Deregulation created a vibrant marketplace for generation over the past decade, which has led to a proliferation of new participants in the electricity marketplace as well as new generators. And this market, in combination with economic expansion in California, has placed ever-greater demands on the state's electric power grid.

To respond to these demands and provide the flexibility expected by the marketplace, Southern California Edison (SCE; Rosemead, California, U.S.) has had to evolve the principles by which new generation is connected to the grid. Getting the grid to respond to contingencies in a “smart” manner is part of this evolution.

Since deregulation has decoupled customer reliability-of-service obligations from generation providers, the door has been opened for owners of new generation to weigh the risk of disconnection from the grid against the costs of constructing redundant transmission. As a result, a principle change during this era has been to allow generation to “trip” when transmission outages occur as a matter of common practice; whereas in the past, it was generally preferable for the vertically integrated utility to construct redundant transmission to avoid loss of generation under such conditions.

PROTECTION PHILOSOPHY

SCE's transmission network consists mainly of 500-kV and 230-kV transmission with multiple long corridors connecting remote generation or connecting to other utilities (Fig. 1). This network has become increasingly congested as a result of ever-growing customer load and significant additions of new generation. Because of long lead times related to new transmission, SCE relies on remedial action schemes (RASs) to manage potential overloads that arise when transmission facilities are disconnected because of faults and other adverse events that can affect facilities. This is also one of the facets of the smart grid that is receiving so much attention in Transmission & Distribution World and throughout the power industry.

The emphasis of reliability criteria has been to preserve the continuity of service to the load, while accepting the practical realities that undesirable events may occur that require mitigation to arrest uncontrolled cascading across the grid. Monitoring the system and recognizing events that have the potential for cascading and implementing appropriate mitigation is a critical function of any power-system protection philosophy to preserve reliability and is central to the RAS method.

REMEDIAL ACTION SCHEMES

Energy flows naturally over the power grid from generator locations, in proportion to the size of the transmission facilities, to locations where the energy is consumed. There is a natural shift or redistribution of power on the grid following a transmission line outage.

Direct-tripping generation is a classic method to protect the remaining operating transmission lines from overloads that could result from the congestion. This method of protecting the grid, by tripping generation, is a RAS, also known as a system integrity protection scheme (SIPS) — and such a scheme has existed since the earliest days of electric power systems.

RAS is widely understood within the utility industry and employed where conditions warrant. The decision to use this technique has its own complexities related to economics, impacts on the generation equipment, implications related to loss of generation and operational considerations unique to each utility.

Up until the early 1990s, for SCE it was a rare situation when the tripping of generation using a RAS was preferable to new transmission line construction. Over the last 20 years or so, the cost of building and the difficulty of siting new transmission has made RASs an economical and timely choice to accommodate the development of new generation.

RASs are comprised of three parts: monitoring, event detection and mitigation. Depending on the scheme's control logic, prior studies are typically performed to indicate “arming levels” for the scheme and mitigation needed for each problem event. Equipment that fulfills each of the RAS functional requirements typically has been deployed within one or two neighboring substations with a local controller that initiates any mitigation signal actions within the substation or neighboring substations.

GOING TO CENTRALIZED RAS

Today, SCE uses 17 existing RASs and expects that 55 new RAS arrangements will be deployed in the next few years. About half of SCE's 500-kV and 230-kV transmission lines are monitored for contingencies and flow levels, and controlled by local RASs (Fig. 1, table). Because of the many problems associated with the limitations of existing RAS deployments, centralization is viewed as desirable from both a planning and an operating perspective.

One key problem associated with existing schemes is the one-size-fits-all dilemma. This dilemma arises because existing logic controllers typically deployed in substations have limited logic capability, sometimes necessitating over-mitigation to ensure that the full range of possible system events is addressed and that under-mitigation does not occur.

Under-mitigation brings its own set of consequences that can include problems serious enough to require immediate operator intervention or to precipitate the tripping of other facilities. Being able to prescriptively mitigate system events with “just-right” mitigation can save money and prevent other more serious system problems.

The technology opportunities that exist today to improve RASs — and make them centralized — include combining the capabilities of new intelligent electronic devices (IEDs) with SCE's high-speed telecommunications assets (which includes more than 5000 miles [8047 km] of high-speed telecommunication capability, of which nearly 60% is fiber based). Additionally, basing the system on IEC 61850 standards will permit SCE to use more sophisticated controls than are possible with today's commonly deployed equipment.

In the centralized RAS (C-RAS), IEDs will be used to monitor current grid conditions and detect events that potentially require mitigation. IEDs will “package” key data and, using IEC 61850 communication protocols, send the data to a central control processor that will decide what, if any, mitigation (generator dropping or load dropping) is needed or allowed under the specified criteria. The central controller will then package mitigation instructions for delivery to key locations, where breakers or other switching on the grid will be performed, accomplishing the needed mitigation action.

