ROV Final Report

Final Report for Remotely Operated Vehicle, Jenny

Submitted December 10, 2014

Neptune Inc. 2609 Draper Dr.

Ann Arbor, MI 48109

Sanjana Belani

Caleb Irvin

Ryden Lewis

Emily Thayer

Ryan Wilkie

Product Engineers

Submitted in response to NASA BAA NNHI4ZDA001N for the development of ROV for

Ross Ice Shelf expedition

1

Table of Contents

Executive Summary………………………………………………………………….2

Introduction…………………………………………………………………………..3

Design Overview……………………………………………………………….…….3

Model Description……………………………………………………………………5

Model Performance…………………………………………………………………..9

Full-Scale Performance……………………………………………………………..10

Conclusions and Recommendations……………………………………………...…12

References…………………………………………………………….......................13

Appendix…………………………………………………………….........................14

2

Executive Summary

NASA issued a BAA requesting a Remotely Operated Vehicle (ROV) design to be deployed

under the Ross Ice Shelf in Antarctica. We are tasked with creating a prototype ROV that is able

to operate in an arctic subsea environment, and perform one of the two tasks requested by

NASA. Our ROV has been designed for the valve turning task.

NASA requires that the ROV has a mass of less than 15 kg and does not exceed the dimensions

of the transportation container (58 cm × 40 cm × 30 cm). The ROV must also be quick and

highly maneuverable in all 6 degrees of motion (forward, backward, up, down, left, right). The

thrusters, payload, and camera must be protected against collisions. The ROV must achieve a

minimum velocity of .5 m/s. Only materials provided or pre-approved by NASA may be used in

the construction of the ROV.

To complete the valve turning task, our team created a prototype ROV, Jenny. Our ROV uses a

rectangular prism shaped frame, which allows for easy attachment and adjustment of

components. The dimensions of our ROV are 49 cm × 31 cm × 29 cm, which fit within NASA’s

requirements. There is an inner frame to secure the payload, and an outer frame to attach the

thrusters, camera, buoyant material, and ballast. The open space between components allows us

to easily slide the payload into the frame since the payload is to be attached on location. We

designed the ROV to have a wide face, in order to increase our chances of hitting the valve.

However, we limited the surface area of the ROV face by minimizing the height in order to

minimize drag. All 4 thrusters are placed on the rear of the ROV to provide a high forward

speed. In an attempt to maximize the stability, we placed the payload (bulk of the mass) toward

the bottom of the ROV, and the buoyant material towards the top of the ROV.

The ROV is completed. A complete project schedule can be seen in Appendix A. We tested the

ROV at Canham Natatorium and were able to complete the valve turning task in a time of 5

minutes and 58 seconds. While testing at the Marine Hydrodynamics Laboratory (MHL), our

ROV Jenny achieved a top speed of 0.559 m/s.

A scaling factor of 2.5 will be used when scaling up the ROV prototype to full scale. Different

materials need to be used in the full scale model, due to the extreme temperature and pressure

differences in the actual environment NASA will be using the ROV in. We have decided to use a

frame made of high-strength steel and propellers made of stainless steel due to their high strength

and anti-corroding properties. Because of its high strength and low density, we have decided to

use syntactic foam for buoyancy. Lithium-ion battery would be used as a power source for the

full-scale ROV since it has a high energy density, a good weight to energy ratio, high energy

efficiency and a relatively long life-cycle.

The following report presents our ROV design, summarizes our design rationale, discusses our

model performance, and describes the scaling process and suggestions for the full scale ROV.

3

Introduction

NASA has issued a BAA asking for designs of a Remote Operated Vehicle (ROV) to deploy

under the Ross Ice Shelf in Antarctica. This ROV will investigate the sub-ice sea and determine

how it has changed since the original expedition in 1977. The ROV must not exceed a total mass

of 15.0 kg or the approximate dimensions 58 cm × 40 cm × 30 cm. It must be able to go

forward, backward, up, down, left and right. It should have a rigid, cemented body structure that

is easily flood-able with a quick attachment point for the camera. The payload must be

detachable and should be strategically positioned to provide a clear line-of-sight for the video

camera. The attachment point for the tether should be on the bottom so that the tether can hang

straight down. Up to four thrusters may be used and each thruster must be wholly located inside

the structural frame and protected from bumps and collisions. The ROV must achieve a

minimum average velocity of 0.5 m/s. The ROV must be designed to either turn a valve or

collect a ring. Only approved building materials must be used; additional materials may be

approved by NASA on request.

