P14029: Robotic Fish

Detailed Design

Table of Contents

Design Overview

The head consists of a watertight acrylic box made of 3/8" and 1/4" acrylic sheet. At the bottom of the head box is the centrifugal pump feeding water into the solenoid valves that are directly in front of it. Above the pump and solenoids is the Arduino tray, which holds the Arduino Mega microcontroller and the batteries. There is a layer of Silicone Based Rubber (SBR) gasket between the edge of the main box and the lid, and another layer between the lid and the Arduino tray in order to give the electronics another layer of water protection. A rain detector is also installed to sense any water leaks coming through the box or from the internal water lines and fittings.

The tail sections are made of 1/8" acrylic sheet, but these are only flat plates and the water is free to flow around them. The tail fin itself, will be a sheet of laminated paper, as a the lamination is waterproof and provides the level of stiffness/compliance desired at a low cost. The rest of the fish will be covered by a rubberized fabric skin in order to meet the customer requirements of "feeling like a fish". This skin will be a spandex sock dunked in silicone and made to fit over the fish, secured to the acrylic and to itself with Velcro.

The muscles are hard-mounted to the front of the head box, pressurized by the flow of water from the pump and solenoid valves. Fishing line is then routed from the free ends of the muscles out to the tail sections in order to pull them side to side.

Here is the Bill Of Materials (BOM), as of 12-18-13.

Drawings & Schematics

CAD model
CAD model
CAD model

The full CAD directory may be found here. Note that the skin is not modeled in the drawings, but that the inner workings will in fact be enclosed. The skin attaches to the fish with Velcro on the bottom of the acrylic pieces, and then there is another Velcro strip along the bottom where the skin overlaps after going around the fish and back onto itself in order to secure the fabric's free edge. The plastic tubing and fishing line (actuation cables) are also not shown in the model.

A noteworthy aspect of this design, is the waterproofing features being implemented. Between the lid and the box walls, there are layers of SBR gasket as explained above, and there are additional features to seal the corners of the box. Chemically bonding acrylic glue was highly recommended by the shop staff, and a double-layer silicone caulk method was used successfully during other MSD projects. As sealing is critical to the viability of the project, our design incorporates the acrylic glue at the joint along every edge of the box, with silicone caulk along both the inside and outside. This makes for a triple-seal, minimizing the risk of leaking.

A piping flow chart was also created in order to determine the fittings needed for routing the tubing, and in order to find the best combination to reduce cost. This diagram tracks the path of the water from the outside environment, into the box, through the pump and solenoid valves, and then into the muscles and back out the exhaust port of the solenoids and through the box to the environment.

Piping Schematic

Engineering Analysis

Buoyancy Analysis

One of the questions about the design was how stable it would be. This was addressed with a buoyancy analysis in SolidWorks. The first step was showing the weight, buoyancy force, center of gravity (CG), and center of buoyancy (CB) of the fish before adding weights or flotation. SolidWorks' mass properties features are used to display the CG point and the total mass, and the CB point and buoyant force were obtained by turning all of the parts to the density of water, including the empty volume of the box (which is why the internals don't appear in half of the pictures - they're part of a block of water). Weight was then added in the form of a steel block at the bottom of the fish, and buoyancy was added to the top in the form of arm floaties (the pool accessory for small children to keep them afloat).

Before weight and buoyancy modifications, the model weighed 15.2 pounds and had a buoyancy of 16.7 pounds. The CG and CB points are at the origin of the purple model for each of their respective pictures.

CAD model
CAD model

After adding the weight at the bottom and the air bladders at the top, the model weighed 19.12 pounds with a buoyancy of 19.03 pounds, showing the range of possible corrections that can be made with weights and air bladders.

CAD model
CAD model

Technically only the weights were required to achieve neutral buoyancy, but using more weights and air bladders makes a more stable system by raising the CB and lowering the CG. The final locations of these two points are shown below, the CB on the left and the CG on the right. The separation between the two points is approximately 1.6". For comparison, the distance from the bottom of the battery to the top surface of the head box's lid is 9.25". This analysis also neglects the weight of the water flowing through the system and any dynamic effects, but again there is more than sufficient design space. The floaties are both attached by Velcro to the top of the head box, and the weight is also suspended by a Velcro saddle. Velcro was chosen as the best attachment method because of its ease of assembly and disassembly as well as it's cost effectiveness.

Note that CB and CG have nearly the same position along the length of the fish, and also side to side. This means that the pitch of the fish will naturally be very close to horizontal as opposed to nose up or nose down, and that the roll of the fish will also be near zero. Both of these will also be able to be adjusted by modifying the air in the sides of the bladders or by shifting the weight.

CAD model

Power Consumption

Power Calculations

Prototyping and Testing

A testing rig was constructed in order to bolt a pair of muscles and a tail section together for testing two main objectives.

  1. The first was to prove feasibility of the pump and solenoid design. It was found that the SMC 10mm solenoid valves were capable of running the fish, but that the yellow Clippard drum-shaped solenoid vales were not capable of passing water. Using the SMC valves, the system worked very well.
  2. The second objective of this test rig was to be able to investigate the effects of changing certain parameters on muscle firing timing and actuation speeds.
Tail Testrig
Tail Testrig

A second round of testing was later conducted with two fin sections, giving good results for actuating multiple fins. The video of this testing may be viewed here. Quicktime is required and the link must be opened in the same window/tab.

Tail Testrig

Testing was also done using two power supplies borrowed from Ken Snyder in the Electrical Engineering Department, running the pump and solenoids at less than their specified voltages as explained further below.

Preliminary Coding

A software flowchart was created in order to represent the motion of the fish, as controlled by the Arduino.

Software Flowchart

Electronics Components


Bill of Materials

The Bill of Materials was updated as the design matured, and tailored to the materials available in the lab in order to reduce expenses.

Bill Of Materials updated 12-18-13.

Current Bill Of Materials, continuously updated through MSDII and in Excel spreadsheet format.

Test Plans

Projected Schedule for MSD II

MSDII Test Plan
MSDII Test Plan

Risk Assessment

Current high priority items on the Risk Assessment, 12-18-13.

Piping Schematic

Mini Design Reviews

During the Detailed Design Phase, we had meetings regarding specific portions of the design, with several faculty members who are experts in those fields.

DDR faculty consultations:

DDR Faculty Consultation with Dr. Gomes

DDR Faculty Consultation with Rob Kraynik and Dave Hathaway

DDR Faculty Consultation with Dr. Walter

DDR Faculty Consultation with Dr. Day

Detailed Design Review

The final Detailed Design Review (DDR) was held with Rick Lux (faculty guide), Dr. Lamkin-Kennard (our customer), and Dr. Walter, Dr. Gomes, and Dr. Day (faculty with experience in the relevant fields) present.

Here are the Agenda, DDR Presentation, and a summary of the feedback received incorporating action items.

MSDI Wrap-Up

The team got together and discussed what things could have gone better during MSDI, and any things that we wanted to improve in MSDII. The results are listed here.

Table of Contents MSD I Home

Project Definition Systems Design Subsystem Design Detailed Design

Table of Contents MSD II Home

Subsystem Level Prep Subsystem Level Build & Test Subsystem & System Level Build, Test, Integrate Systems Level Build, Test, Integrate Verification & Demonstration of Results
MSD - The Postseason