Utility experience shows AMR could support outage management
The use of automated meter reading (AMR) as an integrated part of the emergency restoration solution is an exciting prospect that could allow utilities to more proactively address customer outages. Many of today’s advanced metering infrastructure (AMI) business cases are claiming a benefit for this enticing feature. However, first generation AMR solutions have presented unexpected results. This article will discuss results found in two utilities that have full AMR deployment. Before presenting those results, though, an explanation of the outage restoration process with and without AMR is in order.
Emergency restoration is part of any electric utility’s distribution operations. Restoration activities can range from a single residential customer outage to the entire utility’s customer base being out of service. Our focus here is on major outages---those where 50 percent (300,000-500,000 customers) or more of the customers are without power for more than 4 days. All utilities have major emergency restoration plans that permit management to conduct an effective restoration effort in the least amount of time. These plans, usually reserved for medium and large outages, are extremely critical and can have a significant impact on the success of a restoration effort.
Equally important is the use of integrated information systems to capture, analyze, and report outage information to utility managers in order to determine staffing, priorities, material requirements, and restoration times. (See exhibit 1.)
Prior to the 1980s, most utilities used slips of paper to represent individual customer outages, which were then sorted into, what I like to refer to as, the “Shoebox” approach. The “Shoebox” approach was a means of sorting individually reported outages into more coherent groupings that represented the physical world of distribution feeders or circuits. This allowed distribution dispatchers to efficiently dispatch repair and first responder crews.
The Outage Management System (OMS) -- the electronic “Shoebox” – first appeared in the 1980s and is at the core of all of these systems. OMS takes individual outage data from the customer call center, Integrated Voice Recognition Unit (IVRU) or Supervisory Control and Data Acquisition (SCADA) system and displays the predicted outage location and the customers impacted. Many of these systems have the ability to rank outages by the number of customers affected. As a result, utilities can now restore the most customers in the shortest amount of time, effectively replacing the traditional “Shoebox” approach. Today, the output of the OMS is used to inform the different customer communications channels on the progress of restoration efforts.
How does an OMS perform its role? OMSs maintains an up-to-date distribution system connectivity model that reflects the current configuration of the physical electric system. Reported outages are analyzed against the system model compared to the current operating status of key equipment, e.g., substations, transformers, switches, and fuses.
Today’s OMSs have business rules that allow the efficient management of large scale outages and restoration efforts. Proper integration of key systems--including customer information systems (CIS), IVRU, energy management systems ( EMS), and mobile work force management (MWF) ---significantly reduces the need for manual and redundant data entry.
The SCADA/EMS systems supply valuable real-time information about operating conditions and system configuration. When combined with the OMS connectivity model, circuit level outages can be quickly identified and outage reports mapped and analyzed.
Leading OMSs provide a library of planned switching scenarios that the switching coordinator uses to manage outages. Restoration procedures and processes can also be defined in the OMS to help with large-scale distribution outage restorations. The procedure defines the correct sequence of events to safely and effectively restore circuits. The sequencing is coordinated with the real-time system status from the EMS.
Integration between the OMS and a MWF system allows dispatching of OMS results to field personnel. Field information---such as outage validation, cause, and estimated time to restore---are sent back electronically to the OMS, passing seamlessly to the CIS for call center notification and IVR message updates.
Integrating a Geographic Information System (GIS) to the OMS allows electric connectivity data to regularly pass to the OMS for developing the model that reflects the as-operated configuration of the electric system in the field.
An AMR system can replace or supplement much of the granular information received from the CIS and IVRUs. For this article, AMR will refer to first generation one-way meters that transmit data to pole top collectors, which, in turn, send the information back to a server for processing. An AMR system, when integrated with OMS, provides for automated reporting of customer outages using the “last gasp” capability of these hybrid meters. “Last gasp” capability is the meter’s ability to send out a signal that indicates the meter is no longer receiving utility power and is operating on its battery. An OMS can automatically determine if a customer’s meter matches a specific outage report followed by a specific outage status.
When AMR is added into the mix, we generally find that overall reliability statistics appear to degrade initially. This is a direct result of having potentially more accurate time reporting at the beginning of an outage. There are many reasons why the outage clock doesn’t start ticking immediately. One example would be when customers are away from home and don’t realize that power is out at their home until they return to the residence and then report the problem.
The AMR system can be an effective tool for outage restoration verification where meters are “pinged” (interrogated) to determine their energized state. This "pinging" provides an automated capability for systematically verifying power restoration at the customer site without having to contact the customer directly.
Direct utility experience
In KEMA’s review of two utilities’ major storm restoration activities, we came across the same fully deployed AMR system. This AMR system is an early generation system employing one-way meters with a “last gasp” capability. The data back haul for this system is a series of pole top collectors, which capture the individual meter data and send it back to the AMR company’s host servers. In the first company, which does not have an OMS, the AMR tool operated independently of a restoration system as described above. The second company uses the same AMR solution integrated with the OMS. Interestingly enough, both companies had the same results but for different reasons.
