1. Team P14029: McKibben Muscle Robotic Fish
Project Manager: Zachary Novak
Mechanical Design Lead: John Chiu
Lead Engineer: Seaver Wrisley
Controls and Instrumentation Lead: Felix Liu

2. Agenda Project Background Propulsion System Feasibility
Problem Statement
Pressure Drag Calculations
Deliverables
McKibben Muscle Test Results
Updates to Engineering Reqs
Control System Feasibility
Feedback from Last Review
Electronics Schematic
Concept Breakdown
Logic Flow Chart
Internal System Diagram
Microprocessor Options/Selection
Subsystem Identification
Solenoid Options/Selection
Critical Subsystem Identification
Solenoid Control Circuit Schematic
Orientation Control Feasibility
Kinematic Analysis
Preliminary Subassembly Design
Buoyancy Calculations
Design Concept
CAD
Power System Feasibility
Materials
Updated Budget Assessment
Power Flow Analysis
Power Source Selection
Updated Risk Assessment
Battery Life Analysis
Detailed Design Schedule

3. Project background

4. Problem Statement
This project is designed to prove the feasibility of McKibben muscles for use in underwater robotic applications, and to develop core technology and a platform for other teams to use in the future. The project specifically seeks to develop a soft-bodied pneumatic fish that looks, moves, and feels like a fish. The robotic fish should be capable of swimming forward, backward, and turning, most likely using Body Caudal Fin propulsion, and the primary mechanism for generating the swimming motion must be McKibben muscles.

5. Deliverables
A functional prototype which meets all customer requirements, and that may be used as a platform to be expanded upon by future MSD teams
Detailed documentation covering project design, testing, and fabrication
Appropriate test data ensuring all customer needs are met
Detailed user manuals for operation and troubleshooting
Suggestions for future expansion

6. Engineering Requirements
Selected engineering specifications
Maximum turning radius: Two body lengths
Maximum height: 3 feet
Operation time: .25 hour
Corrosion spec: ASTM B-117
Safety
Maximum voltage present: 24V DC
Maximum allowable pressure: 70psi
Maximum pinching force in joints: 10lbs
Body and fin motions: 30% tolerance to published values (next slide)
This will be difficult to see, maybe we should mark this up to focus on important sectors and/or blow up certain part(s) of it

7. Questions from last review
Buoyancy Calculations
How do known components affect the buoyancy? Will we be able to offset this to achieve our neutral buoyancy desired?
Power Analysis
How much power is needed to run the components and how does that translate into expected operating time?
Spatial Analysis
How will all the components fit? How big will the robot be?
Muscle Testing
In the operating range of the centrifugal pump, will the muscles be able to produce enough force/displacement to make the fish “swim”?

8. Concept Breakdown

9. Internal system diagram

10. Subsystem Identification
Skin
Control System
Body Structure
Microcontroller
Voltage Regulation
Frame/Supports
Programming
Locomotion System
Wiring
Linkages
Orientation System
Muscles
Air Bladders
Actuation System
Power System
Pump
Battery
Solenoids
Manifold
Waterproofing
Plumbing

11. Critical Subsystem determination Guidelines
Highest Technical Risk
Most Challenging Technically
Most Important Engineering Requirements
Most Important System Level Behavior

12. Critical Subsystem Determinations
Actuation System
Locomotion System
Control System
Power System
Honorable Mention: Orientation Control System
“Propulsion System”

13. Orientation control system feasibility

14. Buoyancy Calculations

15. Design idea Inflatable Air Bladders Side Mounted (Trim Tanks)
Will allow Roll Control
Front and Rear Bladders
Will allow Pitch Control
Passive System
Manually inflate
Fine tune for neutral buoyancy before run
Butyl/Schrader Valve
Capacity (4 total)
20-25 cubic inches each
Easily fit design footprint
Feasible

16. Power system feasibility analysis

17. Power flow chart

18. Power consumption (Main components)
Thanks Thermoelectric team!

19. Summary
Found that Lithium polymer batteries were the best combination of power, weight, and cost.
Does require the use of a “smart charger” for safe charging.
Presents additional risk item: Battery catching on fire, happens during charging if done incorrectly.
Action to mitigate risk: Design fish such that battery can be removed for charging.

20. Decision/lifetime analysis
Class: Lithium Polymer (LiPo)
Specs: 25.9V, 4 Amp-Hour (4000mAh)
Weight: 1.41 pounds
Cost: $52
Expected Battery Life Analysis

21. Propulsion system feasibility analysis

22. How much force is required?
Pressure and friction drag forces act to slow the fish down
The muscles, in order to move the fins, must overcome:
Pressure drag of the fin due to rotation
Pressure drag on the fin due to apparent incoming fluid velocity
Reactions from the other fins
Friction drag slows the fish, but is NEGLIGIBLE as far as muscle force is concerned

23. Preliminary Drag Calculations

24. Free body Diagrams

25. Drag Calcs (continued)
The pressure drag force is dependent on the perpendicular velocity squared.
The torque is found by integrating the drag force times distance along the fin section.

