1. Team Solar PDR Solar Energy Predictor Capstone Spring 2009 Daniel Seltzer Jim Love Rob Chadil Eric Dickey

2. Brief Overview A portable device used to accurately estimate the amount of solar energy that can be harvested at an arbitrary location. The device should take into account solar panel tilt, elevation, and geographic location. System output will show the power a solar panel could theoretically harvest over the course of a year at the chosen location.

3. Purpose The need for this device comes from rapid growth in the commercial and home solar Photo-Voltaic (PV) industry. At present, the margin between making or loosing money with a PV installation is quite small. More accurate information means better decisions.

4. Theory – Hemispherical Photo

5. Theory – Sky Detection (Estimated White Sky/Black Object ) N EW S

6. Theory – Sun Path (1/29/2009) W N E S

7. Sun Path Results (January 29, 2009 – Clear All Day – 4.89 kW·h / M 2 )

8. The User Perspective GPS Tilt Compass Personal ComputerData Capture Device

9. System Organization Sensor Module Micro- controller GPS Compass Tilt Sensor Camera (IP) Camera (IP) Circular Fisheye Lens Battery Power System SD Card/USB ThumbDrive Character Display Number Pad Single Board Computer Pic Buff (SRAM) Data Capture Device Software Modules Windows Laptop Sun Tracks Algorithm Image Processing – Sky/Obstruction Alg. Final Power Calculation Camera Orientation Compensation User Interface

10. Single Board Computer Small (3.8’x4.8’) Single Board Computer. ARM9 Based Light weight Embedded Linux OS preloaded, with support for Debian if needed. Used to coordinate between Sensor Module, Camera, and removable media for data storage.

11. Tilt Sensor SRAM 1 MB SRAM 1 MB 32 Bit Parallel Microcontroller 32 Bit ARM7 50 MHz Microcontroller 32 Bit ARM7 50 MHz GPS Receiver Compass Module User Interface LCD Screen Keypad 3 Channel Analog 8 Bit Parallel UART I2CI2C Sensor Module Board Image Capture Module User Interface 4x20 character LCD for displaying user messages Expansion space for graphic LCD Allows image preview capability Capable of displaying more detailed messages 16 button keypad for user input Compass Module Honeywell HMC6343 single chip package Outputs heading message to predict orientation of sun relative to solar panel Sensitive to the presence of magnetic materials Must avoid using iron or steel components in design GPS Receiver Acquires position fix and outputs message to microcontroller Provides a time stamp for the image Data used to predict sun’s position Tilt Sensor Single package, 3 axis accelerometer Analog Devices ADXL330 w/analog output Used in combination with compass data to compute orientation of the device relative to Earth fixed frame SRAM 1 MB (2 256x16 chips) 32 bit data bus interface Used for image buffering (if graphic LCD is used) Microcontroller NXP (Phillips) LPC2292 series 32 bit ARM7 core 50 MHz CPU frequency 24 bit address space + 32 data bit external memory interface

12. Image Capture Small camera fitted with fish eye lens. Communication over Ethernet simplifies interface. Gives 180° view of horizon.

13. User Interface 1/24/2009[8:52:01] Solar Predictor Console 1/24/2009[8:53:12] Users See Status messages here 1/24/2009[8:53:56] Not an interactive prompt 1/24/2009[8:54:03] Basically just a status message box. Solar Energy Predictor v1.0 Sky boundaries Sun path Diffuse light approximation Sky Threshold Some other thing to select Gamma pre-distortion Show… File Edit Import Help Textual Output Area. Numbers such as total yearly expected, total daily expected. Used to compliment graphs. 6/23/2009 2012 Day Select Year Select

14. System Verification Need a way to tell if system is giving reasonable results. Hand-Calibrate Estimate for one day, then borrow a solar luminosity meter and take the full day measurements by hand. Calibrate using Charts Estimate for a year, at a location with few horizon obstacles, then compare to similar charts from NOAA, NASA, etc.

15. CDR 2/24 Sensor Module PCB Rev 1 Schematic, Layout, and Ordered Linux Functional on SBC. Image Capture/Processing on Debian Machine functional. Algorithms completed in MATLAB Milestone 1 3/17 Sensor Module powered and with component interface code written Image capture from single board computer Basic user interface on windows computer, algorithms converted to C# Milestone 2 4/14 Sensor module software complete Sensor module fully integrated with single board computer User interface completed Expo 4/30 Full system test completed

16. Budget Fisheye lens$114.99 IP Camera$250.00 Single Board Computer$337.00 PCB's (2 with stencils)$250.00 Compass$143.95 GPS$59.95 Misc/Other IC's$35.00 LCD$17.99 Keypad$25.00 Total $1,233.88

17. Risks and Solutions Have started work early, and results look good Modular testing, starting with windows, then Linux, then SBC Have chosen an IP camera for OS compatibility Cant get camera to talk to Embed Linux computer. If sensors are not accurate enough, could throw results off enough to make whole device useless Can use redundant measurements and average results, either by sampling single device multiple times, or by using multiple devices. Tilt/Compass/GPS not sensitive enough to give results that we need Scheduling time and budget for a second board revision Full Team review before ordering boards. PCB layout mistakes or board population mistakes

18. Going Further Graphical image preview on unit. Built in rechargeable battery power pack. Remote camera trigger Account for single/dual axis sun tracking Always more variables to account for. work these into GUI to give better estimates

19. Questions?

20. Project Schedule