Data Con LA 2019 - Location Analytics For Smart Grid Reliability by Vivian Sultan
1 like1,236 views
This document discusses using location analytics to enhance smart grid reliability research. It aims to advance reliability through tools that analyze data related to geographic position. The research focuses on distribution systems, which account for most customer outages. Location data on outages, assets, and weather will be analyzed in GIS to identify outage hotspots and inform vegetation, equipment, and storm management. Methods include loading data into ArcGIS, cleaning layers, analyzing patterns, and identifying emerging hotspots over time. Prescriptive challenges are outlined around targeting storm preparation, vegetation trimming, and infrastructure inspection based on the spatial analyses.
Data Con LA 2019 - Location Analytics For Smart Grid Reliability by Vivian Sultan
- 1. Vivian Sultan, PhD DATA CON LA August 17, 2019 Location Analytics For Smart Grid Reliability A Spatially Enhanced Analytical Model For Power Outages
- 2. Goal And Research Direction Grid reliability research aims to address challenges and remove barriers to integrating high penetrations of distributed energy generation at the transmission and distribution levels (Office of Energy Efficiency & Renewable Energy, 2017). The objective is to advance Smart Grid reliability through the use of location analytics - a class of tools for seizing, storing, analyzing, and demonstrating data in relation to its position on the Earth’s surface. - GIS fostered a new approach to forecasting and data analytics. - GIS applications include recognizing site locations, mapping topographies and also developing analytical models to forecast events. - GIS is not limited to any specific field, only restricted by the availability of geospatial data. This research is concerned with Smart Grid reliability, specifically the reliability of the distribution system. - Distribution systems account for up to 90 % of all customer reliability problems. Main research question “How may location analytics be used to enhance Smart Grid reliability research?” - What current knowledge exists related to smart-grid reliability? - How may location analytics enhance the understanding of power outages and be used to improve the reliability of the smart grid?
- 4. Data Selection and Acquisition EPRI Data Mining Initiative The data sets include data from advanced metering systems, supervisory control and data acquisition (SCADA) systems, geospatial information systems (GIS), outage management systems (OMS), distribution management systems (DMS), asset management systems, work management systems, customer information systems, and intelligent electronic device databases Weather Data Georgia Spatial Data Infrastructure (GaSDI) and the Georgia GIS Clearinghouse is the data source for the monthly temperature and precipitation data The National Oceanic and Atmospheric administration website (NOAA) is the data source for the storm events and storm details
- 5. Methodology Step 1: Loaded data files from EPRI’s Data Repository along with weather data to ArcGIS. – Created a folder (geodata set) and imported the data files and basemaps (counties, tracks, roads, etc.) into the geodata set – Imported weather shapefiles into the geodatabase Step 2: Ran initial power outage events data exploration analyses in excel and GeoDa software. Step 3: Merged and related various data files in ArcGIS. – Merged outage events layers into one combined layer and linked to customers called and customers interrupted data layers – Related the forestry data and the Asset Management data with the combined events layer Step 4: Changed the projection of all maps to WGS 1984 projection system. Step 5: Cleaned the outage events map layer. – Started with 80,839 total records in the outage events map layer attribute table – Ended with 76,848 total records Step 6: Defined and created a study area for throughout project.
- 6. Methodology Cont’d… Step 7: Created a separate dummy variable for each cause of outage and Joined tables Step 8: Created new map layer for tree caused events by selection from the combined events layer – Wrote a query to select all the events under cause (Wind/Tree, Limb on Line, Tree Fell on line, Tree Grew Into Line, Vines) – Exported data into the geodatabase – Named new map layer “Right Of Way Outage Events” Step 9: Repeated the previous step to create additional map layers for weather related outage events, equipment failure, and System overload events. – Weather related outage events (events under cause category Wind/Tree, Wind, Ice, Major Storm, Lightning) – Equipment failure (events under cause category Failed in Service, Deterioration) – System overload (events under cause category Thermal overload, Overload, Load shed)
- 7. Methodology Cont’d… Step 10: Used the average nearest neighbor tool to find the average distance between outage events and if events are likely to cluster in certain areas Step 11: Calculated transformers age and joined to the transformer table in ArcGIS Step 12: Used the Convert time field / data management tool in ArcMap to convert outage event time to day of year – Repeated the same step for the storm events on storm details map layer Step 13: Using ArcMap ModelBuilder tool, three models were designed to spatially join the 48 map layers of weather data with the outage map layer – Model 1 to spatially join the outage events with the weather data – Model 2 to rename the output field (contour field) from model 1 – Model 3 to join the outage events data with the 48 fields of weather data
