Managing the GPS/GIS Function in an Electric Utility
By Michael Michelsen, Jr.
Anew period of higher significance has arrived for the GPS/GIS function at electric utilities. To a degree never equaled before, utility managers are looking to their GIS programs, filled with increasingly accurate data collected by GPS technology, before making many decisions. With this capability comes an expectation for GIS/GPS professionals to provide higher levels of planning and management of their data collection process. At Duke Power, managers rely on GPS mapping systems to fill their data collection equipment needs.
Management by Objective
“In June 1997, Duke Power was asked by the city of Charlotte to provide a more complete accounting of services provided to the city for street lighting. Until that time we were only able to bill the city in a summary fashion. The city wanted a more complete accounting of the services provided, and we wanted to provide it,” said Rob Manning, Duke Power`s electric distribution engineering, construction and operations manager. “We saw this as a good thing because we consider this to be a trend nationwide. Fortunately, Duke Power was already studying implementation of a program of this type. The request we received from the city only caused us to speed up the implementation.”
When the city of Charlotte requested the more detailed billing system, Manning joined with Emmy Lou Burchette, Duke Power`s marketing vice president, retail services, to co-sponsor the street light inventory project, a comprehensive program implemented to fully account for street lighting facilities within the billing area. “This use of technology provides valuable benefits to both Duke and our customers,” said Burchette.
The Mission and the Method
“Duke Power was one of the original power companies in the country to map its electrical facilities in an electronic format,” said Brent Richardson, the project manager charged with implementing the street light inventory project. Duke originally estimated that there may be 62,000 street lights in the city of Charlotte, with as many as 11,500 being lights that should be privately billed. The final count revealed approximately 55,000 lights, slightly more than 3,000 of the lights were found to be billable to private customers rendering a final number of nearly 52,000 lights billable to the city. Providing this volume of lighting facilities is in addition to providing the energy needs of approximately two million customers across the service area. Duke Power knew that it had numerous types of different lights, but until now it was difficult to determine where some of those lights were, and what, in detail, the utility was billing customers.
In June 1997, the city of Charlotte requested more details regarding street lighting billed to the city by Duke Power. “In order to accomplish this we knew that we needed to take a physical inventory of the street lights, then tie it to our proprietary billing system, which would give each billable facility a point on the earth,” Richardson explained. “We wanted to be able to point to a given light and tell our customer how much they are paying for that light. Prior to this, many electric utilities have put their facilities in an electronic asset database, but none could specifically identify billable items from which they could generate a statement.”
None of this posed a major problem, Richardson thought, until the city requested that the new billing system be implemented in time for the city`s next budget cycle–March 1, 1998–only eight months away. “At the time we weren`t sure what we were agreeing to, but we went forward with the plan because some thought had already been given about this type of a program as an option,” Richardson explained. “Stokes White, a Duke Power teammate, had been studying options for such a project and had originally put together a two to three year time frame for implementation. This agreement significantly sped up the process.”
Richardson teamed with Dennis Morton, a Duke data manager, to recruit other Duke teammates as a first step to develop an inventory system that would give the city the information they needed, by the time they needed it. “We weren`t sure what we were dealing with, so this made us consider all kinds of options, for people as well as other resources,” Richardson said. “Our second step was to put together RFPs (request for proposals) for a number of vendors that might also be able to help us with their services, once we decided what product we actually wanted to get back.”
Richardson explained that one of the keys to the company`s new inventory program was the ability to display information graphically. Duke Power`s corporate GIS system is undergoing a conversion from IBM`s GFIS system to Smallworld. “We already had a GIS program implemented as part of our efforts to better map our electrical facilities, but we had not previously mapped our lights. The timeframe of this project would demand a GIS solution prior to the major conversion and we also needed to include a data capture method,” Richardson explained. “We selected MapInfo and were pleased with the GIS capabilities it brought to this application. Their system was basic and flexible enough, with good functionality to meet our needs.”
All of this was important to Richardson and Morton, who stressed the need to know the attributes of the company`s services–especially those attributes that were billable to the city. “Beyond these concepts, we really hadn`t given a lot of detailed thought to how we would capture those data items, with what tools, or how we would tie the eventual capturing of that information back to the billing system,” Richardson explained.
At the beginning of the program, Richardson and Morton both felt that although they possessed the management abilities they needed to implement the inventory system, they lacked the personnel as well as the full technical abilities to carry out the program–especially with less than a month to begin the program. “Data collection with a GPS seemed like the natural way to go with this program,” Richardson said. “Unfortunately, since electric distribution didn`t own any GPS units at the time, and the skills necessary to use the units weren`t consistent with our core business, we didn`t feel like we could do the job without hiring outside help in the timeframe required.”
Despite the need to contract for outside GPS data collection services, Richardson and Morton felt compelled to purchase their company`s own Trimble Pathfinder ProXR GPS units in order to audit the contractor`s work.
Enter 3001 Inc.
“When we were faced with the task of implementing this program, we had only one fellow who was dedicated part-time to GPS data collection,” said Morton. To make matters worse, Richardson and Morton were also working on other projects at the time, so in order to implement the program, both were assigned full-time to develop a data collection program within two weeks.