In the first company (a million plus customer combination utility) 700,000 of its customers lost power due to extensive damage to both the transmission and distribution systems during back-to-back storms. Many of the pole top collectors were also rendered unusable. Complicating matters, entire circuits were locked out for an extended period of time (well beyond the backup battery life of both the meter and the pole top collector.) As a result, the AMR system was unable to provide any useful information to the emergency operations center (EOC) management. EOC management had to shut down the entire system until late in the restoration effort (several days later).
Near the end of this 10-day outage, EOC management again tried to use the AMR system to identify which customers were still out (service failures) with little success. Determining meter status was accomplished by pinging a meter and waiting for a response or by signaling a group of meters to read themselves, in which case a no-read would indicate that power was still out. Management was unable to ping groups of meters since it would take hours to get a response. Individual pinging would also take too long to complete when there were over 30,000 meters. For these reasons, EOC management ignored the AMR solution in favor of the conventional method of calling back customers. Our team learned that this was not the first time EOC management had problems with the AMR technology. As a result, management generally doesn’t use the AMR during major emergencies.
Specifically, our KEMA team found the following issues with this application:
In discussions with the company and AMR provider, our team learned that the system was designed to read meters for revenue purposes. This means that the configuration was not optimal for restoration support.
The meter groupings followed the traditional meter reading books that are inconsistent with the physical feeder-lateral relationship with the meters.
With the number of poles lost during the storm, many of the pole top collectors were rendered unusable. There was no backup means for capturing the meters’ “last gasp” signal when its assigned collector failed.
The second company (a large combination gas and electric utility) uses the same AMR technology but links it to its home-grown OMS. During small events the technology works well. It provides solid input to the OMS, which in turn provides direct input to first responders via an on-board computer using MWF. It is during medium- and large-scale events that the system does not work. There were two events in 2006 that uncovered problems with this AMR solution. The first event left 650,000 customers without power and the second left 270,000 customers in the dark. During these events, the following issues surfaced with the AMR system:
An important issue was identified during feeder lockout situations. Within seconds of a feeder lockout, the SCADA system would update OMS with the event. System dispatchers would then react by trying to reclose the circuit, many times successfully. In fact, 12 to 17 minutes later the AMR system would report that all the customers on that feeder were out. The AMR data was not time stamped, which confused dispatchers since the SCADA system reported normal operation. Management determined there were several screens the meter data must pass through to ensure that that there were no false readings. These screens, combined with the normal processing of the meter data, added a delay of 12 to 17 minutes in getting the data to OMS.
During the two major storm events, the AMR technology collapsed under the shear volume of “last gasps” received in a very short period. The meter provider’s servers could not handle the volume of data being received and the data collisions that resulted. This forced management to shut down the system relying on customer calls and SCADA initiated data.
As stated earlier, this AMR solution is an older, one-way system that was designed to read meters first and provide outage information second. The AMR provider indicated that the system was not configured to fully support “last gasp” functionality of meters other than the prime mission of providing highly reliable consumption usage readings.
Observations
While both utilities use the same AMR technology, they process the information in very different ways with the same end results for major outage events. Both management teams, at the beginning of the events, had to abandon the systems in favor of the more conventional use of customer calls and SCADA data. Compounding this situation, they were unable to effectively use the system to evaluate the customers that remained out near the end of the restoration event.
The newer AMI technology with two-way communications may help alleviate some of these issues. However, there is still a concern that the shear volume of data received during these events could continue to render the information useless unless a protocol is developed to better manage the volume of data received. Further, the data back haul needs to be strengthened to address the potential loss of field collectors, which is inevitable during major storm events.
Glossary
Advanced Metering Infrastructure (AMI): Automated meter reading, meter data management, meter “last gasp” outage reporting and processing, and automated remote interrogation (pinging) of meter for power restoration verification.
Customer Information System (CIS): Managing information about customers, customer services, metering and billing with supporting Interactive Voice Recognition Unit (IVRU), web posting and other customer and public communications.
Customer Service Representatives (CSR): the call takers in a call center.
Geographic Information System (GIS): Detailed geographic mapping of utility transmission and distribution facilities and equipment, network connectivity, equipment information and field configuration.
Interactive Voice Response Unit (IVRU): In the context of outage management, the IVRU routes calls to CSRs and allows customers to self-report and receive outage information.
Mobile Workforce Management (MWF): Automates field crew operations with mobile workforce dispatch, scheduling and routing, remote electronic connectivity and automatic vehicle location.
Outage Management System (OMS): Managing trouble tickets, outage analysis and assessment, crew dispatch and restoration process
Systems Operations Supervisory Control & Data Acquisition (SCADA), Energy Management System (EMS) and Distribution Management System (DMS): Real-time monitoring of the electric transmission and distribution network, energy supply, equipment operating status and remote switching and control.
Work Management System (WMS): Work order processing and management, resource assignment, job status and completion tracking.
About Automation Insight
Automation Insight is a complimentary monthly publication designed specifically for the utility industry and those serving the utility industry. For comments or suggestions on future article topics, please e-mail automation.insight@kema.com.
To receive Automation Insight by e-mail, please e-mail automation.insight@kema.com with 'subscribe' in the subject line.
Automation Insight is an opt-in subscription. KEMA does not sell or otherwise make public subscriber information and honors all ‘unsubscribe’ requests. To unsubscribe, please e-mail automation.insight@kema.com with 'unsubscribe' in the subject line.