The deep, turbulent waters of the Antarctic present several limitations to the design of the ROV.

The temperature of the waters in the Antarctic is approximately -2.16˚C (Patton, 1999). Because

of the increased depth, the pressure acting on the ROV will be much greater. Also, the density of

the salt water in the Antarctic will be higher compared to the density of the freshwater of the

testing facility. This will lead to greater buoyancy acting on the ROV. These extreme

environmental changes must be accounted for in the full-scale design of the ROV.

Design Overview

Our ROV, Jenny, uses a rectangular prism shaped frame, which allows for easy attachment and

adjustment of components. We decided that making our ROV neutrally buoyant would be the

best approach to the valve turning task, as well as the maximum speed test. In an attempt to

maximize the stability, we placed the payload toward the bottom of the ROV, and the buoyant

material towards the top of the ROV. Seen below in Figure 1 is our ROV, Jenny.

Figure 1. Rear Perspective View of our ROV

4

Frame The frame of our ROV is in the form of a rectangular prism. It is 49 cm long, 31 cm wide

and 29 cm high. There is an inner frame to secure the payload, and an outer frame to attach the

thrusters, camera and buoyant material. The frame was designed with a wide face so that the

valve can be hit easily, but with a minimal height to reduce drag.

Payload The payload is inserted in the back of the vessel and it rests on two supporting pipes

running along the bottom of the ROV.

Thrusters The four thrusters are located in each of the back side’s four corners. The thrusters in

the four corners are controlled individually, and allow the ROV to be easily turned and pitched.

Buoyancy For buoyancy, we added two buoys to each side of the top and one buoy in the center

of the top. We also placed two empty 20 fl oz. Gatorade bottles below the buoys on each side.

This allows the center of buoyancy to be above the center of gravity, resulting in a stable vessel.

Camera The camera is located in the front corner of the frame of the ROV, to provide a good

viewing angle, as well as protect it from collisions. The camera is secured to the frame of the

ROV with the help of zip-ties in such a way that there is minimal movement. This helps in

maximizing the quality of our video feedback.

Control System The control system uses two switches and four buttons. The two switches

control the bottom thrusters, forwards and backward. Each top thruster has two buttons, one for

forward and one for backward. The switches are placed on the top of the controller, and the

buttons are on the front of the controller so that they are easy to reach. This makes the ROV

highly controllable because each motor can be turned on in either direction, so the ROV may be

turned in any direction using any combination of motors.

The components and their placement on our ROV can be seen in Figure 2 and Figure 3.

Figure 2. Perspective CAD Model View (Rear) of ROV

Buoys

Camera

Thruster

Gatorade Bottle

Payload

5

Model Description

Dimensions and Properties The ROV frame measures 49 cm × 31 cm × 29 cm. The frame is

constructed entirely of PVC pipe, and cemented together. Five buoys are connected to the top of

the frame with zip-ties. Two Gatorade bottles are attached beneath them, one on each side, by

zip-ties. Thrusters are attached at each of the back corners. The payload canister slides into the

frame, where it is supported by two PVC pipes running down the center of the frame. The

payload and the camera are both zip-tied to the frame for easy installation. Our CAD models

seen in Figure 4 and Figure 5 display the dimensions.

Figure 4. Rear CAD Model View of ROV Figure 5. Top CAD Model View of ROV

Figure 3. Perspective View (front) of ROV

6

Mass Budget

Below in Figure 6 is a table of our detailed mass budget. Most of the mass of our ROV comes

from the payload and tether.

Item Quantity Mass(g) Total(g)

Payload/Tether 1 < 10000.0 < 10000.0

PVC Pipe 400 cm 976.4 976.4

Thrusters 4 236.2 944.7

PVC T’s 16 28.5 456.0

Buoys 5 46.5 232.6

PVC 4 Ways 6 37.2 223.2

Camera 1 212.5 212.5

Zip Ties 33 1.8 58.7

PVC Corners 2 29.2 58.4

Gatorade Bottles 2 28.5 57.0

Total Mass 13291.5

Center of Gravity

The center of gravity was found by finding the moment of each mass in relation to each

dimension and dividing the sum of the moments by the total mass. The center of mass relative to

our ROV can be seen in Figure 7 and Figure 8.

