Data Conversion Fast Track at Allegheny Power

Data Conversion Fast Track at Allegheny Power

By James D. Black, Black & Gorman L.L.C.

Allegheny Power is a conglomeration of Monongahela Power, Potomac Edison and West Penn Power companies which overlaps parts of Pennsylvania, Ohio, West Virginia, Virginia and Maryland. The utility serves 1.4 million customers in a mostly contiguous territory of about 29,000 square miles. Recently, Allegheny Power and Duquesne Light Co. agreed to merge. Duquesne Light serves more than 580,000 customers in a service territory of about 800 square miles, including the city of Pittsburgh. When approved by the state public utilities commission, the combined companies, called Allegheny Energy, will serve nearly 2 million customers and have a combined generating capacity of 11,000 MW.

While a merger such as this would put pressure on any utility to speed up the process of managing its data assets and outside plant facilities, Allegheny Power must also keep a close eye on recent state legislation aimed at restructuring electric utilities for retail customer choice. Gov. Tom Ridge signed The Electricity Generation Customer Choice and Competition Act in late 1996, making Pennsylvania the fourth state to voluntarily comply with federal mandates for industry deregulation by the year 1999. Suddenly, customer retention is a key issue.

Racing to stay ahead of these increasing market forces, the Operations Dept. of Allegheny Power, under the direction of vice president of operations Karl Pfirrmann, is renovating and expanding its information technology (IT) systems beginning with the AM/FM system. Carl Livingood, project lead for AM/FM, under the direction of Dana Rogers, planning and AM/FM; and Jim Haney, director of operations services; is handling the day-to-day coordination of these efforts. As project lead, Livingood`s responsibilities include helping his team chart the future of IT at Allegheny Power.

Livingood sees AM/FM as the bedrock supporting the work management system, outage analysis system, engineering analysis and other vital systems which will eventually link to business systems beyond engineering and operations. “AM/FM is a system that you build other systems on top of,” he explained. “The kinds of things you can get out of work management and outage analysis systems are not possible without a solid foundation of AM/FM in place.” Livingood said his department aims to impact business decisions and make a substantial contribution to customer service.

If the AM/FM system supports the enterprise, then its the data that supports the AM/FM system and all other IT applications within the organization. Rapid, accurate data conversion is essential to the goals of Livingood`s team at Allegheny Power. One of the first and most critical tasks Livingood undertook was to select a data conversion methodology to reduce project implementation time to less than two years.

“We have two main goals: The first is to put an AM/FM system in place which will support outage analysis and work management. The second goal is to convert all legacy data over to the target data format by the end of 1997,” Livingood said. “This project in its current form was started in April 1996. So that`s pretty aggressive to get all that data converted in less than 24 months.”

Livingood`s team stresses that the data asset created must be immune from any subsequent changes in platform. By converting the legacy data to a neutral format, the team was in a position to take advantage of emerging technology. “To a certain degree you can separate the conversion process from the target platform,” he said. “If you take care to convert your data to a neutral format, you can take it to any platform.”He cites for example that Allegheny Power began converting the data to a previously selected platform, but near the end of that initial conversion effort the team decided it was in its best interest to convert to a new target platform. “We asked Apex Data Services, our conversion vendor on the West Penn project, to retool for our new platform.”

Concurrent Conversion

Apex was selected in part because it had adopted a different approach to data conversion–an approach that would make up for lost time in migrating from one platform to another. In the past, Livingood explained, data conversion could not begin until data definitions and a physical data model specific to the new AM/FM system were available. This sequential linkage between the development of a specific technology and conversion forced a protraction to the overall project schedule.

Having evolved from the perspective of “graphics-first” data maintenance procedures, traditional conversion methods follow a prescribed sequence of tasks:

1. Create the graphics either manually or with automated pattern recognition techniques.

2. Enter attribute data either by keystroke or from pick lists.

3. Associate the attributes to the graphics.

4. Validate the data for network connectivity and logical relationships.

5. Plot the graphic and attribute data on paper.

6. Manually check the plot for errors.

7. Rework the process to correct the errors.

8. Repeat as necessary until the specified quality is attained.

The burden of manual digitizing and the cost of rework cycles demanded that the data be compiled from the beginning in strict conformance with the data structures of the target system. The process did not have the flexibility to allow a recompilation of the data to an alternative format after the fact. Allegheny Power`s West Penn project time lines were too stringent for traditional conversion (see figure).

As a “data-first” approach to conversion, the Apex DATAWORKS conversion process is arguably completely different:

1. Capture all data alphanumerically including attributes, graphics and the relationships defining connectivity.

2. Capture the data twice and use the computer to trap errors at the point of entry.

3. Compile the data into a generic data model.

4. Validate the data model for network connectivity and logical relationships.

5. Generate graphics automatically.

6. Edit the graphic presentation where necessary, particularly in congested areas.

7. Recompile the data model in conformance with the data definitions of the target system.

By capturing the data in a generic logical data model, conversion can begin early in the project. The data can be converted concurrently with the selection and development of a new system, what Apex calls “Concurrent Conversion.” So far, the results at Allegheny Power indicate that this process can accelerate project implementation by months, if not a year or more.

Data Prioritizing

Livingood`s team made decisions about how much information is needed early in the process. Recognizing there is a certain level needed by engineers, a different level needed by accounting and also a “super-engineer level” needed for overall project management, it determined that there are mappable and non-mappable items, defining mappable items as poles, transformers, primary conductors, regulators, reclosers and all the in-line devices–anything that has connectivity. Non-mappable items are cross arms, anchor guys, insulators and other support devices. “This was an enormous amount of detail not needed by an engineer,” he said. “We`ve made some tough decisions to track only items that support the outage analysis, engineering analysis, maintenance systems or work management systems.”

