Presentation is loading. Please wait.

Presentation is loading. Please wait.

A Survey of Energy, Water, and Environment Complex Networks Present By: Eric Klukovich Date: 10/21/14.

Similar presentations


Presentation on theme: "A Survey of Energy, Water, and Environment Complex Networks Present By: Eric Klukovich Date: 10/21/14."— Presentation transcript:

1 A Survey of Energy, Water, and Environment Complex Networks Present By: Eric Klukovich Date: 10/21/14

2 Overview  Complex Networks  Nevada Solar Nexus  Studies in Energy Complex Networks  Studies in Water Complex Networks  Studies in Environmental Complex Networks  Conclusion

3 Complex Networks  Relatively new concept and is being actively researched to understand their full potential.  Complex networks are based on graph theory.  Model real-world data in a much more accurate way.  Analyzed from a different point of view.  Shows different trends and features within the data.  Can apply different metrics  Degree distribution  Closeness/betweenness  Clustering coefficients

4 Uses for Complex Networks  Can help solve real-world problems such as improving resiliency and robustness in a network.  Can model energy, water, and environmental data.  Used to see how reliable and efficient the current energy and water distribution systems are.  Can show the impact if a power outage occurred in different regions.  Rainfall, climate, and seismic data have been modeled to identify patterns in the topology and dynamics.

5 Nevada Solar Nexus  Creating renewable energy resources has become a national priority.  Nevada is focusing its research into this area  Currently three-quarters of all energy production is from fossil fuels.  Increased dependency on global markets  Creates greenhouse gases  Focus in these areas:  Solar Energy  Water  Environment

6 Solar Energy Goals  Create renewable solar energy generation in Nevada.  One of the best locations for solar energy generation in the world.  Has the potential to diversify the economy of the state.  Solar energy generation should have a small impact on water resources and the desert environment.  The relationship between solar, water, and environment should be understood for renewable energy to be beneficial.

7 Water Goals  Ideal locations for solar energy generation is in arid lands.  Water resources will be limited.  Need to maximize limited water use at the facilities.  Explore the use of lesser quality water in solar energy development.  Minimize the impact of moving water/wastewater to and from the facilities.  Extraction, treatment, distribution and disposal require energy and impacts the environment

8 Environment Goals  Study the impact of the solar facilities on the environment  Minimize construction, operation, and decommission impacts  Study the impacts on organism populations  Microclimate change in planet communities  Impact of solar arrays on the balance of desert soil  Impact on landscape patterns  Before, during, and after construction

9 Energy Complex Networks  Energy distribution affects large amounts of people on a regular basis  Electricity  Natural Gas  Oil  Transferring energy is usually done through wires or pipes in a grid configuration.  Can grow to be very complex and difficult to analyze.  Complex networks can find patterns and help solve problems in the current system.

10 Modeling the Power Grid  There have been many major blackouts in North America within the last few decades.  Difficult to determine what happened with the all the interconnections within the grid.  Can show where power grids are vulnerable to blackouts or outages.  A few papers have modeled the power grid and show how well the grid can function if generators shut down.

11 Power Distribution

12 Study 1 - Power Grid Vulnerability  Modeled the North American power grid using data in the POWERMap mapping system.  Modeled every major substation and 115 − 765 kV power lines.  14,099 nodes (substations)  19,657 edges (transmission lines)  Three main types of substations  Power generators  Transmission stations  Distribution stations

13 Study 1 - Power Grid Vulnerability  Power generation vulnerability was tested:  Nodes were removed by their degree and randomly.  It was found that the loss of connectivity when removing power generation nodes did not alter the overall connectivity of the grid.  There is a high level of redundancy at the generating subsystem level.

14 Study 1 - Power Grid Vulnerability  Transmission substation vulnerability was tested:  Removed nodes randomly, highest degree, highest load, and removing the top 10 highest loads.  When the nodes were selected randomly, then the loss of power was proportionate to the number of nodes lost.  The degree and load based removal showed a higher increase of connectivity loss.

15 Study 1 - Power Grid Vulnerability  Issues with the study  The data set only identified the generator nodes and the rest were identified based on criterion.  This may or may not accurately model the power grid and could lead to different results.  The authors assumed that each distribution station only had one transmission line going to it.  There could be more than one transmission line going to each distribution station, changing the degree distribution.  Therefore if it failed it could lead to a greater loss of power.

16 Study 2 - Power Grid Reliability  Modeled the North American power grid using data for the western and eastern power grids.  Western Electricity Coordinating Council (WECC)  North American Electric Reliability Council (NERC)  Western - 78,216 nodes  Eastern - 235,907 nodes.  Uses the Barabási-Albert Network Model to find quantify the grid’s resilience.  Data sets are very accurate.  Contains all data needed to accurately test resilience.

17 Study 2 - Power Grid Reliability  Uses loss of load probability to create a failure propagation model.  Power flows to node from an edge (propagation unlikely)  Power flows from the node to an edge (propagation likely)  Calculates the probably of removing edges or nodes.  The Eastern and Western power grids were scale free.  The loss of load probability for was found to be 0.026.  This value was compared to the loss of load probability of the Bonneville Power Administration’s region of the western grid (0.027).  The Barabási-Albert model accurately predicts the reliability.

