| GISdevelopment.net ---> GITA 2003 ---> Network Operations Management |
Outage management system helps keep the lights on in east texas
Keith Staples
Manager of Information Systems
Sam Houston Electric Cooperative, Inc.
Shane A. Gebert
Project Manager M.J. Harden Associates
Keith Staples
Manager of Information Systems
Sam Houston Electric Cooperative, Inc.
Shane A. Gebert
Project Manager M.J. Harden Associates
Abstract
This presentation will focus on all aspects of successfully setting up an outage management system. This will include planning, data collection/conversion, quality control and customization.
The planning aspect will focus designing an appropriate model to store the data and how the legacy and GPS collected data would fit into the model. The data collection/conversion aspect will summarize what procedures were put in place to insert data into the model. The quality control aspect will describe the different steps that were put in place to create data that was as accurate as possible. Finally the customization aspect will point out all of the changes that needed to occur, whether it’s the OMS software or the model itself, for a successful completion of this project. Along with showing all the successful parts of our project, problems will be pointed out. This will include incorrect planning, software incapabilities, unforeseen pitfalls and user errors. This presentation will give the audience invaluable knowledge of what steps worked and what steps didn’t work
Company Backgrounds
Sam Houston Electric Cooperative (SHECO) ranks as one of the largest of the 67 electric cooperatives in Texas. They now maintain more than 5,780 miles of line and serve over 48,000 members. Headquartered in Livingston, their area encompasses 10 counties in East Texas. Sam Houston Electric Cooperative, Inc. was founded in Polk County on August 27, 1938 and received its state charter on May 16, 1939.
M. J. Harden Associates (MJH) is a geographic information technology firm providing solutions for geospatial data management and analysis. Located in Kansas City, Missouri, M. J. Harden Associates provides GIS solutions for many clients located throughout the country. MJH was founded in 1956 and began pioneering the use of photogrammetry. In the mid-80s, services expanded into GIS and IT and later grew into geospatial software and internet development.
Project History
In an attempt to be more efficient in responding to electrical outages, SHECO decided to implement an Outage Management System(OMS). This system would pinpoint the problem using customer locations that have an outage. The locations are derived from customers calling in during an electrical outage. The phone number is matched against a customer service database that is related to a location on a map. Once the OMS specifies where the problem exists, dispatch crews are sent to the location of the problem. To arrive at this solution, an accurate GIS database of their electrical system needed to be created.
Creating an outage management system
Planning Phase
At the beginning of this job, SHECO had all of their existing GIS data in ArcFM 7 coverage format. The first thing to do was find a geodatabase model that met their needs and then convert the legacy data they had to this new geodatabase model. A software developing and consulting company stepped in and provided the necessary services to get this off the ground. After an appropriate model was designed, we were ready to implement a conversion plan to create a GIS database the accurately represented SHECO’s electrical system. The diagram below shows the approved conversion plan. The ArcFM 7 data is migrated into the ArcFM 8 geodatabase using customized conversion software. From here the data is exported out to shapefiles so the survey crews can load the data into the field data collection units. The data is then brought out into the field where it verified. This is also the point at which missing data is collected. The missing data is collected to their own shapefiles. After a feeder is completed in the field, the data is processed in the office. Once the data was approved by SHECO, it was sent to MJH where it would be put through more quality control checks and eventually inserted back into the database.
Data residing in a geodatabase cannot be used in an OMS. The data needs to be converted to format that can be used by the OMS. MJH wrote the software to perform this conversion. Not only does the software convert the data, but it also creates an error file that reports any problems that might exist in the data. Each feeder will have to be converted, edited for any errors and reconverted before the data can be inserted into the OMS.
Production Phase
Due to a large backlog of work orders and the need to have a service to transformer connection, a decision was made to perform a full field survey of SHECO’s entire system. To save time in the survey work and maintain important attribute information, existing data that resided in the geodatabase would be exported out to the GPS units and taken out into the field for verification. Upon approving a means of creating accurate data, we began the survey data collection phase of the project. Immediately, we ran across some hardware performance issues. The intent was to load all of the data for one feeder into a GPS device to take out into the field for data collection and verification. Using the current devices, we could only get about 25% of the data for an average feeder on the devices. This required us to either split up the data within a feeder or upgrade the GPS device to something with larger processing and memory capabilities. An attempt to split up the data proved to be very time consuming and confusing to the survey crews, so the hardware was upgraded to allow more data to be stored.
After a few days to get the survey procedures set, we were ready to start a large field survey effort. Approximately 15, one man, survey crews were assembled. Each one provided with the necessary equipment to efficiently collect and verify the field data. The area that we were dealing with is mostly rural and forested. This created some unusual situations. Each survey crew was equipped with an ATV. This allowed for easy passage on questionable roads and also allowed the surveyor to collect and verify much of the data without ever having to get out and walk. Another unusual situation was the large number of aggressive dogs that surveyors had to deal with. For this, each person was given dog repellant spray. Other problems that were encountered included locked gates and unhappy residents. Eventually all these issues were overcome and the data made it to the quality control phase of this job. Once the data made it into the office, each feature shapefile was thoroughly examined. If any data was missing, incomplete or inaccurate, the surveyor would be sent back into the field to recollect it. Once SHECO was satisfied with the data, it would be sent to MJH for more quality control and eventually be loaded back into the database. The data would be looked at three times before it would be loaded into the database. A project manager that oversaw the day to day operations of the survey crews would make sure the data was complete and accurate. Then the data would be sent to SHECO where a GIS analyst would open up the attribute tables and make sure that everything was valid. SHECO would also randomly check some of the data in the field to ensure that the survey crews were collecting the data correctly. Once SHECO was through with their checks, it would be sent to a project manager at MJH. There the data would be put through a series of checks that included connectivity, attribute and visual accuracy. If errors were encountered that couldn’t be resolved over e-mail or the telephone, the data would then be sent back to the project manager of the survey crew for corrections.
After several feeders were done with the collection, verification and quality control checks, a geodatabase load would be performed by MJH. The decision was made to input around 10 to 20 feeders at a time. The time it takes to perform the geodatabase load is usually the same whether you did one feeder or twenty feeders at a time. So a large number of feeders would be prepared and then loaded in at one time instead of loading each one individually. The process was to first load all the shapefile data to a personal geodatabase, create the units tables with a bank tag relationship then load the personal geodatabase and unit tables to the SDE geodatabase.
OMS Insertion Phase
The data that has been collected, verified and been through endless quality control work still is not ready for the OMS. Software written by MJH needs to be run on each feeder to perform the conversion and error reporting. The error file that is created will find problems that will prevent the OMS from working properly. Incorrect phasing and disconnected features would be common errors that show up in the error file. All valid errors need to be corrected before the converted data can be inserted into OMS. This phase of the project was performed by SHECO. The error files contained x and y coordinates that allowed the user to find the exact location of the problem. This process was further simplified by creating an error shapefile from the coordinates. The shapefile could then be used to locate and fix the problems. Once all of the errors are corrected for a feeder, it can be loaded into the OMS.
Conclusion
This project proved to be challenging for all of those involved. The success hinged on planning and tireless efforts by all. Many steps were taken to go from an idea to an operational OMS. During the way, many problems were encountered and eventually solved. This was a learning experience for all of those who took part in this large undertaking. The result of this effort is an effective tool to deal with electrical outages, which will not only result in a more efficient electrical cooperative, but more satisfied customers.