XCG = ((236.175 × 49) × 4+10000 × 16.3)/(944.7+10000)

XCG = 19.122

YCG = (212.5 × 7.75)+2(236.175 × 15.5)/(944.7+212.5+10000)

YCG = 0.804

(31/2) + YCG = 16.304

ZCG = (212.5 × 29)+(236.175 × 29 × 2)/(212.5+944.7+10000)

ZCG = 1.78

Figure 6. Detailed Mass Budget of Our ROV

7

Circuit Design

When designing our control box, we had to create a circuit to power the four thrusters. We

decided to power each thruster individually, and to use two single pole double throw switches

and four buttons. The bottom two thrusters are each wired to a single pole double throw switch,

meaning the operator presses forward on the switch to engage the thruster forward and

backwards for reverse. A colored wire and its corresponding black wire are attached to each one

of these switches. The switches are operated with the user’s thumbs. The top two thrusters are

controlled with buttons. We opted to use buttons rather than switches because they were easier

for the operator to press with his/her index finger (an image of the control box is seen in Figure

10). Because the buttons are single pole single throw, each of the two top thrusters required two

buttons each: one for forward and one for reverse. For each set of buttons, the thrusters’ colored

wire was attached to the forward button and the corresponding wire was attached to the reverse

button. Finally, the last remaining black wire was attached to each switch and button in series. A

circuit diagram is seen below in Figure 9 with thrusters A-D labeled on their switches/buttons.

Figure 7. Center of Mass From Front View Figure 8. Center of Mass from Top View

Figure 10. Control Box of our ROV Figure 9. Circuit Diagram Showing the Wiring

Configuration of our Controller

8

ROV Calculations

Weight

The mass of the ROV (not including the tether), 𝑚, is 13291.5 grams. The gravity acceleration

constant, 𝑔, is known to be 9.81 m/s2. To calculate the weight of the ROV, we multiplied the

mass of the ROV by the gravity constant, to get a weight of 130.38 N.

Buoyant Force

Our ROV is neutrally buoyant so we know that the force of buoyancy is equal to the weight of

the ROV. We know this through the equation of the sum of the forces.

Top Speed and Drag

The drag (D) acting on the ROV is given by the equation:

D = T = 1

2CD AROV V

2 (11)

where T is thrust, is the density of freshwater, CD is the coefficient of drag, AROV is the cross

sectional area of the ROV, and V is the top speed of the ROV. Using equation (13), the value of

forward T was found to be 25.22 N and backward T was calculated to be 8.19 N. The value of CD

is taken to be 1.28, AROV is 0.0899 m2, and is assumed to be 998 kg/m3. Using these values,

the value of forward top-speed was calculated to be 0.663 m/s and reverse top-speed was found

9

to be 0.377 m/s.

Max Power

The maximum required power is given by the equation:

Prequired =𝑉 × 𝑇

𝜂𝑜𝑣𝑒𝑟𝑎𝑙𝑙 (12)

where V is the top speed of the ROV, T is the thrust and ηoverall is the overall efficiency of the

thrusters. The value of V was calculated to be 0.663 m/s, forward T is 25.22 N and backward T is

8.19 N and ηoverall is assumed to be approximately 0.25. The value of forward Prequired was

calculated to be 66.88 W and reverse Prequired was found to be 12.35 W.

Thrust

We tested the thrust of the 4 thrusters in the GFL tank, and calculated the total thrust to be 25.22

N. To calculate thrust, we attached each individual thruster to a lever arm, attached a wire from

the lever arm to a load cell, and recorded the voltage output from the load cell. This voltage was

then converted to force (Newtons). The calculated thrust resulted in an calculated top speed of

0.663 m/s. Through rudimentary testing at tank in GFL, we were able to record a average

velocity of 0.595 m/s. The length and width of the test tank severely limited the accuracy of the

data we collected, and we were therefore unable to do testing involving the maneuverability of

the ROV. The thrust equation used is shown below.