Expected benefits from the successful development and deployment of the C-RAS project include more-precise mitigation, simplified design and maintenance, easier implementation of control logic changes to respond to varying system conditions, improved automated testing, simpler operation methods and better event data archival. SCE's franchise service territory covers 50,000 sq miles (129,500 sq km). Because its electric grid spans an equally large geographic area, implementing RAS logic changes sends key technical personnel traveling hundreds of miles to logic controllers in remotely located substations. Centralization will facilitate the implementation of logic changes, with capabilities to manage multiple operating and planned RAS arrangements, while saving valuable time and expense.

THE TESTING PHASES

A speed requirement of three cycles, or 50 msec, was established based on existing RASs. These systems perform event and fault detection, processing and mitigation, with a total elapsed time of approximately 16 cycles, or 267 msec, including circuit-breaker operating delays. Existing schemes allow for approximately two cycles for processing, or 33.4 msec. Adding a cycle to the processing time requirement for margin puts C-RAS on a 50-msec time-delay limitation.

In the first phase of testing, using General Electric's new N-60 universal relays and SCE's high-speed communications channels, laboratory results showed an achievement of 11 msec to 12 msec within the laboratory local area network (LAN) and 20 msec to 22.5 msec over a wide area network (WAN), covering physical distances up to 660 miles (1062 km). In this phase of testing, generic object-oriented substation event (GOOSE) data packets were used to transmit breaker status information to verify that the speed of the communications infrastructure over the physical distances involved could achieve the time requirements for C-RAS. Although not obvious, this aspect of testing was critical since the communications infrastructure included multiple GOOSE hops from fiber to microwave and back. IEC 61850 defines GOOSE profile for Ethernet-based fast communication of substation events. Each GOOSE data packet contains commands, indicators and alarms as user-configured messages to and from IEDs.

In the test over the WAN (Fig. 2), an N-60 relay in the lab was used to initiate an event GOOSE, simulating a line outage detection event. This was sent over the WAN to SCE's distant Magunden substation, which was set up to echo the GOOSE back to the lab.

There, an N-60 relay, simulating the central logic processor, received the event GOOSE and sent a mitigation GOOSE to Magunden substation. This again echoed the GOOSE back to the lab, where an N-60 mitigation relay received the GOOSE. The entirety of this communication speed (Fig. 3) was timed with a GPS clock.

GE's N-60 relays perform multiple functions of send, receive and self-check, in addition to transmitting GOOSE data packets containing multiple electrical parameters, the primary CPU can process up to 512 lines of programmable logic. Newer products extend this range to more than 4500 lines of logic.

In the second phase of testing (Fig. 2), a high-performance Dell processor was used to host the analytics of the central logic controller and a high-performance Dell server was used to host the data historian for event archival. Off-the-shelf SISCO software was used for the GOOSE interface and off-the-shelf OSIsoft's PI software was used for the network representation and the data historian.

Laboratory results with the central logic controller and historian, in combination with the GE N-60 universal relays and SCE's high-speed communications channels, achieved speeds of 40 msec over a WAN, again covering physical distances up to 660 miles (Fig. 4).

In the third and final phase of testing, yet to be completed, SCE will use customized software in place of the off-the-shelf software and equipment arranged to mirror an existing scheme. SCE will deploy the N-60 relays in remote substations and set them up to perform real-time monitoring and event detection, although mitigation will occur through the existing RAS. SCE expects to achieve the target 50-msec processing time for the C-RAS project.

In the future, the C-RAS project is expected to provide a bird's eye view of multiple events affecting the larger grid. These smart-grid developments open the door to new automation and control philosophies that can serve as broad safety nets to protect system reliability well beyond today's reliability criteria limits.

ACKNOWLEDGEMENTS

The author is indebted to the ingenuity, creativity and unwavering support from Shashi Pandey and Doug Dawson, consultants; Herbert Falk, SISCO; Norma Tam, OSISoft; Richard Schmittdiel, Skip Perry and Wayne Wong, SCE Telecommunications; Curtis Sanden, SCE Protection; and Pui Lau and Mark Adamiak, General Electric.

Patricia L. Arons is manager of Transmission & Interconnection Planning for Southern California Edison (SCE; Rosemead, California, U.S.). After earning a BSEE degree from Rensselaer Polytechnic Institute and working at Power Technologies Inc., a consulting and software company in upstate New York, Arons joined SCE in 1983 as a planning engineer in the system development department. She has helped develop SCE's transmission and subtransmission facilities and has been concerned with issues involving open-access transmission tariffs, generation interconnections and general business policies related to transmission. Arons is responsible for planning SCE's electric grid facilities.
patricia.arons@sce.com

Line Lengths Monitored

	Miles of transmission circuits
Voltage	Total	RAS monitored
500 kV	1183 miles	1069 miles (90%)
230 kV	3574 miles	1181 miles (33%)
115 kV	1846 miles	350 miles (19%)
All	6603 miles	2600 miles (40%)	Observation: Almost all bulk power lines bringing generation/imports into the greater Los Angeles basin load area are being monitored for contingencies and flow levels with remedial action schemes.