- 8. Methodology Cont’d… Step 14: Merged and related additional data files in ArcGIS – Added four additional columns to the outage map attribute table to show the weather data for each outage event – Joined by date the storm events with the outage events – Joined the storm events details with the outage events – Joined the outage events with the forestry file – Added a field “Adjusted_TransfAge” and a field “Adjusted_PoleAge” - Used Field Calculator to calculate the difference between the outage event year and the year the equipment was installed or modified – Added columns to show “Forestry Expected Pruning Man Hours”, “Average Climbing Tree Pruning Miles”, “Actual Pruning Man Hours/Circuit Mileage” Step 15: Conducted exploration and correlation analysis In SPSS Prior to statistical analyses, the following steps were taken to prepare the data: – For variables forestry expected pruning man hours, average climbing tree pruning miles, and actual pruning man hours / circuit mile , a value of zero (0) was input for missing data – Values for transformer age was substituted for missing data on pole age
- 9. Methodology Cont’d… Step 16: Ran Optimized hotspot analysis In ArcGIS – When the Input Feature is power outage events data and you do not identify an Analysis Field, the tool will aggregate the power outage events and the outage events counts will serve as the values to be analyzed. - one level of analysis – Another level of analysis is when you provide an Analysis field Step 17: Ran Emerging Hot Spot Analysis In ArcGIS - Two Steps processes – Create Space Time Cube By Aggregating Points – Run the Emerging Hot Spot analysis
- 12. Analyses and Finding Reported Power Outage Events Percent Count by Cause Reported Power Outage Duration by Cause
- 13. Analyses and Finding Cont’d… Inadequate data for analysis and many null fields – Asset Management folder showed inspection data for only two types of equipment – “Last Date Installed” and “original date installed” fields for equipment were mostly null values Not all files in the data set appeared useful considering the scope of this project work – Jets data file is about the field jobs – Circuit “Load” data do not include longitude/latitude data – “Load” data appeared to be overall feeder data
- 14. Analyses and Finding Cont’d… Descriptive Statistics
- 15. Analyses and Finding Cont’d… Correlation Results
- 16. Analyses and Finding Cont’d… Spatial Pattern Analysis in ArcGIS
- 17. Analyses and Finding Cont’d… Optimized Hot Spot Analysis Level 1 - The aggregation process resulted in 1296 weighted polygons Incident Count Properties Min: 1.0000 Max: 598.0000 Mean: 59.2955 Std. Dev.: 81.2320
- 18. Analyses and Finding Cont’d… Optimized Hot Spot Analysis Level 2
- 19. Analyses and Finding Cont’d… Optimized Hot Spot Analysis Level 3 Input Features: Weather Related Outage Events 2013 -2015
- 20. Analyses and Finding Cont’d… Optimized Hot Spot Analysis Level 3 Input Features: Equipment Failure Outage Events 2013 -2015
- 21. Analyses and Finding Cont’d… Optimized Hot Spot Analysis Level 3 Input Features: Right Of Way (Trees Related) Outage Events
- 22. Analyses and Finding Cont’d… Optimized Hot Spot Analysis Level 4 - Pole Age Analysis Map Output
- 23. Analyses and Finding Cont’d… Emerging Hot Spot Analysis Level 1 Time step interval 1 Month Number of space-time bins analyzed 41472
- 24. Analyses and Finding Cont’d… Emerging Hot Spot Analysis Level 2
- 25. Analyses and Finding Cont’d… Emerging Hot Spot Analysis Results Right of way (Trees Related) outages has the highest number of locations with hot trends (259 total count of locations) – Include the 40 consecutive locations with a single uninterrupted run of statistically significant hot spots - The utility company can use this information to reduce the risk of wildfire and keep customers safe Weather Related outages (160 locations with hot trends ) – Considering the availability of weather forecasts, this analysis can help a utility firm prepare should a storm is anticipated Equipment Failure outage (129 locations with hot trends ) System Overload (27 count of locations with hot trends)
- 26. Prescriptive Research Challenge #1 / Storm Scenario Link The National Weather Service issued a Red Flag Warning for the region, cautioning of extreme risk of a storm. The challenge that the utility is trying to answer is “Where should we preposition workers, and equipment in preparation of storm?” Challenge #2 / Vegetation ScenarioLink The grid has so many poles and wires that are vulnerable to falling trees and flying debris. The challenge is “Where should a utility improve tree cutting and trimming-related initiatives to foster operational excellence and reduce the risk of vegetation coming into contact with power lines?” Challenge #3 / Aging Infrastructure Scenario Link Considering the utility goal to reduce labor and cost of Inspection contractors, the research question in this case is “ Which infrastructure should be inspected to reduce the risk of power outage?”
- 27. Prescriptive Research Cont’d… Solution / Artifact GIS-based solution in Insights for ArcGIS. A web-based, data analytics application with the capability to work with both interactive maps and charts at the same time So easy to use, everyone at the electric utility, from the personnel in the field to the chairman of the board, can take advantage of its capabilities Capability to record workflows, utility personnel can rerun this analysis whenever Inspection budget becomes available or whenever a storm is expected to hit the service area All relevant data is imported from previous analysis sourced from the SCADA/OMS/DMS systems at a power utility into Insights for ArcGIS The idea is to connect to virtually any type of streaming data feed and transform the GIS applications into frontline decision apps, showing power outages incidents as they occur
- 28. GIS offers a solution to analyze the electric grid distribution system
- 29. Together…Shaping the Future of Electricity
Editor's Notes
- #5: `