T =−6.8493+.019(𝑚𝑉)

2.5 (13)

Model Performance

Seen below in Figure 11 is a graph of our speed testing trials at the MHL. Our best trial recorded

a speed of 0.559 m/s, meeting NASA’s requirement of at least 0.5 m/s. A table of our trial data

can be seen in Appendix B.We had an average speed of 0.469 m/s and a standard deviation of

0.07. Our predicted speed of our ROV was 0.663 m/s. Possible reasons for not reaching our

predicted speed can be attributed to our estimated coefficient of drag being lower than actual,

and the fact that our battery wasn’t fully charged when we started speed testing. There are

several reasons why 3 of the top 4 speeds do not meet the required 0.5 m/s. During 2 of our

trials, the ROV struck either the bottom or side of the pool. In the third trial, our ROV ran into

the test valve that was placed in the MHL. While we were unable to verify the consistency of our

ROV’s speed, we were able to confirm that when tested without hitting obstructions, our ROV

met the required speed.

Figure 11. MHL Testing Velocity Graph

10

We did our initial testing at the Marine Hydrodynamics Laboratory (MHL), seen in Figure

12. The water there was about 15.6 ˚C with a density of around 1000 kilograms per cubic

meter. The final competition was held at the University of Michigan Natatorium, seen in Figure

13. The water there was warmer than that of the MHL and contained chlorine. The addition of

chlorine causes the water in this facility to be more dense than previous testing locations.

During competition, our fastest and only completed trial of the valve turning task was 5 minutes,

58 seconds. We were disqualified on our first two trials due to hitting the bottom and breaching

the surface. There are several factors that lead to these disqualifications. For the first trial, we

were unable to keep the ROV off the bottom was because our Gatorade bottles used for

buoyancy were compressed from the high water pressure in the deep pool. Another issue was to

pitch our ROV, we had to be moving forward; our design did not allow the ROV to be pitched

vertically while it remained horizontally stationary. In the second test the ROV breached due to

the buoyancy being positive because ballast was removed after hitting the bottom.

Full-Scale Performance

We have designed the ROV model to be scaled-up for use in the Ross Ice Shelf, where it will be

collecting data. The full-scale dimensions of our ROV are 1.23 m x 0.78 m x 0.73 m. The ROV

will be operating in seawater where the temperature is estimated to be around - 2.16°C. It should

be capable of running continuously for 3 days. It should also be able to withstand the strong

currents, high salinity content, high corrosive properties and high pressure conditions present at

the depths of the ocean. Due to the harsh Antarctic environment, we will need to take into

account the materials we will use for the full-scale ROV.

For the material of the frame of the full-scale ROV, we have decided to use high-strength steel.

The major advantages of high-strength steel are that it has a good resistance to corrosion, is

Figure 12. MHL Testing Facility

Figure 13. Canham Natatorium

11

relatively cheap and the fact that it is commonly used, so there is much knowledge of it

(Tavakkolizadeh, 2002). It also has large yield strength, which would allow the ROV to

withstand the large pressures it would experience if it were to travel to a depth of 1000 meters.

To provide buoyancy and counteract the weight of the ROV, we have decided to use syntactic

foam. Syntactic foam provides high strength and low density, and offers very low moisture

absorption.

For the propellers, we have decided to use stainless steel since it is not prone to corrosion. Due to

the lack of lighting underwater, we have decided to put headlights next to the payload. This will

increase visibility and allow the camera to capture clearly visible images.

We have decided to use lithium-ion battery as a power source for the full-scale ROV since it has

a high energy density of 144 Whr/kg, a good weight to energy ratio, an efficiency of 95 percent

and a relatively long life-cycle of roughly 500 cycles, enabling it to work continuously for 3

days. An additional benefit is that it is not prone to outgassing if subjected to overcharging

(Bradley et al., 2001).

When our ROV is fully scaled up, it needs to be able to accomplish the task of turning a valve

180 degrees, and capable of observing the subsea antarctic environment. A major strength of our

ROV is the ease of operation. Each thruster is activated by one switch or set of buttons, allowing

intuitive control of the ROV to move forward, backward, right, left, up, and down by activating

the thrusters in different combinations. This will allow the full scale ROV to efficiently

accomplish its required tasks.

