Senior Design Submarine

My senior design project was a competition between four teams - each designed a remote control submarine to race. Our submarine used thrust vectoring instead of conventional control surfaces. I was the team lead and built the fiberglass nosecones.
- Skip to video gallery -
The submarine geometry was designed using CATIA V5 (Hull and Nosecones) and Pro/Engineer (propulsion). We used 6" PVC pipe for the hull and pressurized it to provide compressed air to ballast tanks for ascent with a payload. Fiberglass composite noscones were molded in four steps.
The nosecones contained integrated balast tanks and vertical thrusters. The nose and tail were identical. Shown here is an initial iteration in which the nosecones were to be attached to the hull using tabs. The final iteration had integrated threads which were molded directly into the epoxy resin of the nosecone.
An exploded view of the hull assembly. Part A600 was originally a threaded PVC endcap but leaked pressure out of the hull even with liquid teflon paste applied to the threads. We ended up using a rubber plug instead and used the threaded endcap to retain it axially because it was only rated to 5psi. We were pressurizing to 60psi+.
Exploded view of the assembly as-built showing the nosecone, retainer nut, rubber end-plug (yellow), electronics and hull.
The nosecones were optimized for low drag using a half-symmetric 3D FLUENT analysis along with a CATIA V5 parametric model. Shown here is turbulent kinetic energy on the surface of the sub which indicates regions of energy loss and hence parasitic drag. Hand calculations were used initially assuming that the hull was comprised of a sphere (pressure drag) and tube (skin drag).
Two different nosecone lengths were compared (8" & 9"). The wake was triggered by the vertical thruster ports at the top and bottom of the tailcone. Suprisingly, although the 8" nosecone (more blunt) had more energy dissipation at the forward thrust port it also had a smaller wake structure and overall lower energy dissipation.
8" nosecone velocity plot showing stagnation zones and vortices within the thrust ports. The vertical thruster motors were not included. A 1 m/s freestream velocity was used. The actual sub could go almost that fast.
9" nosecone velocity plot showing larger wake structure.
9" nosecone 3D wake visualization showing isosurface of 0.5 m/s velocity.
The PVC hull was analyzed for stress under internal pressure. Regions where we modified the PVC (holes, mounts etc.) were refined and checked for safety at higher pressures.
The objective of the submarine competition was originally to pick up an object on the pool bottom. We simplified our device into a metal hook attached to the bottom of the sub.
Molds were cut using a CNC 3-axis mill out of high density foam. Dowel pin holes were located in the same CNC run for alignment accuracy. The molds were coated in automotive primer and lubricated with Vaseline prior to layup.
Strips of medium weight fiberglass cloth were laid down around the periphery of the nosecone mold. A halogen lamp sped cure times in the sub-freezing weather.
I made the first nosecone far too thick, which made it very hard to separate from the mold. The primer had fused with the epoxy, even through the thick coat of Vaseline. I used scalding hot water to soften the part enought to pry it out of the mold. Remnants of the primer can be seen on these two parts. We decided to use the thicker part as the forward nosecone for impact durability.
Once the main form of the nosecone was complete, the ballast tanks were carefully formed within the cones using a multi-step process of taping an internal form wall and picking it out from behind the cured section of fiberglass. I thought about using foam and removing it with acetone but this worked.
This is the mandrel I used to form the vertical thruster ducts. It was inserted through holes in the cones, lubricated, laid up and removed after cure (using torsion). Fillets were formed using sculpted Playdough.
Finished nosecone after four step featuring and trim. Feature steps were:
  • Cone body
  • Ballast tank wall
  • Vertical thruster tube
  • Integrated epoxy threads
The leads are for the vertical thruster and there is also a blow tube for the evacuating the ballast tank. The blow tube is opened by a solenoid switch within the pressurized hull.
A frontal view of the unpainted complete nosecone. You can see some crazing on the tip from an accidental drop (10" fall, ~40 lb sub). No permanent damage was visible. This nosecone had about 6 plies of fabric near the tip.
Two completed nosecones after primer and wet sanding for smoothness. Here you can see the integral epoxy threads. This fastening mechanism simplified interfacing between the hull and nosecones and led to a refined and streamlined look.
The finished nosecone including awsome warfish paint scheme applied by Jason. I had previously painted the sharkskin fade on and the face went over it really well. We had to have bold colors on the head to maintain orientation when piloting from above water.
The sub in a "developemental" pose. Note that the nose is missing and the tail of the sub is at the top. Horizontal thrusters are shown with safety shrouds installed. The propellers were quite dangerous when turning at full speed.
The complete sub.
We had to ballast the submarine to neutral bouyancy, so we used my bathtub. Ballast and balance are two separate and equally important issues.
Our sub had wiring issues at the same time as the A380. It gets complicated pretty quickly. We had six thrusters, each drawing about 3 A for a total peak current of around 18 A. Each thruster set was controlled by an independent waterproof reversible voltage regulator. We found out after purchase and planning that the regulators gave less voltage in reverse. This fact coupled with our propellors forward thrust bias made reverse very weak, especially in the vertical thrusters (so, the upwards direction).
Our sub in the water, trawling just beneath the surface.
Being acrobatic came a little too easy for our sub. It's an effect called "dolphining" whereby the submarine won't stay level in forward flight causing a tumbling (pitch) instability. They make inline control mixers that correct for it but we didn't anticipate the handling dynamics well enough. We did consider the position of the CG relative to the center of bouyancy and the center of drag, but failed to anticipate this instability.
The ballast tanks were filled prior to operation by inverting the sub to let all the air out of the tanks in the nose and tail. This made the sub neutrally bouyant until the ballast tanks were blown by remote control to lift a payload.
Triage. We had loose electrical connectors which led to all kinds of difficulty. Current would leak throught the highly ionized pool water and they would also come loose. Gven more time, better connectors would have been found.
The sub breaching the surface at high speed. You can see the ripple behind it where it surfaced.
A few of videos of the sub.

Our sub was quite the crowd pleaser and it retired to a podium in the Purdue machine shop. Its internals have been canabalized for future ME senior design projects.
Resume - Lebenslauf