volume between both planes is opened at its after end. Its not an enclosed volume.

The tail section, from the same material, has also been sanded, for turbulence avoidance. The vertical plane shape is designed that way for decreasing surface dragg.

The pseudo-conical part ahead of the motors come from a model airplane catalogue. Are the parts used as propeller cones.

Similar cones have been fitted ahead and astern the battery cylinders. The dragg reduction has been quite big.

Both sides of the ROV are a lot opned in order to allow water flow to the props while decreasing the body surface as well.

All curves are elliptical or parabolloidal shaped.

A few drawings of this "hull":

Comparative results:


Force(Y) Vertical dragg. Note it has been reduced almost ten times.

Force(Z) Horizontal dragg. Reduced in a 50%

Propeller flow rate: Water propelled though the ducts. Also improved.

Hope to have the building plans in two days.

Moderators ..... Where is the place for sharing those plans ?


Regards and ........... Happy 2014.



This is truly remarkable work- I can't describe how thankful I am that you're part of this community!

In short- I agree that the current design is nowhere near hydrodynamically-optimized, and that a more vertically-symmetrical laminar flow into the horizontal trusters would greatly improve performance at higher speeds.

My design considerations so far have not optimized hydrodynamics because I've assumed most operations would be at slow speeds, and I've instead optimized for strength and simplicity.

That being said, looking through your post made me realize that it wouldn't be very hard to construct an experimental shell with slits on the top (on either side of the vertical thruster hole) and on the side (just aft of the tabs that hold in the internal structure) in order to empirically measure differences in performance. I suspect that even the simple act of allowing pressure to equalize better behind the main bulkhead will greatly reduce parasitic drag there.

I have to admit that the internal structure for the new 2.6 revision actually positions the motors even further forward (closer to the bulkhead) which I expect lowers high-speed performance slightly compared to 2.5. I once again made this decision because I wanted to improve durability, simplicity, and ease of construction, and I assumed that slow probing movements with the ROV would be more typical then long linear motions.

With a simple modification like slits in the shell though, perhaps we can have our cake and eat some of it too!

I'm looking forward to seeing your building plans. If the design can be built from laser cut parts, I'll cut out a prototype of whatever you send!



Thanks a lot Erik.

Im proud of contributing to the OpenRov project. May be Im not very usefull in the robotics and electronics part. But ...... hydrodynamics and hull design is my daily work. Software development is a collateral activity, as many times there is no comercial one developed for our needs. In any of those fields, don't hersitate counting on me as much as you need.

Ok, Im on the work for the construction plans.

About the "lids": Opening lids at both sides, will increase induced dragg and does not grant for a better water supply to the propellers.

As an example, ballast lids at submarine sides have always been one of the problems regarding their design, due to that dragg increase, and derived noise generation. The better solution always goes for smoothing forms.

Im with you. Strenght and durability are the first conditions to comply in this project.

Moving motors closer to the center, of course will improve manouverability (The lesser the vertical axis Inertia, the better) But, if propellers are moved as well, hydrodynamic particulars will result seriously damaged.

Would it be possible to supplement the tail shaft for holding propellers at possition while moving motors to the center ?

It coud improve ROV's performance.

As a conclusion, my worries around hydrodynamics (appart from, as we say, "professional obsession") are about battery consumption and ROV range, more than for speed.

OK. Once again, thanks for your words, and happy for being useful for this project.



We are happy to produce a aluminium prototype if this is interesting?



Ion thank you for this post, this is truly awesome! I just finished a project at work and I think I will have some spare time to try and start messing around with this. I have no formal knowledge of hydrodynamics but rather just an intuitive sense of what works and doesn’t from experience. That’s why I made the tail for mine which works pretty well for now but could use some improvements.

I like the idea of moving the props back but moving the motors forward, maybe by extending the prop shaft?

I will post anything fun that I come up with.




You always read that the teardrop is the most hydro-dynamically ideal vessel. It’s good to see the data. Thank you very much!


Pretty much all commercial ROV’s have a fairly boxed shape, and all ROV’s I have worked with have vectorized propulsion systems, maybe something to consider? I can send some pictures tomorrow of the one I have standing on deck right now.


This is a great idea. My motors are on the outside of the housing, so I am not as worried about that.

I drew up a battery end cap in Inventor and printed it tonight on a MakerBot 2X printer.

It fits very tight. May try to make it smaller in dia. before printing more.


919-batterytubeendv1drawing.pdf (80.1 KB) 920-batterytubeendv1makerbot.jpg (181 KB) 921-batterytubeendv1b.jpg (146 KB)


Hi all:

Im still drawing the building plans, hope to finish today.