Full-Scale Calculations

The full-scale calculations can be found by scaling up the thrust, top speed and maximum

required power of the model using Reynold’s scaling.

To find the full-scale thrust, we use the following scaling ratio:

λT=λ⍴ λν2 (14)

where λT is the thrust-scaling ratio, λ⍴ is the density-scaling ratio and λν is the viscosity-scaling

ratio. We multiplied the model thrust with the calculated value of λT and acquired a value of

86.12 N for full-scale thrust.

To find the full-scale top-speed, we use the following scaling ratio:

λV=λν / λL (15)

where λV is the velocity-scaling ratio, λν is the viscosity-scaling ratio and λL is the length-scaling

ratio. We multiplied the model top-speed with the calculated value of λV and acquired a value of

0.483 m/s for full-scale top-speed.

To find the full-scale power, we use the following scaling ratio:

λpower=λ⍴ λν3 / λL (16)

where λpower is the power-scaling ratio, λ⍴ is the density-scaling ratio, λν is the viscosity-scaling

12

ratio and λL is the length-scaling ratio. We multiplied the model power with the calculated value

of λpower and acquired a value of 166.3 W for full-scale power.

Conclusions and Recommendations

We were challenged by NASA to create a model ROV within the dimensions of 58 cm × 40

cm × 30 cm, that weighed less than 15 kg and had a floodable frame. It had to reach a minimum

velocity of 0.5 m/s and be able to turn a valve. To meet these requirements we designed an ROV

with dimensions of 49 cm × 31 cm × 29 cm. The total system weighs 4.47 kg without the battery

and tether. Knowing that the battery and tether together weigh less than 10 kg, our total mass is

less than 15 kg.

Our ROV is designed to be quick and maneuverable, with a large frontal area to turn the valve

plate with. We have four rear facing thrusters to maximize the speed of the ROV. Our best top

speed recorded is 0.56 m/s, with an average speed was 0.47 m/s. Each of the thrusters are wired

individually to contribute to the maneuverability of the ROV, so that each may be individually

fired forward or backward. When the thrusters are used, the vehicle has neutral buoyancy. Using

the thrusters at a thrust less than positive buoyancy allows controlled ascent and enables the

ROV to come to the surface in case of a power failure (Serrani & Conte, 1999).

During the competition we found that our ROV was hard to control, in both the vertical and side-

to-side directions. This is due to the fact that our pitch is controlled by rear facing thrusters, and

the thrusters’ water displacement is used to move forward and try to pitch at the same time. We

found out that our tether placement on the ROV was not optimal, and was affecting our ability to

control turning the ROV.

For future use, we recommend a creating a PVC housing for the camera, to further secure it in

place and give it a more centered view. We also recommend having a vertical thruster, so the

pitch of the ROV can be controlled while it is stationary. Another issue we need to address is the

tether placement. We are unsure of where a better attachment point for the tether would be, but

through more testing, it could be determined. A final recommendation we have for our ROV

design is floodable ballast. This would allow the ROV to be neutrally buoyant at all times,

despite the changing tether weight.

13

References

Bradley, A.; Feezor, M.; Singh, H. & Sorrell, F. (2001). Power systems for autonomous

underwater vehicles, IEEE Journal of Oceanic Engineering, Vol. 26, No. 4, 526–538.

Patton, Edward, et al. (1999). "A New, Highly Efficient Deep Water ROV Buoyancy System."

Offshore Technology Conference.

Serrani, A. & Conte, G. (1999). Robust nonlinear motion control for AUVs, IEEE Robotics &

Autonomation Magazine, Vol. 6, 33-38.

Tavakkolizadeh, Mohammadreza, and H. Saadatmanesh. "Strengthening of steel-concrete

composite girders using carbon fiber reinforced polymers sheets." Journal of Structural

Engineering 129.1 (2002): 30-40.

14

Trial Time (secs.) Velocity (m/s)

1 18.67 0.49

2 21.97 0.42

3 22.24 0.41

4 16.35 0.56

Appendix B. Speed Testing Table From MHL

Appendix A. Gantt Chart of ROV Project Schedule