SLIDE 1
PIXELS
kyle bradbury, phd from
KNOWLEDGE
to extracting insights from energy data through visualization
ENERGY DATA ANALYTICS LAB
PIXELS to KNOWLEDGE extracting insights from energy data through - - PowerPoint PPT Presentation
PIXELS to KNOWLEDGE extracting insights from energy data through - - PowerPoint PPT Presentation
from PIXELS to KNOWLEDGE extracting insights from energy data through visualization kyle bradbury, phd ENERGY DATA ANALYTICS LAB milestones in energy through VISUALIZATION Early Power Plants Technical Drawings 1869 Source: Babcock
SLIDE 2
SLIDE 3
VISUALIZATION
milestones in energy through
SLIDE 4
Early Power Plants
Technical Drawings
1869
Source: Babcock & Wilcox Company. Steam, its generation and use. Babcock & Wilcox., 1922.
SLIDE 5
Thermal Energy Efficiency, 1898
Sankey Diagrams
Source: https://en.wikipedia.org/wiki/Sankey_diagram#/media/File:JIE_Sankey_V5_Fig1.png SLIDE 6
U.S. Energy Use, 2014
Sankey Diagram
SLIDE 7
Source: http://www.firstgreen.co/2013/08/graph-
- f-the-day-map-of-u-s-electricity-consumption-in-
1921/
Electrification of the U.S., 1921
Cartogram
SLIDE 8
Oil Trade 1969
flow diagram
SLIDE 9
Source: Hubbert’s Peak, from M. King Hubbert, “Nuclear Energy and the Fossil Fuels,” presented at a meeting of the American Petroleum Institute, 1956.
Hubbert Curve 1956
line plot
SLIDE 10
Actual Production Hubbert’s Predicton
Hubbert Curve 2000
line plot
Source: https://en.wikipedia.org/wiki/Hubbert_peak_theory#/media/File:Hubbert_Upper-Bound_Peak_1956.png SLIDE 11
Actual Production Hubbert’s Predicton
Hubbert Curve 2014
line plot
Source: https://en.wikipedia.org/wiki/Hubbert_peak_theory#/media/File:Hubbert_Upper-Bound_Peak_1956.png SLIDE 12
Atmospheric CO2 2006
line plot
Source: NOAA. http://climate.nasa.gov/vital-signs/carbon-dioxide/
SLIDE 13
Atmospheric CO2, 2006
line plot
Left: https://filmefuerdieerde.org/en/films/climate/an-inconvenient-truth Right: http://www.moviesteve.com/wp-content/uploads/2013/09/inconvenient_truth1.jpg
SLIDE 14 Source: U.S. Energy Information Administration – Annual Energy Review 2009
Petroleum Hydroelectric Wood Nuclear Coal Natural Gas
History of U.S. Energy Consumption, 2009
line plot
SLIDE 15
U.S. ELECTRICITY
generation
SLIDE 16
Energy Information Administration (EIA)
http://www.eia.gov/state/maps.cfm?src=home-f3
SLIDE 17
LINK
Capacity
SLIDE 18
Generation
SLIDE 19
CO2 Emissions
SLIDE 20
Generator Age
SLIDE 21
LINK
U.S. Generation
SLIDE 22
REMOTE SENSING
energy resource assessment from
data
Jordan Malof Rui Hou Leslie Collins SSPACISS Laboratory Kyle Bradbury Richard Newell Energy Initiative
SLIDE 23
Oahu, Hawaii New Solar Arrays after 2008
LINK
Kyle Bradbury & Mengyang Lin
SLIDE 24
Oahu, Hawaii New Solar Arrays after 2008
Kyle Bradbury & Mengyang Lin
SLIDE 25
Oahu, Hawaii New Solar Arrays after 2008
Kyle Bradbury & Mengyang Lin
SLIDE 26
Oahu, Hawaii New Solar Arrays after 2008
Kyle Bradbury & Mengyang Lin
SLIDE 27
the problem
Interest exists in quantifying U.S. distributed solar power capacity Capacity estimates are difficult to obtain
– Large-scale audits are conducted with questionnaires
Theoretical illustration of solar panel capacity by region
more less Malof, J. M., R. Hou, L. M. Collins, K. Bradbury, and R. Newell, “Automatic solar photovoltaic panel detection in satellite imagery,” in 2015 International Conference on Renewable Energy Research and Applications (ICRERA), 2015, pp. 1428–1431. SLIDE 28
machine learning solution
High resolution satellite images are increasingly available Algorithms may automatically estimate power capacity from images
Example e of
- f imag
agery ery data ta Solar r arra ray
Malof, J. M., R. Hou, L. M. Collins, K. Bradbury, and R. Newell, “Automatic solar photovoltaic panel detection in satellite imagery,” in 2015 International Conference on Renewable Energy Research and Applications (ICRERA), 2015, pp. 1428–1431. SLIDE 29
detection algorithm development
100 images were extracted from US Geological Survey satellite imagery (right) ML algorithms were developed to automatically locate the solar panels
50 without solar panels 50 with solar panels (red)
… …
Our development dataset of 100 house images
Malof, J. M., R. Hou, L. M. Collins, K. Bradbury, and R. Newell, “Automatic solar photovoltaic panel detection in satellite imagery,” in 2015 International Conference on Renewable Energy Research and Applications (ICRERA), 2015, pp. 1428–1431. SLIDE 30
A snapshot of the algorithm
Malof, J. M., R. Hou, L. M. Collins, K. Bradbury, and R. Newell, “Automatic solar photovoltaic panel detection in satellite imagery,” in 2015 International Conference on Renewable Energy Research and Applications (ICRERA), 2015, pp. 1428–1431. SLIDE 31
92% of panels were identified, with 4 total false alarms
solar array detection algorithm performance curve
Malof, J. M., R. Hou, L. M. Collins, K. Bradbury, and R. Newell, “Automatic solar photovoltaic panel detection in satellite imagery,” in 2015 International Conference on Renewable Energy Research and Applications (ICRERA), 2015, pp. 1428–1431. SLIDE 32
visualizing results – houses with panels
Black polygons are labeled solar arrays Yellow ellipses are detected regions
Malof, J. M., R. Hou, L. M. Collins, K. Bradbury, and R. Newell, “Automatic solar photovoltaic panel detection in satellite imagery,” in 2015 International Conference on Renewable Energy Research and Applications (ICRERA), 2015, pp. 1428–1431. SLIDE 33
visualizing results – houses without panels
+
Black polygons are labeled solar arrays Yellow ellipses are detected regions
Malof, J. M., R. Hou, L. M. Collins, K. Bradbury, and R. Newell, “Automatic solar photovoltaic panel detection in satellite imagery,” in 2015 International Conference on Renewable Energy Research and Applications (ICRERA), 2015, pp. 1428–1431. SLIDE 34
Data+ team created a ground truth data set
- f over 19,000 solar
array locations
SLIDE 35
ENERGY
data analytics lab