My goal is conserving Erik's design philosophy, simplicity, low price, good performance, durability ..... not an easy task.

Thaks for all your replies. Is the first time I received so many of them. May be my Christmas gift ¡¡


An aluminum ROV of what kind ? Do you mean the same ROV but changing construction material ?

The whole project is based on the idea of producing an easy to home-build ROV. Aluminum welding is quite a complex and expensive technology for most home builders.

At the same time, electricity and Aluminum-copper mixing in a wet environment, is not an easy thing to work with. I've worked on some Aluminum ships building projects (A Ferry-1100 passengers, and a Patrol ship 35m LOA) you'd have to see the large ammount of problems regarding bi-metal pairs and electrical isolation togheder with structural ones.

Anyway, the idea can be developed. If things are well done, we'd get a stronger ROV.

Could you please explain me your idea a little bit more ?


Yes, the very famous tear drop. :-)

Its a curious fact, that subs Hydrodynamics were not really studied until the end of the WW-II. Those sutudies, based on the XXI german U-Boat class, arrived to that "Tear Drop" concept, that is now the general shape of all serious submarines (Have a look to the USA Albacora project).

For our case, the ideal proportions make no sence. The ROV must be kept simple, in design and construction, and service speed does not justify for increasing complexity too much.

But, that ideal form can be approached, and that's what I've tried.

For your information: Two main dragg figures must be taken into account for a submerged body. Shape (form dragg), and wetted surface dragg. The more slender the body, the better, but as lenght increases, wetted surface grows as well. At the same time, pressure distributions must be taken into account. At the bow end, pressures abruptely grow at the stagnation point (Relative flow speed=0), moving around the hull for finally meet at the after end. The pressures distribution development along that path, must achieve ambient pressure at that point. This transition must be achieved in the smoothest possible way.

After a long study and many sea trials, it was found that best shape is given by a revolution body, which profile is given by an elliptical section ahead, extending to the maximum beam, and a parabollic section, from that maximum beam to the after end. Best practical shape (propulsion arrangement volume requiremens astern) was found to be achieved by means of two ellipses, better than the parabolla:

Ellipses parameters:


LOA= Lenght over all.

A and AA = minor ellipses semiaxis

A= AA = 1/2 Maximum beam.

B and BB = mayor ellipses semiaxis.



LOA / 2A =7

BB / B = 1.75

Hope this information can be of your interest.


Of course Im sure we all are interested on having a look to those pictures.

Yes, most commercial ROV are box shaped (Once again, simplicity, endurance and low speeds). But, knowing that OpenROV is mostly a research project, may be we could test different solutions.

Are you from Norway Roy ?

Jhon Stager.

Nice work ¡¡ Im sorry to write a but, but ...... An elliptical end would work a lot better than the spherical one.

Try plotting this shape. Excel is a good tool for that.

If batteries external diameter is 32mm:

R=16mm(minor semiaxis)

Ms= 41.5 (Mayor semiaxis)

If R is different than 16, Ms will keep the same proportion. Hence just change both values proportionally.

For values from 0 to Ms, calling X to the elements of the range column.

Plot this function Y= 16 x SQRT(1- X²/Ms²)

The resulting (X,Y) series, will give you a very good practical hydrodynamic shape.

Kind regards.


Great work. One question about the proposed hull design. What parts have changed exactly?
ex. there are now upper and lower plates with an oval cut-out for the vert thruster.
ex. the motors appear to be moved up a little to make them inline with the electronics tube
Q. Are the propeller enclosures still the same or are they wider with tapering?
Q. What are the dimensions of the upper and lower plates?


I’m thinking about how to apply your findings to those of us who already have the ROV. Slight modifications to our hulls won’t get to efficiencies of your new design, however they may improve our down time and diving issues.

Would it help to cut out a rectangular vent in the upper plate? How much does just adding cones on the battery pack help?
Is there “low budget” software available so that we can do similar tests?


Hi friends:

Im almost finished with the building plans. Just solving some matters regarding how and where to fit wiring.

CFD software:

With no theoretical backgrownd, this kind of software is useless. Any way, If you have that knowledge, or you are for studying, OpenFoam, for Linux, is quite a good free CFD software. It will perform the fluid study, another soft is required for data visualization. (There are a lot in the net).

If you are used to Linux, try "CAELINUX" compilation. It's specially directed to engineering and FEA (Finite Elements Analisys) and includes all the required software.

A simple but good 2D analisys tool for starting and getting the main concepts may be "JavaFoil". Its designed for Aerodynamic profiles, but as far as Reynolds and Mach numbers can be adjusted, the analisys results are useful. Its not a FEA but a Panels software. I mean, its not as accurate and complete as others, but works fine.

