While the Seeds of Automation Innovations are being Sown Across the Country, some of the most intriguing projects are coming from independent municipal utilities in search of solutions to their unique situations. In Orangeburg, South Carolina, U.S., a picturesque town of around 14,000, the Orangeburg Department of Public Utilities (Orangeburg DPU) began a project to deliver top-notch customer service. Soon, the project grew to include the integration of self-healing feeders and a distribution automation system.

The project began with an effort to increase customer satisfaction by reducing the utility's system average interruption duration index (SAIDI). Orangeburg DPU set a goal to cut the number and duration of its customer outages. The first step was simple: implement a tree-trimming program. This lowered the utility's SAIDI from 3.6 to 2.6. But Orangeburg DPU wanted to do more. With a goal of reducing SAIDI to 1.0, the public utility decided to implement a self-healing feeder solution.

REQUIREMENTS OF SELF-HEALING OPTIONS

Orangeburg DPU launched an extensive study of the available self-healing options. While all promised to reduce the number and duration of outages on unfaulted sections, not every solution could integrate well with other technologies or applications. Ultimately, the utility decided the solution it would choose must have the flexibility to work with other technologies and applications, be cost justifiable and be capable of being deployed across the entire service area.

A multiservice utility, Orangeburg DPU is the largest municipal electric utility in the state of South Carolina. Orangeburg DPU owns and operates 22 electric substations, with service available at 115 kV, 46 kV, 25 V, 8.32 kV and 480 V for large customers. Its distribution system operates 60 feeders on 800 miles (1287 km) of line. The largest circuit has 1600 customers, but most feeders operate with 500 to 800 customers per circuit.

To avoid costly mistakes down the road, Orangeburg DPU used a phased approach in the build-out of its strategy. The utility identified its top-four priorities for system architecture and feeder application:

  1. Cost

    The solution had to be cost effective. For Orangeburg DPU, this would mean reusing the common infrastructure shared by all its applications. By using common infrastructure such as platforms, sensors and communications, the utility prepared to reduce ongoing maintenance costs.

  2. Scalability

    Orangeburg DPU needed a technology that would be scalable across the entire system, with the potential to expand from a pilot system to serving customers across 340 sq miles (881 sq km). Orangeburg DPU's ongoing objective is that all new technology be fully expandable and easy to add incrementally.

  3. Interoperability

    The utility required applications that would operate smoothly across a suite of applications, in which each application would be adaptive to the current state of the system, whether the network state was normal or abnormal.

  4. Daily accountability

    Orangeburg DPU needed a system that would deliver daily updates and maintenance of the distribution network to be performed from a single data source to all applicable functions.

CENTRALIZED OPERATION

The Orangeburg DPU selected these priorities to work within its architecture, through which system functions perform from a centralized distribution control center. The dispatch center monitors and controls the electric system under 24/7 operations. A state-of-the-art fiber-optic communications network supports the distribution management system (DMS), which provides advanced visualization of the distribution network.

All applications use the distribution network topology model to provide and display the real-time network status. The network map supports dynamic features such as a direction of power flow indicator, network tracing to find key network elements, color by phase, loop detection and coloring line segments with violations. All of these features help to improve situational awareness for Orangeburg DPU operators.

Orangeburg DPU has 30 Cooper Power Systems (Waukesha, Wisconsin, U.S.) NOVA reclosers on its network and 12 S&C Electric Co. (Chicago, Illinois, U.S.) Scada-Mate switches. During a recent outage, Orangeburg DPU operators were able to visually identify the location of the break through the dynamic network colorization features.

The system provides operator switching support with a 3-phase dynamic, unbalanced load flow by supporting the operator's ad hoc analysis. Before an operator closes a switch, a click on the device can run a load flow from a pop-up menu. Any violations that may result from the operation will display in a pop-up box (for example, “operation will cause 109% overload” or “operation will cause a 32% phase imbalance.”)

THE SELF-HEALING FEEDER SOLUTION

The initial cost justification of the system was to improve SAIDI, the active application in achieving this objective is a self-healing circuit. Various technologies were available, but a load-flow-based self-healing feeder solution, developed by Advanced Control Systems (ACS; Norcross, Georgia, U.S.), was selected. Functioning on ACS's PRISM DMS, the application is called Fault Detection Isolation and Restoration (FDIR). The model-driven FDIR application offers several advantages over a script-driven solution.

FDIR does not assume a normal network state, so it will adapt readily to the current topology and provide an optimum restoration solution to restore unfaulted sections, even with abnormal feeder topology. The major advantage of this network-adaptive capability is the provision of a switching solution even if a second or third fault occurs on an abnormal circuit. It also means operation without removing or changing protective schemes already in use, such as overvoltage relays, which often results in changes to the network topology.

FDIR is also able to operate in both radial feeder networks and in looped networks. This is an important feature, as Orangeburg DPU expects to see increased usage of distributed resources in the future, essentially causing a multisource or looped topology. In this case, it uses fault direction in addition to fault detection.

