Let's Make Robots!

Autonomous Underwater Vehicle

Path Movement Underwater, Autonomously maintains a stable position in the water.
AttachmentSize
IMAG0139.jpg1.36 MB
IMAG0143.jpg1.2 MB
IMAG0135.jpg1.29 MB
IMAG0142.jpg1.11 MB

The aim of the project was to design and build an underwater vehicle with an emphasis on stability such that it may possible to perform light well intervention. This evolved into a project with a (relatively) streamlined shape and a control system capable of maintaining a stable position, such that, upon rotating due to external force, the craft is capable of actuating back to it's original orientation. Some of the initial concepts can be seen here. CFD anaylsis was carried out to determine the drag coefficients of the various concepts.

Initial Concepts for the Craft

The concept on the bottom left was chosen for further development. The craft's frame is made from 12mm mild steel bar, model shown below.

The skin of the craft is 2mm mild steel sheeting. The final concept can be seen below.

The sensor array of the project is sparkfun's 9DOF razor IMU, which has a 3DOF accelerometer, 3DOF gyroscope and a 3DOF magnetometer. This feeds into an arduino mega 2560, which has control of six 12v DC motors which were cannibalised from a Rule 750GPH Bilge Pump. They are water proof, and have 50mm model boat propellers attached to them. They draw about 5.5~6.5A. They are controlled via. six motor controllers. These are custom made high current H bridges. They're currently a work in progress. The craft also has a data logger capable of logging any changes in orientation in pitch, yaw and roll. The basic idea of the electronics can be seen below.

The project is entering the final construction stage and hopefully will be finished in the next fortnight. 

H-Bridge

H-Bridge

 

Profile

Current State, Profile

Current State, showing external mounts. 

Update 04/04/11:

 A flotation test was carried out in a water tank. As you can see the craft floated with roughly 55% submerged. 0.5, 1 and 2 kg bags of sand were made up and used to appropriately weight the craft, to make the AUV as close to neutrally buoyant as possible. 

Flotation Test

Notice the M8 lifting eyes used to aid lifting the AUV in and out of the water. Upon submerging it was discovered water was entering the inside of the craft. The threads of all the bolts were vaselined, and rubber washers were used to guarantee a seal. The holes where the motor wires entered the craft were epoxied again, and siliconed. Subsequent testing confirmed the craft was water tight. 

Motor Driver

 

Shown above is the motor driver boards and weeks of painstaking effort. At first it was hoped to use power mosfets to construct an h-bridge. However after several melted mosfets, on-fire transistors and shorted PCBs, the decisions was taken to opt for mechanical relays instead. Two SPDT relays were arranged in an H bridge configuration, shown below:

H-Bridge Schematic

Six motor controllers were built, on three PCB boards. The only deviation was using relays rated at 12v/10A instead of 5v. Each board is powered by one 12v Lead acid battery. Tracks on the PCB are sitting at just over 3mm. 

PCB schematic

Note mistakes on this board. Three of the relays do not recieve power from the power rail, this was correcting by soldering wires on. Also some of the tracks (where the motors will be drawing their current) are far too thin and were made thicker before manufacture. Close up of boards shown below:

mdclose

The final electronics tower can be seen below:

electronic tower

Show is the small fuse box, obtained from ebay. It used blade fuses. The size of fuse is yet to be determined. The power wires (blue/brown) can be seen, as can be motor wires (red). At the rear of the picture is the arduino control wires. On top is a piece of copper strip board. The IMU / Datalogger and Arduino was secured onto this strip board. The Arduino was held in place using velcro. 

Final testing will be put up soon.

 

Update 08/04/2011 - Final Testing Complete. More Videos to follow...


Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

is it true that neutral buoyancy has many variables that change depending on things like water temp, craft temp, salinity, depth, etc.?    as I recall submarines are constantly adjusting things to maintain neutral buoyancy.

the 1 word I noticed in your description was "Intervention".   lets hear more about that   :-)

We based our buoyancy calculations from Archimedes, such that the buoyancy force acting on the craft is equal to the weight of the water it displaces. This has been a bit tedious to get right but we are just about there.

As far as I'm away submarines have ballast tanks which they flood / empty to make the submarine positively buoyant or negatively buoyant as it dives / surfaces.

When I said intervention I was referring to 'light well intervention'. That's basically an interruption in the operation of the equipment used to retrieve oil offshore on the seabed for maintenance.  For example, often the equipment will not be operational and ROVs will be used to actuate different valves and test various functions. Ultimately, it would be beneficial to have an autonomous craft capable fo this and more.

Looks badass. Did you use TIG welding on the skin?

We got the welding done by a technician, think it was MIG but not sure. Good point though, I'll find out.

Cool!

I think your build time is exagerated. 3000 hours would be 375 days if you only worked on it 8 hours a day :P

It seems excessive but I should probably have said it's between 5 people.

That also includes all background research, initial design, detailed design, manufacturing, testing, report writing and website building - which comes out at around 15 five day weeks each, 8 hours per day, which seems a bit more reasonable.

 

This sounds like a very interesting project. How deep do you intend this robot to dive? Are you going to install ballast tanks?

I think your steel hull will prevent your compass module from working as the magnetic flux lines will be directed around your compass module. Electronic interferance from the motors and drive circuitry could also be an issue.

You may need to mount your 9DOF razor IMU outside of the steel hull in order for it to function well.

Well, structurally it should be sound to dive quite deep, it's really a question of the integrity of the hatch cover and subsequent seal arrangement we have. I'll post some picks of that soon. But the real limiting factor will be the length of time the batteries last. The motors draw around 5-6A and two motors are paired up to a single 12v 4.5 amp hour lead acid battery. So the battery life won't be too great. The only saving grace is that the motors won't be operating continously. 

It's only a prototype though and for testing purposes it's unlikely to exceed 2 metres in depth. 

You make a good point about the magnetometer, although I think we might just have to hope for the best...


Very interesting. Don't see many underwater robots on LMR. One step closer to a design to use in space or in the space station.