Customer Satisfaction Will Remain High if we minimize the number and duration of outages a customer experiences. To continue already high customer-satisfaction ratings, PPL Electric Utilities (Allentown, Pennsylvania, U.S.) is commissioning its first two fully automated distribution circuits with a system it calls Supervised Intelligent Switching to Reduce SAIDI (SISRS). PPL believes this is the first distribution automation (DA) of its kind installed in North America.
PPL has approximately 1.4 million customers in its 10,000-sq mile (25,900-sq km) service territory located in central eastern Pennsylvania. The utility has about 1100 12-kV circuits, the vast majority of which are overhead construction. PPL has several metropolitan areas, with the biggest cities being Harrisburg, Allentown, Lancaster and Scranton. Additionally, the territory is mostly rural with rolling hills and mountains.
DISTRIBUTION AUTOMATION JUSTIFICATION
PPL tried unsuccessfully for many years to justify DA on an economic basis. When it chose to find ways to increase reliability for its customers, DA showed its benefits.
A team was commissioned to examine ways to lower the system average interruption duration index (SAIDI) and to set up a plan to go forward. The SAIDI team investigated many alternatives to reduce the duration of outages. The SAIDI savings of each alternative was calculated, as well as an estimated cost to achieve those savings. The cost was divided by the SAIDI savings to come up with a dollar/SAIDI-minute value.
Alternatives with lower values were implemented immediately. Some of the alternatives were: 24/7 trouble men, restoring as many customers as possible before repairing the trouble, increased sectionalizing (fusing), increased line maintenance and inspection, equipment replacement, increased tree trimming, increased load-transfer capabilities and DA. This is not an all-inclusive list of the ways PPL addressed lowering its SAIDI, but enough to give an idea of what the utility tried. The dollar/SAIDI-minute method is not meant to indicate how much it costs to reduce SAIDI. It is meant to put each alternative on a relative basis to see which projects can provide more relative investment benefit and which ones should be done first.
SAIDI is the product of the number of customers affected by a permanent outage and the duration of that outage divided by the total number of customers in the company. From this equation, to reduce SAIDI, the number of customers who see a permanent outage must be lowered or the duration of that outage must be reduced, or both. The total number of customers is fixed. PPL saw SISRS as a way to drastically reduce the number of customers who see a permanent outage. However, the reduction in those who see a permanent outage comes at a cost, those customers who are restored will still see a momentary interruption. So a reduction in SAIDI with SISRS may lead to an increase in the momentary average interruption frequency index.
DESIGNING WITH AUTOMATIC DEVICES
Preventing an outage in the tree-rich Pocono Mountains area is a daunting effort. PPL accepted that outages are going to happen, but that it would try to limit the number of customers who experience a permanent outage. The utility designed its system with enough automatic devices to limit any given permanent outage to 500 customers.
Once PPL decided to embark on an automation effort, it needed to find the circuits that would show the most benefit. The search began with the top 20% of the worst-performing circuits. These are circuits where history indicates that many outages have occurred. After all, why install the equipment on a circuit that has never had any outages?
Next, PPL discounted single-phase outages because the utility doesn't use single-phase ties. PPL then looked for outages that occurred at the circuit breaker. Outages at the breaker were in the first section of the circuit and all the customers were out. This condition allows for the largest percentage of customers to be restored with automation.
Lastly, if all the aforementioned conditions were met, we checked to see if there were any existing ties to the circuit, and the tie had to have the capacity to handle most of the adjacent circuit being transferred to it during restoration.
This analysis left about 20 circuits. PPL averaged the historical outages over the last five years to get an estimated annual SAIDI for each circuit. The utility assumed 50% of the permanent outages could be eliminated. If PPL used four automated devices per circuit, and assumed a normal distribution of outages, some outages would restore 75% of the customers, some would restore 50% of the customers, some would restore 25% and some would restore 0%. So PPL assumed the average of 50% even though it targeted the circuits where the circuit breaker operated most of the time. Two circuits were selected for the pilot project: the Blooming Glen 06-2 circuit and the Lanark 23-1 circuit.
LOCAL VERSUS GLOBAL
When PPL did due diligence on the automation systems that were available, the utility noticed there were two kinds of systems: local and global. These are not industry terms but PPL observations.
Local systems tend to be individually programmed devices that perform required tasks based on input from each other. The switching devices are intelligent and work quite well if every scenario is determined beforehand on each circuit. If the circuit is rearranged, they must be reprogrammed. At that time, the systems did not provide notification to anyone about what they had done. The customers are back in service, but the operator or dispatcher does not know it.
Global systems are more complex and require robust communication networks. The switching devices are intelligent and all of them communicate back to a single large computer with their status and alarms. This computer is usually centrally located back at the main office or operations center. The restoration solution is determined at the central computer and the instructions are sent to the individual switches, which communicate back after they are done. This system has the advantage of being able to show the operator or dispatcher what is going on.
PPL liked the global solution better, but did not have the funding for large computers and big communication networks. The utility came up with a hybrid design that can grow into the global system in the future. This will allow for a gradual build out and spread the cost out over 20 years. An approach such as this requires a standard protocol (DNP3) so that many different devices can be used.
PPL selected Advanced Control Systems (ACS; Norcross, Georgia, U.S.) to develop the system, which consists of a host computer located at a substation. This computer has an operational model of the circuit and adjacent tie circuits. The host performs a load flow on the model before composing a restoration solution. Since the host knows the condition of the circuit at any time, it can even attempt to restore customers if a second fault occurs based on abnormal conditions.