INTEROPERABILITY, FLEXIBILITY, EXPANDABILITY

Another requirement for Orangeburg DPU was that the system be able to operate using the various reclosers and intelligent electronic devices from different manufacturers already in service on the network. The PRISM DMS with FDIR was an optimum solution because it is able to seamlessly interface to each of these different controllers and devices without the need to replace hardware or add intermediary controllers. A profile library enables a user to mix any combination of switches or overvoltage relays/overcurrent relays in order to perform the restoration function.

During complex scenarios where the load pickup of the unfaulted line section of a faulted feeder is greater than the second feeder's capacity, a load-transfer switching plan will be included in the switching solution in order to transfer load from the second feeder to a third feeder. This action will relieve the capacity problem on the second feeder, enabling it to pick up the load on the unfaulted sections of the faulted feeder. If a plan exists, FDIR is also able to break the unfaulted line sections into load groups by using more than one tie feeder for restoration.

This is all expandable to any sized network. Orangeburg DPU expects to model the entire system as a single network model that will support all other planned applications. There is also a capability to generate a return-to-normal topology switching plan. Restoration phases can be configured to operate in two modes: automatic or advisory. The system is configured so that both the isolation switching phase and the upstream restoration switching phase are automatic. However, the downstream restoration switching phase is presented to the operator for review and implementation.

The switching steps are accessed in a switching plan management application. The operator may ask for a plan that considers manual switches as well as remotely controlled switches. Once approved, the plan can be implemented in a single-step mode or will run by executing all of the steps sequentially.

COMPLEXITIES BEYOND COSTS

Orangeburg DPU faced many challenges when selecting a vendor and equipment. As for many small utilities, for Orangeburg, cost was generally an important factor, but other factors also carried weight:

  • Management of the system

    Who would be responsible for keeping the units functioning and up-to-date, and how would that be accomplished?

  • Environmental concerns

    Would the unit perform when it was 100°F (38°C) with 90% humidity during a South Carolina summer, or when there was a half-inch of ice outside during a January ice storm?

  • Dependability

    What was the failure rate of the units? Was the communication solid so that it was available 99.99% of the time?

  • Deployment

    Where was the best place to put the units to provide optimum benefits?

These are among the many questions Orangeburg DPU engineers needed to answer before the utility could move forward. They are very typical and may be some of the same questions another utility will need to answer, but each utility is unique and the answers will vary.

Often, the most costly aspect of operating applications is the maintenance of the network model and application database. As Orangeburg DPU examined maintenance costs in detail, it became clear that without an integrated approach to this implementation, costs would be repeated to support private databases and multiple models (of essentially the same distribution feeder network).

To trim both ongoing costs and the complexity of maintenance tasks, the utility needed a solution in which the model and real-time database would be common to all network-based applications.

With the selected applications, the source for Orangeburg DPU's network model is the geographic information system (GIS). Daily updates are made from an ESRI (Redlands, California, U.S.) GIS, the same source used by the outage management system. ACS's DASmap is used to extract the GIS model changes. Orangeburg DPU uses this to build connectivity and export a node-arc-node network model, automatically add the points to the real-time database and create the dynamic map display.

The new file changes are loaded daily into the real-time on-line DMS without downtime. This process takes from 30 minutes to four hours, depending on the size and complexity of the update. A new program version will soon handle incremental updates, speeding up the process as it provides automatic detection of GIS changes from one update to the next.

A Bright Future

To realize the SAIDI gains of the pilot program on a system-wide basis, Orangeburg DPU plans to increase the area under automation control. In the next phase, a feeder-loss minimization application will automate the control of the system's 57 capacitor banks. The power factor for each feeder will be calculated from the telemetered kilowatt and kilovolt-amp reactive. The load flow will be used to determine reactive power requirements at various capacitor bank locations, as well as for the entire feeder. This will be accomplished in real time, and the resulting capacitor switch commands will be presented as either operator action recommendations or automatically executed. Orangeburg DPU plans to add other applications in subsequent phases, including optimized switching, as well as intelligent switch-plan generation for planned outages.

Beginning with self-healing feeders, Orangeburg DPU is deploying an integrated suite of applications. Through careful management of adaptive, common model-derived applications, Orangeburg DPU is now poised to adopt additional applications to capture and sustain the gains in customer satisfaction along with increased reliability, flexibility and efficiency.


John Bagwell (jbagwell@orbgdpu.com) has served as director of the City of Orangeburg Department of Public Utilities (DPU) electric division since 1998. Bagwell came to Orangeburg DPU in 1987 as a control systems engineer, where he installed a SCADA system covering 23 electrical substations, two peak shaving generator plants, a gas distribution system, a 19-MGD water treatment plant and a 9-MGD wastewater treatment plant. He also installed a fiber-optic network consisting of OC3 backbone and 50 miles (80 km) of fiber cable, and was instrumental in installing the department's first personal computer network. Bagwell earned a BSEE degree from Clemson University and is president of the South Carolina Association of Municipal Power Systems.