18 Water Complex Networks  Water distribution is infrastructure that must always be available.  Can analyze the efficiency, vulnerability, and create plans for alternative resources.  Rivers can also be modeled  Monitor the water flow.  Take protective action if the river is being depleted.

19 Study 1 - Water Distribution Analysis  Modeled four different water distribution networks  East-Mersea, United Kingdom  Colorado Springs, Colorado  Kumasi Town, Ghana  Richmond, Virginia  Nodes represented source, control, and storage/processing facilities.  Edges represented by pipes.  The weight was the diameter of the pipes.

20 Study 1 - Water Distribution Analysis East-Mersea Colorado Springs RichmondKumasi

21 Study 1 - Water Distribution Analysis  Each network’s density was calculated  All networks were sparse and resemble the urban areas.  The Colorado Springs network had many loops.  The degree distribution and central point dominance for each network was calculated.  Determined which nodes were the most important.  Found that large clustering was in the town’s center.  The efficiency of the water distribution was measured  Topographic measurement for efficiency was not accurate.  Construction and cost has a major factor on how the network is created.

22 Study 1 - Water Distribution Analysis  Route factor is a better way to measure efficiency.  Distance between the supply node and the demand source  The network was found to be highly efficient in the four graphs.  The robustness was measured by random removal of nodes.  42% removal for Colorado Springs caused failure.  37% removal for Kumasi caused failure  32% removal for Richmond caused failure  22% removal for East-Mersea caused failure  An extreme event would make water distribution vulnerable.

23 Study 2 - Modeling River Networks  Modeled the Haihe Basin River network in China  565 nodes (319 natural and 246 engineered nodes)  Two types of nodes  Natural – source, bifurcations, confluence, and outlet.  Engineered - hydro power plants, reservoirs, pumping stations, and transfer plants.  Edges  Natural or artificial water channel that connected two nodes.  Directed – flow of the river.

24 Study 2 - Modeling River Networks River network River Node/Edge Example

25 Study 2 - Modeling River Networks  The degree distribution was calculated to categorize the different nodes.  The river’s sources and outlets could be easily determined.  The nodes that can be used to regulate the flow were also found.  This study acts as a foundation for more in depth studies in river networks.  Could find potential sources of drinking water.  Could model pollution spread in the water system.  Find the impacts on surrounding communities if the river dried up.

26 Environment Complex Networks  Complex networks can be used to analyze data and find new information and patterns.  There have been several studies in areas related to climate dynamics, rainfall and seismic activity.  Spatial grid points as nodes  The edges represent if an event occurred in both the linked nodes.

27 Study 1 - Modeling Earthquakes  Modeled earthquake data for Southern California.  Nodes – small cells that divide up the geographical region.  Edge – seismic activity that occurred different cells.  Loop – seismic activity that occurred in the same cell.  Found that the aftershocks of the earthquake tended to have a loop back to the original node.  Earthquake data is scale-free and is a small world network.  The connectivity distribution follows the power law.  The average path length is small and has a high clustering coefficient.

28 Study 1 - Modeling Earthquakes Degree Distribution Earthquake Network

29 Study 2 - Modeling Precipitation  Modeled extreme rainfall in different areas of the world.  South America  South Asia  Indian subcontinent  Entire globe  Nodes were based on small divisions of geographical locations.  Edges were created between two nodes if the amount of precipitation met a certain threshold.

30 Study 2 - Modeling Precipitation  A highly accurate network was created and many different aspects were observed.  The degree distribution accurately placed the most arid places at the nodes with the lowest degree.  Islands are disconnected in the network and they form their own micro-networks.  The larger land mass does not affect the islands.  Rainfall patterns have changed rapidly more recently compared to previous decades.

31 Study 2 - Modeling Precipitation Degree Centrality

32 Conclusion  Complex networks can be used to model realistic datasets in different domains.  Energy, water, and environmental data can be analyzed using complex network modeling and metrics.  Many studies have accurately modeled data in distribution system to determine how efficient, robust, or vulnerable the system is.

33 References  http://venturebeat.files.wordpress.com/2010/10/grid.jpg  http://regan.med.harvard.edu/pictures/Hier/3D_hier.jpg  “Structural Vulnerability of the North American Power Grid” http://arxiv.org/pdf/cond- mat/0401084.pdf.  “Evaluating North American Electric Grid Reliability Using the Barabási-Albert Network Model ” http://arxiv.org/ftp/nlin/papers/0408/0408052.pdf.  “Complex network analysis of water distribution systems” http://arxiv.org/pdf/1104.0121.pdf.  “Modelling and analysis of river networks based on complex networks theory”  “Small-world structure of earthquake network” http://arxiv.org/pdf/cond-mat/0308208.pdf

34 Questions


Download ppt "A Survey of Energy, Water, and Environment Complex Networks Present By: Eric Klukovich Date: 10/21/14."

Similar presentations


Ads by Google