Other free programs that can help for starting and understanding fluid dynamics are Xfoil and XFR5 (a 3D version of Xfoil).

Your questions about dimensions and parts can not be answered until plans are finished.

About making modifications to the existing ROV.

Modifications to anything that already works quite well, from no other information than a theoretical study, is not a good thing.

Let's wait and check "seatrials" results before taking this new model as better than the former one.

Anyway, I'll post plans here for anybody who could want to test it.

About prop ducts, Im testing some profiles, always having in mind that they must be built, in most times, by only a sand paper and patience.


Once again .... the smother, the better. Cutting holes would probably make the ROV work worse at the same time that it would be weaker. For being sure about the Hydrodynamic performance with that hole, a study is required.

About the improvement that would suppose only adding cones to the battery tubes, ... I'll make some numbers. I'll answer later.


Hi quantumquark:

For your information:

Flow speed= 1.08 m/s

Along longitudinal axis

Battery tubes particulars:

Tube OD= 32mm

Sectional area: 0.000804 m²


L= 300mm

Reynolds number(Re)= 280398

Dragg= 0.449 N (Each tube)

Cd(Dragg coefficient)= 0.96


L= 356mm

Re= 332739

Dragg= 0.089 N each tube.

Cd= 0.19

Of course interactions with other parts will distort flow and change those figures. But provided the possition where this tubes are fitted (working as appendages and not as a part of the hull), this figures can be taken as correct.


Cd is reduced by a factor of five.

Total dragg goes from 0.9 N to 0.18 N.

Quite a good improvement.

By the way, removing almost 1 Newton of drag from below, will decrease the tendency for diving when full ahead.

Example cone at:




Wow! Thank you very much.


I tried the battery cone a 2nd time with close to the dimensions given.

Used the MakerBot 2X again. It took a little over 2 hours to print the new design. It has 4 shells for thickness on the outside and 15% in fill. It will probably take on water. On this version I have a cutout for the wires. It still fits nice and tight on my 2.5 version battery tubes.


916-batterytubeconev2p1.jpg (166 KB) 917-batterytubeconev2p2.jpg (171 KB) 918-batterytubeendv2.pdf (88.5 KB)


Quite a nice job ¡¡¡

I think that the cone and battery tube external diameters are the same. Right ?

Water filling the inside of the cones has no effect once filled. Only a little "added mass", but as the hole is quite small, once filled there will be no flow.

Its something similar to what happens to the lateral propeller bow transversal tunnel in a real ship. Once water filled, and over a given speed, there are no flow conditions inside the tunnel, that behaves almost as being solid.

Now, togheder with the dragg reduction, the diving tendency has to be a lot smaller.

First effect is hard to be measured without a towing tank, but the second would have to be obvious. Would you please confirm us that it really behaves as calculations predict ?

It would be quite useful for the whole new design validation.

I posted a comercial cone becouse not everybody has a 3D printer. That cone approaches the "ideal" one, but yours is a lot better.

Kind regards and congratulations ¡¡


Hi Erik:

Im in the final stage of plans drawing.

Could you please tell me what 2D file format do the lasercutter accept ?

Or .... What would be the easiest format for sharing ?

Best regards


Dear friends:
I must apologize for my delay posting the construction plans for this “hull-chassis” design. I have had to travel to a yard for a survey. Just back home and back to the work on those plans and assambling instructions.
Final shape is going to be slightly different from the drawing above, due to wiring, besides skin friction reduction concerns (Providing fair wiring paths and total surface area reduction).
Regarding maximum depth, I think it could easily be improved by fitting two internal rings, made from two 6 mm thick-10mm wide rings, obtained from an acrylic tube.
It would also help fitting a hose-clip at each end of the E_Tube with a rubber band between the clip and the tube for protection and correct coupling.
With only those clips, watertightness would be highly improved. By the way, they can be found made from titanium (lighter and a lot durable).


Download and check the pdf catalogue (REF-0300960-9)

Once again, my apologizes for the delay.

Best regards :slight_smile:


Hi Ion,

I have been following this thread closely, was wondering if you have an update on your construction plans. Looking forward to testing them.

Kind Regards,



Hi Peter:

Thanks for your interest.

Im working on the building plans and parts cutting templates. I was intended to complete the work in two days, but ....... I've had no time.

As soon as I get it finished, I'll post the plans here.

A few days ago, I asked about what the better plans file format would be for sharing. With no further instructions, I'll post in .DWG AutoCad format.

Regards and thanks again :-)