All of the switching devices on the trial circuits are ABB electronic vacuum reclosers. Some of the reclosers interrupt fault current and the others are used as switches to isolate the fault and restore load. PPL decided to use reclosers for all switching devices to keep the development to a minimum, even though the cost was somewhat higher. Once PPL figured out how to program one, the others were easy. The utility has already begun testing motorized air-break switches at single locations with operator control. These will be phased into the SISRS program as PPL becomes more familiar with them.
The reclosers communicate back to the host by a 900-MHz spread-spectrum radio. PPL selected Microwave Data Systems (MDS; Rochester, New York, U.S.) as the radio supplier and radio survey consultant. The radio network uses repeaters to get to each device. The host polls each recloser for its status about every five seconds. If a fault occurs, the fault flag is set and passed back to the host at the next poll.
When a fault condition exists, SISRS gathers the fault flag from each switch and determines which section of the line has the fault. It then isolates the fault by opening the switch on each side of the fault. After fault isolation, it attempts an upstream restoration if possible. Upstream is done first since the circuit was supplying that portion of line before the fault. Then it attempts downstream restoration after checking to see if the adjacent circuit can accept the load. This can be done using real-time load information from the adjacent sub or with load curves that simulate it.
If the receiving circuit can accept the load, SISRS will close the normally open recloser to finish the downstream restoration. The reclosers switch to alternate settings and protect in the backfeed condition. Tests show a total restoration time of 2 to 3 minutes.
The results are communicated back to the main SCADA computers by a digital cellular link through a firewall and ModBus interpreter. PPL uses a Telemetric (Boise, Idaho, U.S.) modem to communicate back to its SCADA system. The operator has complete control of every switch in the system through the SCADA link. There are one-lines that the operator can view to determine what has happened. Load data is also available to display on the operator's screen. The load data is transmitted hourly. The operator can query SISRS for the latest load information on each device. The system can be disabled locally at the host computer cabinet or remotely through SCADA.
Once data is transmitted to the SCADA system, it can be passed on to the outage management system as a confirmed outage to aid the dispatcher in sending crews to the proper location.
THE RADIO NETWORK
It is easy to try and shortcut the radio network. Each repeater costs money and takes more time to design and build. PPL has learned that the radio network is not the place to try and cut costs. The utility was on a tight schedule and rather than installing several high antennas as repeaters, it decided to try and use the store-and-forward method to get communication to the normally open point. This involved each device passing the message on to the next device serially until the message got to the proper device.
PPL did the radio survey in November, when there were no leaves on the trees, and installed the system in April, when there were still no leaves on the trees. The system was communicating quite nicely. Then May came and so did the leaves. The line-of-sight 900-MHz radios didn't like the leaves. SISRS lost contact with one or more devices on a regular basis and started sending loss-of-communication messages frequently. It was particularly bad when communication was lost with the first device, which resulted in lost communication with all the devices. The system was corrected and this lesson learned.
Change is difficult. Although installing a system as complex as SISRS was technologically possible, every department had to be onboard to make it work. PPL could have done a better job educating all of our operators and linemen.
One mistake was commissioning the system before the SCADA one-lines were complete. When the system operated, the operator had to go to a website consisting entirely of text messages. These messages were mixed in with dozens of other remotely controlled devices and were cryptic in nature. The operators were not familiar with the nomenclature they saw, because it was not the standard nomenclature they were used to seeing on SCADA. This could only lead to confusion in a time of crisis.
PPL learned that thorough operating instructions for SISRS should be issued before the systems are commissioned. The operators need the knowledge to understand how the system works.
The utility needs to do a better job of communicating the use of reclosers as simple switches and as fault-interrupting devices. The linemen as well as the operators are used to seeing reclosers work as reclosers. When the recloser, acting as a switch, failed to isolate a fault, it was confusing to them, even though it was operating as designed.
Another lesson learned is that there is a lot of work required to design, build and install the systems, and a dedicated team of engineers, technicians, operators, IT personnel and linemen is necessary. Coordinating all of them requires a full-time project manager or engineer. This person also needs to keep up with the pace of evolving technology to make sure the investment that has been made can continue to work with the newer technology.
THE NEXT IMPLEMENTATION
PPL's next step will be to take what it has learned from the two independent circuits that are already installed and apply it to a fully integrated system. This system will have four substations with six circuits and 24 switching devices. All of this will be controlled with one host computer at one of the substations. The radio network will contain several main repeaters that can see all the devices.
PPL wants to build a lab with a working SISRS installation that has actual recloser controls in series with lights that simulate actual load. Flat-panel displays will show what is happening both graphically and with text alarm messages. Users will be able to apply any kind of fault they can think of and watch how the system reacts. Instructors will answer any questions about how the system works. Operators, engineers and technicians will be able to spend time there to become acquainted with the system.
The radio network has been designed, the SCADA displays are in progress, the reclosers are being installed and the operating instructions are being developed.
Howard Slugocki is a supervising engineer in the distribution planning department of PPL Electric Utilities. He has more than 25 years of experience in engineering, distribution planning, customer service and information technology at PPL. He has developed planning tools and techniques to help PPL make informed decisions regarding capital improvements and maximizing system utilization. Slugocki is a system storm director during storm restoration. He has spent the last two years developing and implementing a distribution automation program for PPL. He is a licensed professional electrical engineer in Pennsylvania and has a bachelor's degree in civil engineering from Michigan Technological University.