In 2004, transformer fires raged through Maricopa county's westwing substation, resulting in tens of millions of dollars lost. To prevent such calamities and improve service reliability, Salt River Project (SRP; Phoenix, Arizona, U.S.) tackled the challenge of integrating dozens of sensors and intelligent electronic devices (IEDs) at its 500-kV/230-kV Browning Substation in early 2006.

Although Browning already had numerous sensors and IEDs producing large amounts of raw data, SRP personnel had no consistent means of getting useful information from them. Most of the devices had their own proprietary software. Some devices allowed remote communication, while others required site visits to download data. SRP needed to integrate the systems in a way that would convert data into useful information that would be securely available to SRP engineering and maintenance staff on their desktops.

NO HIGH-PRICED PATCHWORK

SRP staff recognized that attempts to modernize existing systems often result in failure. Frequently, these attempts involve developing or purchasing new technology and grafting it on to an existing system. But then it turns out the existing system was not designed to work with the new technology, so engineers must develop an integration system to connect the two. Unfortunately, the next time another change is needed, still another “one-off integration” has to be created to accommodate it. Each new addition or upgrade is integrated in an ad hoc manner, without an overall direction that will allow growth without constant re-engineering. As one part in an ad hoc system becomes obsolete, it may be difficult or impossible to come up with a custom patch to fill the gap or to interface the remaining pieces. Maintaining this patchwork quilt can become hugely expensive.

Although any competent engineer can find a way to glue any two systems together, it takes a little more work and thought to glue those systems in a way that is extensible, scalable, manageable and secure. Without a plan — an architecture — projects face continuing added expense, as each new system has to be redeveloped.

REQUIREMENTS FIRST

One common problem is failing to clearly define a system's requirements in advance. A requirements-based approach allows one to understand the information from the start and create policies to manage it. However, too many companies gloss over or skip this step entirely. The result is products that don't perform as desired (both practically and in the marketplace) and products that require costly reworking. Furthermore, the expense of reworking a product increases exponentially as later errors are discovered.

To be effective, requirements must be captured and communicated. Requirements are useless if only one person or group knows them. Someone must be responsible for implementing, testing and supporting them.

MODERNIZING THE PROCESS

SRP's first step in modernizing the Browning Substation was to update the process used to identify project requirements. To that end, SRP introduced processes advanced by the IntelliGrid consortium. This EPRI-led international initiative is creating the technical foundation for a smart power grid that links electricity with communications and computer control to achieve substantial gains in overall performance.

A major early consortium product is the IntelliGrid Architecture: an open-standards, requirements-based approach for integrating data networks and equipment. The IntelliGrid Architecture provides the methodology for specifying intelligent systems, so that today's investments are not wasted on equipment or systems that must be abandoned or re-engineered later.

A key element of the architecture is a process adapted from the software industry to identify system requirements. The process employs “use cases” or scenarios that describe how actors interact with a system to accomplish a specific goal. Software and other high-tech companies have employed use cases for years to identify and document project requirements. Now, SRP and other IntelliGrid team members are showing how use cases can benefit the power industry as well.

Developing use cases involves identifying the actors, which may be human beings or pieces of hardware, and their goals. Participants in use-case development then identify the steps each actor takes and the system's response to the actor's actions (see IntelliGrid Basics). It may take one or more scenarios to describe how the goal is achieved.

Done properly, use cases translate into a project's functional requirements — what the system must do. For example, a functional requirement might state: “The meter shall time stamp measured values.” But projects also need to capture nonfunctional requirements — what a system must be — such as constraints, behavior and performance targets. For example, a nonfunctional requirement might state: “The meter shall produce data time stamped to 10-ms resolution with 1-ms accuracy.”

For projects such as SRP's Browning Substation, capturing requirements may take one or two work weeks to create the use cases, followed by several more weeks to conduct workshops, brainstorming sessions and reviews. Finally, the information must be documented. The total amount of time varies greatly with the size and scope of the project.

Although some members of the SRP project team were initially skeptical of the use-case approach, they soon came to accept the process. The enhanced network was one benefit that won them over. Consulting with users and stakeholders on system requirements can help build consensus in an organization for the type of underlying project that is needed. Users from more than a half-dozen departments teamed up to participate in the Browning Substation project use-case development.

FROM REQUIREMENTS TO REALITY

With the use-case process completed and the requirements defined, SRP moved forward with the integration project. The project team added new hardware, including an additional server connected to the enterprise server, and software to create a data backbone using a common information model that lets each application communicate with the new system. The data are integrated into a common database for logic processing.