According to Richardson, the timeframe required by the city for implementation of the new billing process was the most strenuous objective of the program. “We felt from the very beginning that at best we had a project that would take as long as 18 months to complete,” he explained. “All of a sudden we had to have it fully implemented within six months. We knew that we had a monumental data collection job to do. The second most challenging feat of this job was pulling together the resources and the technologies, making them fit together, working with a contractor we had never worked with before, and managing the whole project.”
Duke Power hired 3001 Inc. on Sept. 11, 1997. To complete the data collection job for Duke Power, 3001 Inc. began on September 22. Project managers at 3001 Inc., implemented specially designed mountain bikes, which would travel more than 2,600 miles of city streets collecting up to 54 physical attributes about each of the estimated 62,000 utility light poles throughout the city. Richardson commented that the average data collection per pole consisted of 18 to 24 attributes such as the type of pole, the bracket length, the type of light and the type of lamp. Before beginning the project, Duke Power and 3001 Inc. established milestones based on the average collection of 800 to 1,000 lighting facilities per day to meet the original time commitment.
Correct identification of facilities by the data collection crews was a critical element in the success of the project. As a result, each crew member was given a two-day training class which taught them how to identify poles and associated lighting equipment prior to beginning the job. “We had to teach each crew member what they were seeing and how to translate all of this equipment into their data dictionaries which offer menu driven options for collection of specific data,” Richardson explained. “All of these elements and more affect the billing of the service, and each was accounted for by using the data dictionary on the ProXR.”
“Throughout the data collection project, one of our roles was to compile, review and audit the collection process,” Morton said. “Data was delivered in batches consisting of about 3,500 lights each. A batch was delivered on the average of once every two weeks. We reviewed each batch then audited the collected data (about five percent of each batch) using our own ProXR to ensure accuracy. These samples were not totally randomly selected. They were specifically identified from areas that we had questions about after we examined the data from the MapInfo GIS. For example, we might have audited an area that showed streets but showed no data or lights.”
Morton used the ProXR for the audit sample with no training and only limited reference to the equipment documentation. “I think the equipment was very intuitive to use. The Pathfinder Office software was very straight forward for manipulating the data,” said Morton. As each batch of data was collected, it was downloaded into the MapInfo GIS. “This is where we felt MapInfo did a particularly good job by allowing us to perform this downloading process in a batch mode, which was most convenient for us. It also allowed us to manage this project as we went along.”
The Lowdown on Standard Operating Procedure
“As best we can determine, we are the only utility company that has ever done this type of inventory with specific products such as lighting,” Richardson said. “Our contractor worked eight to ten hours a day, five days a week, and in many different types of conditions. Despite all of this, we always felt like we were getting good geographic positions on all of our data. Most of the technicians we dealt with told us that we should wait ten minutes to get a really good positional reading. The problem was that we didn`t have 10 minutes. We could only be at a given point for a maximum of 4 to 6 minutes to meet our milestones. Despite this, our audits never turned up more than a few positions with any significant error. We needed to be accurate because we were dealing with billable items, as well as our reputation with other municipal customers. This meant that we couldn`t allow ourselves to just accept the data as delivered. We had to ensure the data being collected was very accurate.”
Who`s Responsible for What?
Duke provides outdoor lighting under three rate schedules: public lighting, outdoor lighting for private customers and flood lighting. According to Richardson, a high-level analysis of the lighting facilities was called for after the data collection process was completed. “The analysis that we performed corrected other data collection errors as well,” Richardson explained. “During the inventory a small percentage of lights were identified as billable to the city, but were not; they looked like public lights, but they were privately owned or billable to private citizens other than the city. With our high-level analysis, we were able to identify the correct billable party for all of the lights collected. This results in accurate bills, meeting the city of Charlotte`s expectation for correct billing, based on an accurate inventory of facilities provided.”
A Common Database
“One of the key precepts that we kept in mind when carrying out this project was that we had to keep a common database,” Richardson explained. “As the database was created and viewable via GIS, we now had a common database from which we could create a bill as well as see the data. We can run a billing calculation routine against the database to generate a bill or use MapInfo to see a graphical picture. Prior to the creation of this database capability, the difference between the databases for a display tool and billing system was a potential source of discrepancy. The new system eliminated this discrepancy.”
More than Just a Bill
“Creating this database resulted in allowing us to do much more than to just create a bill for the city,” said Richardson. “This process allows us to work better with the city by improving our billing accountability, and allowing us to provide better service as well. For instance, in a given area there might be several 100-watt mercury vapor lights, which might not be the best application for lighting this particular area.
Now, we can identify this equipment and its location in the office and point to areas to upgrade this equipment to better serve our customer. The city can also use this data to correlate crime statistics and traffic accidents to lighting. That`s the value of the tool, and the value of how the data we`ve collected can be used. We`ve created a win-win situation: providing accurate billing to customers while maintaining a robust system that allows us to offer upgrades and options to our customers.”
Using a GPS unit to pinpoint the locations of street lights, Duke is able to accurately inventory the lights and bill the city of Charlotte or private customers.
Duke Power is able to identify every street light on its system by type of light and location.
Michael Michelsen, Jr., is a marketing writer for Trimble Navigation Ltd., Sunnyvale, Calif. Readers interested in obtaining more information about Duke Power`s GIS/GPS Asset Management Program should contact Brent Richardson, Duke Power, at 704-382-7435.