Source Data Preparations

One way to fast track data conversion is by proactively eliminating source document errors and updating map annotation before converting the data. The West Penn project data is being “scrubbed” and annotated by a group of about eight people at Allegheny Power`s Jeannette Service Center near its Cabin Hill headquarters in Greensburg, Pa. The inventory includes about 3,500 paper maps representing nearly 10,000 square miles of land base and facilities. The maps are of two major types: 500 feet/inch distribution maps that cover the entire distribution territory representing 10,000 square miles and 100 feet/inch maps for areas that are too dense to see all the information clearly. The maps represent pole locations, primary and sub-transmission conductors, transformers, protective devices and substation locations, as well as all the land base.

As with many utilities, Allegheny Power records reflect many different engineering and drafting practices. These variations stem from the consolidation of former independent operating companies and the personal drafting styles of generations of record keepers. Keeping this hand scrub work to a minimum is another way to save time and speed up the data conversion process. One way of achieving this is to use the computer to update old drafting conventions. To illustrate an example, Livingood points to a map from the 1930s. “Apex uses people to interpret the source documents and computers to build consistency in the data.” He explains that conversion workers capture the data literally as they see it, without taking valuable time updating 67-year-old symbols by hand. The computer, preprogrammed with a “Rosetta Stone” translator, replaces the out-moded attributes with the current standard.

People vs. Machines

One common misconception about modern data conversion practice that dies hard is that automation is always better than hand work. For example, much effort has been made in recent years to automate data conversion with pattern recognition techniques. Yet there has been no effort to automate quality control.

However, automation is no match for the intuitive power of the human mind to interpret old distribution maps. “For instance, we`ve set up rules with Apex so that when its staff sees certain features, it can interpret on the spot rather than use a semi-automatic process which cannot make these assessments.” Livingood goes on to say that when map annotation is very close to several attributes, such as a pole with multiple transformers, a person can readily intuit which text goes with which feature where a machine cannot. “Humans can instantly recognize the difference between the `noise` on map sheets, such as faint pencil marks vs. actual light pencil annotation supplementing a feature. I think humans can also determine connectivity better than machines.”

Quality control, on the other hand, is ideally suited for automation. “Strangely enough, we`ve come almost full circle from the conversion methodologies used years ago. Then, when you wanted very good accuracy, you made a double pass of your data and then compared the two collections. That`s what Apex`s methodology includes,” said Livingood. By capturing the data twice and using the computer to trap errors at the point of entry, and by validating the data model before creating graphics, Apex avoids the expense of manual check and rework cycles necessary with traditional conversion processes. This more than offsets the cost to capture the data twice, thereby lowering the overall cost of conversion. Moreover, automated quality control dramatically increases the accuracy of the data.

Apex`s process is based upon the Oracle relational database. This neutral relational database format facilitates the migration of legacy digital data and its merger with data captured from map records. This capability is becoming increasingly important to utilities as they integrate departmental data into corporate information infrastructures.

Interestingly, many leading AM/FM system vendors are porting their graphics applications to Oracle. Also, Oracle data is readily integrated with the relational databases typically favored by corporate information systems managers for management information systems and data warehouses.

Using the Data

Many users within Allegheny Power are benefiting from the river of data being created for them. Livingood explains that the operations section is organized into about 100 teams. Each team has a team leader, one or two designers, an office associate (usually shared with another team) and between six to 10 linemen. There are a total of 135 designers for which Allegheny Power purchased 86 licenses of the new target platform and an equal number of designer software tools. “Designers are our main customers for AM/FM design or update software,” Livingood said. The team leaders and office associates use mostly viewing software. In addition, the Customer Support Center has about 200 employees working shifts who will use the viewing software.

“Mapping in and of itself does not pay for the AM/FM system,” Livingood warns. “You need to build AM/FM into your business processes and that`s what we`ve tried to do.” He cites the example of preparing to take as-built information and semi-automatically update the AM/FM system with new annotation, symbol and connectivity information. “We can make pole placements and facility placements very fast through the AM/FM system and use the underlying intelligence to estimate customer loading and optimization of facilities to make sure the design we come up with is the best design for the job–it`s not overbuilt, it`s not underbuilt, but just right.”

Livingood also cites the benefit of providing AM/FM data to the analysis systems. “The main target initially is the outage management system,” he said. The customer service personnel who take the initial call from the customer reporting an outage will use information from the AM/FM system to determine the most probable cause of outages. The fast tracking of data conversion means achieving results sooner for utilities like Allegheny Power. Faced with the certainties of deregulation and competition, the organization did not stand still, but quickly focused on technologies and methodologies that would impact its business needs for today and tomorrow.

Author Bio

James D. Black is a founding partner in Black & Gorman L.L.C., a communications firm specializing in information technology. Black is a well-known industry writer in the field of GIS and is currently the education director for the A/E/C Systems International Conference and Exhibition. He can be reached at (e-mail).

Click here to enlarge image

Western Pennsylvania`s rolling hills and farm lands are served by Allegheny Power.

Previous articlePOWERGRID_INTERNATIONAL Volume 2 Issue 4
Next articlePOWERGRID_INTERNATIONAL Volume 2 Issue 5
The Clarion Energy Content Team is made up of editors from various publications, including POWERGRID International, Power Engineering, Renewable Energy World, Hydro Review, Smart Energy International, and Power Engineering International. Contact the content lead for this publication at

No posts to display