A key software product the team settled on is My IEDs from Subnet Solutions (Calgary, Alberta, Canada). This IED connection-management tool gives users a list of substations and devices they have permission to access, and provides the users with information on the desired device. With this capability, employees in the field and in SRP offices can log in to the SRP intranet and get the information they need.

The software also includes intelligent notification features. For example, if a critical threshold has been reached for a particular device, My IEDs knows the proper person to contact to address that specific problem.

Another decision that had to be made in the integration process was which communications protocol to use. The IntelliGrid Architecture normally steers people toward the IEC 61850 standard. SRP, however, uses DNP3 (distributed network protocol) and decided to remain consistent. Here again, the IntelliGrid requirements-based approach paid off. DNP3 and IEC 61850 are just two of many candidate technologies in this space. The IntelliGrid philosophy is about capturing requirements first and then mapping them to suitable technologies second.

REDUCING COSTS AND RISK

Although little hard data is available, use cases promise significant cost reductions, especially over the long run, because well-developed requirements reduce the need for later reworking. Initially, however, the use-case approach requires additional training, making it more expensive than the traditional approach. In SRP's experience, developing use cases added some extra time and cost, but the requirements are captured in more thorough detail using the IntelliGrid process. The requirements are so well documented that it will be easier for SRP to apply the approach to the next substation project. SRP has several receiving stations that may be candidates for IED integration and expects to employ use cases for those projects, driving costs down dramatically.

The cost of the hardware needed to complete a project doesn't change. In the case of the Browning Substation upgrade, the equipment cost alone was US$250,000. Preventing the failure of a high-voltage transformer can save more than $3 million, justifying the project based on risk mitigation alone.

ACKNOWLEDGMENT

The author would like to thank the Salt River Project personnel involved in this project, particularly John Blevins, for providing the information upon which this article is based.

---

Don Von Dollen is the program manager of the IntelliGrid program at the Electric Power Research Institute. The program's activities focus on accelerating the transformation of the power-delivery infrastructure into the intelligent grid needed to support our future society through a unique collaboration of public and private stakeholders. DVONDOLL@epri.com

TWELVE STEPS TO EFFECTIVE USE CASES

1. Name the system scope and boundaries.
2. Brainstorm and list the primary actors.
3. Brainstorm and exhaustively list the user goals.
4. Capture the outermost summary use case for each primary actor.
5. Reconsider and revise the summary use cases.
6. Select one use case to expand.
7. Capture stakeholders, interests, preconditions and guarantees.
8. Write the main success scenario steps.
9. Brainstorm possible failure and alternate success conditions.
10. Write the alternate scenario (extension) steps.
11. Extract complex flows to sub-cases and merge trivial ones.
12. Readjust the original set of summary cases.

INTELLIGRID BASICS

The EPRI's IntelliGrid initiative is creating the vision and technical foundation for a smart power-delivery system that links electricity with communications and computer control to achieve tremendous gains in reliability, capacity, and advanced customer services. The IntelliGrid Architecture provides an open, standards-based framework that ensures the interoperability of grid components, now and in the future as utilities' needs evolve.

The SRP substation integration project offers a small-scale example of the type of requirements-based systems integration that will enable the development of the smart 21st century power grid. Another example of requirements-based systems integration is the systems engineering approach documented in the “SCE Approaches Massive AMI Rollout” (T&D World, January 2007).

The IntelliGrid Consortium is a worldwide alliance of utilities, manufacturers and public agencies. Current partners include:

Utilities:

Arkansas Electric Cooperative Corp.

CenterPoint Energy Inc.

Central Hudson Gas & Electric Corp.

Consumers Power

Dairyland Power Cooperative

Duke Energy Corp.

EdF

FirstEnergy Service Co.

Golden Valley Electric Assn. Inc.

Great River Energy

Hawaiian Electric Co. Inc.

Hoosier Energy Rural Electric Cooperative Inc.

Kansas City Power & Light Co.

Korea Electric Power Corp.

Lincoln Electric System

LIPA

NYPA

PNM Resources

Polish Power Grid Operator

Salt River Project

TXU

Manufacturers:

ABB

Hitachi

Public agencies:

Association of State Energy Research and Technology Transfer Institutions

International Brotherhood of Electrical Workers

National Association of Regulatory Utility Commissioners

National Association of State Energy Officials

National Conference of State Legislatures

National Governors Association

State Energy Offices and Research Programs