26. Drag calcs (continued)

27. Drag calcs (continued)

28. Muscle testing Test Rig Components: LabVIEW Interface Load Cell
Air Compressor
Data Gathered
Force vs. Pressure
Deflection vs. Pressure
To get Strain (%)

29. 1st Round of Testing - lessons
Dead zone due to space between tubing and fabric mesh. 30 psi was the pressure required to take up the initial slack between the tubing and mesh.
Slope inversely related to rubber stiffness, and directly related to the ratio of inner circumference over wall thickness

30. 2nd round of testing
Assembled new muscles with existing tubing and fabric mesh.
Used tubing with high inner circumference to thickness ratio (it was thinner).
Made sure there was no space between tubing and mesh.
Tested the effect of using a slightly smaller mesh than needed, on the same tubing.

31. 2nd testing session results
Tighter mesh nearly eliminated the dead zone
Obtained a force of approximately 4 pounds at 20psi

32. 2nd testing session results
No significant difference seen between mesh types as long as they are tight to the tubing, ~13% 20psi

33. Force feasibility
Force required due to overcome pressure drag with a muscle lever arm of 4cm (1.57”): 1.83 pounds
Force produced by first set of muscles: 4 pounds
Reaction forces from other fin sections are significant, but also actuate out-of-phase. Are ultimately due to the drag as well, so should be on the same order.
Clearly within feasibility, using a muscle assembled from a limited selection of scrap materials.

34. Strain feasibility Strain level of 13% at 20psi found during testing.
A lever arm of 4cm, and maximum angle of 30 degrees requires a 30.8cm (12.2”) muscle
Has to actuate the section 30 degrees, as well as accommodate 30 degrees of motion in the other direction
Larger muscles can be used, making it possible to lower the lever arm length, decreasing the required muscle length.

35. Control System Feasibility analysis

36. Electrical control Schematic

37. Logic diagram

38. Microcontroller options

39. selection: Arduino mega 2560
PROS
Vast amount of resources available
54 Channel for future expansion capability
8 KB SRAM
Ease of coding and debugging
Shields compatibility
Used by Roboant team
CONS
More expensive ($50)

40. Solenoid valve options
Key requirements:
Inexpensive
Compatible manifolds
Ease of system integration (electrical connections)
Reliability
Additional considerations: Prior experience (reputation)

41. Solenoid selection: Pugh analysis
Clippard valves were selected as the best option
Superior for all critical aspects except for power consumption

42. Selection: Clippard 15mm 3-way Valves
12V and 24V choices
Moderate but acceptable power consumption
Several manifold options
Variety of wiring options
Least expensive, total cost around $160
Used in previous air muscle projects at RIT, such as muscle test stand
Professor John Wellin has been using them for several years controlling flow of water
* There are some similar valves in the lab currently that we will can use for testing purposes. If they end up working well enough we may not need to purchase these at all.

43. Solenoid control circuit

44. Fish Motion Analysis Forward Tuning Parameters
Turning Tuning Parameters
Value
l1 [in] 12
l2 [in] 6
l3 [in] 3
l4 [in]
Theta_1 [deg] 20
Theta_2 [deg]
Theta_3 [deg]
Theta_4 [deg] 10
Theta_2 Phase Delay [Rad]
Theta_3 Phase Delay [Rad]
Angular Frequency [Rad/Sec]
6.28
Forward Tuning Parameters
Value
l1 [in] 12
l2 [in] 6
l3 [in] 3
l4 [in]
Theta_1 [deg] 30
Theta_2 [deg] 45
Theta_3 [deg] 75
Theta_4 [deg] 10
Theta_2 Phase Delay [Rad]
3.5
Theta_3 Phase Delay [Rad]
Angular Frequency [Rad/Sec]
6.28

45. Preliminary Subassembly Design
Side View
13.5in
36in

46. Isometric view

47. Preliminary Subassembly Design
Front View

48. Preliminary Subassembly Design
Top View
Air Bladders
Tail Segments
Sealed
Compartment
Pump
Air Muscles

49. Preliminary Subassembly Design
Sealed Compartment
Arduino Microcontroller
Foam or other structure
Battery
Solenoid Block

50. Preliminary Subassembly Design
Materials
Outer skin and fish structure
Wire mesh for contoured outer shell
Larger gauge wire to support the wired mesh
Skin can be made of molded silicone, waterproof fabric, etc.
Sealed Compartment
Made from acrylic/plexiglass walls for visibility of internals
The walls will be sealed with waterproofing silicone filler along the seams.
Tail Segments
ABS plastic or HDPE (high density polyethylene)

51. Updated budget projection
This leaves ~$200 for:
Plumbing
Body/Skin Structures
Wiring
McKibben Muscles
Air Bladders
Linkages

52. Risk Assessment scale

53. Risk Assessment updates

54. Schedule: Next gate

55. Questions? Concerns? Feedback?