P14254: Underwater Thermoelectric Power Generationedge.rit.edu/edge/P14254/public/FinalDocuments/FinalPresentation.pdfProcess Problems and Lessons Learned Design and Deliverables Quality

P14254: Underwater Thermoelectric Power Generation

Team Members: Charles Alexander, Tom Christen, Kim Maier, Reggie Pierce, Matt Fister, Zach Mink

Guide: Rick Lux

Agenda● Results● Current Status

○ Mechanical System○ Electrical System

● Process● Problems and Lessons Learned● Design and Deliverables Quality● Opportunities for Future Work

Results

● Majority of engineering requirements were met● We were not able to:

○ Efficiently use MPPT○ Charge a battery

Current Status: Mechanical

Assembled System Enclosure Lid Removed

Current Status: Mechanical

● System is working, results are repeatable, 20W of power can be generated at 3.5% conversion efficiency.

Date Heat In (W)

Power Out (W)

Mar 18 500 14.4

Mar 18 630 20.17

May 07 500 13.8

May 07 630 20.3

May 07 630 20

Current Status: ElectricalSubsystems successfully tested:● MPPT (Converter, control circuit, code)

○ Modified breadboard converter achieved 47% efficiency

● 5V Regulator● Load Relay

Not tested: ● Battery monitoring● Full electrical system

integration

Description and Sources of Inefficiency

When modeling thermoelectrics during testing, we experienced a total loss of 9.4W out of 17.8W input:

● Switching Losses ~ 7W● Gate Drive Circuit ~1W● Diode ~ 1.4W

Process

● Team Performance○ Team cooperation was excellent○ We stuck to our expert decision model○ We strove to deliver high quality and professional

products

● Testing○ We performed and documented all the tests we

planned to perform (except the integration)

Tests Successfully Performed

● Mechanical: ○ Heat Sink testing○ Clamping test (x2)○ Insulation test○ Thermoelectric

testing○ Waterproof testing○ Initial heated test

using quartz○ Second heated test

using thermoelectrics

● Electrical:○ Regulator○ Controller testing○ Battery voltage

profile test○ ADC testing○ Relay testing○ Op-Amp testing

Process

● Budget: $ 1800○ Total Cost: $ 1719○ Unit Cost: $ 716

● Schedule○ Followed the schedule for most of MSD II

■ Fell behind when we started facing difficulty testing some electrical subsystems

○ Underestimated the risks we would face

Process

● Problem Solving○ Process often followed informally○ Communication of current issues was lacking○ Process was effective when followed○ We have a problem tracking sheet listing all

problems encountered in MSD2(Problem Solving Document)

Problems and Lessons Learned

1. Ordering Issues

2. No Smart Battery Functions

3. GaN FET Assembly Difficulties

4. Waterproofing

5. ATtiny Fuse for 64 MHz Clock

6. Broken ATtiny Debugger

7. Poor Efficiency ZETA

● Ordering Needs Attention

● Perform Focused Tests

● Test Earlier (if possible)

● More “focused” testing

Design & Deliverable Quality

● Mechanical:○ Insulation design works within uncertainty○ Didn’t break any thermoelectrics○ Heat sink thermal resistance better than desired○ Conversion efficiency of 3.5% close to design efficiency of

3.62% (3.3% error)

● Electrical:○ Design needs modification○ Physical product is not available

Opportunities for Future Work

● Second iteration of the electrical system○ Different DC-DC converter topology

■ Improved gate drive○ Better current sense (differential measurement?)○ MPPT algorithm

To generate 20W of power you need...

70% efficient converter 900W of heat

85% efficient converter 741W of heat

95% efficient converter 663W of heat

Questions?

Results