by John McDonald, GE Digital Energy
Reliability will be the top priority for grid operators no matter whom or where they serve. From a tree branch that takes out a single home to a complex, multifactor event that cuts power to 700 million people as a two-day event did in India this past summer, rapid outage detection and speedy restoration play critical roles.
Regulators assess utilities’ performances partly on reliability indices sometimes referred to as “SAIDI (System Average Interruption Duration Index) and her sisters”: SAIFI (System Average Interruption Frequency Index), MAIFI (Momentary Average Interruption Frequency Index) and CAIDI (Customer Average Interruption Duration Index), which are challenging to keep these numbers within acceptable limits. These indices are based on IEEE Std 1366.
As intelligence is extended down into the distribution system, much data can be generated if used properly by an integrated system of outage detection and power restoration and can help utilities improve on critical indices. This can be done using available technologies. In addition, GE Digital Energy recently developed a patent-pending software application that can monitor social media to provide speedy, granular data to assist these efforts. The attractive secondary benefit is customer engagement and satisfaction.
These developments and their application can provide utilities with a clear path through the smart grid woods and maintain their focus on the most important factor guiding their work: power grid reliability.
In describing how to improve outage detection and power restoration, there is an integrated approach to distribution automation that takes advantage of the strong business case, such as fault detection, isolation and restoration (FDIR) and integrated volt/VAR control (IVVR). For utilities that have installed advanced metering infrastructure (AMI) and interval smart meters, this is a logical step; however, AMI isn’t necessary to reap the benefits of the approach outlined here.
Remember, automated substations that can trigger switches to isolate faults without operator intervention still leave the utility to rely on customers to call about outages. That’s why advocating a more integrated approach to full distribution system automation will improve multiple benefits for the business case and utility performance on outage-related indices. Distribution automation’s business case improves markedly if the driving focus is speedy outage detection and power restoration.
Replacing the Phone Call
Smart meters record electricity usage for billing, measure end-of-line voltage and, in the case of an outage, emit a last gasp as they lose power. Capacitors are designed to hold enough charge to get that crucial message out. That’s almost instant, more precise and quicker than waiting for customers to call—sometimes the difference between an outage’s falling under MAIFI: a momentary outage; and SAIFI: a sustained outage.
For utilities without interval meters and AMI, voltage-sensing meters can be placed strategically at the ends of feeders to ensure compliance with American National Standards Institute (ANSI) standards for delivering 114 V to 126 V of power. Those meters can play a role in outage detection, though a less granular one than full AMI metering. That’s where GE Digital Energy’s new software application for social media will be useful.
A traditional customer phone call could be linked to a physical address by tapping into the customer information system. With the widespread use of social media, we can link customers’ Twitter messages to an address to obtain similar information. This can be done in various ways. The utility could incentivize customers to link their Twitter tags to their account information or to turn on their mobile devices’ geo-tagging function, which provides latitude and longitude that indicates their locations when they tweet instead of call. A cluster of tweets tied to addresses—the greater the sample, the greater the accuracy—can be subjected to automated analysis to provide the precise location and extent of an outage. Grid IQ Insights from GE Digital Energy is a new software platform with geospatial coordinates for automated systems that can connect tweets with an outage management system.
Getting customers to work directly with their utilities takes time and effort. Customers are more likely to tweet one other to complain of outages. The new software application uses text mining to understand whether a flurry of tweets that mention “outage” and “power” really refer to a “power outage,” and if so, can mash up that data with end-of-line sensor data to identify the location and extent of an outage rapidly. Because each household has multiple members and an electricity account holder might not be the one tweeting, this method provides an attractive alternative for utilities without AMI. The analytics software that determines the location of outages in both instances resides in the utility’s outage management system (OMS) platform but remains a separate function.
Replacing Truck Rolls
Crowdsourcing the cause of an outage by leveraging customers’ use of still and video camera capabilities on ubiquitous mobile devices is another social media benefit to utilities. Incentivizing customers to post their pictures and videos of downed lines to a utility’s Facebook page while emphasizing traditional cautions about approaching high-voltage situations could provide images to be analyzed by automated image-assessment software to better inform field crews prior to truck rolls.
Even where AMI is installed, there’s only one meter per household. Data show that each household has multiple social media accounts, potentially multiplying the number of resulting data points. The current 200 million Facebook accounts in the U.S. represent about one account per household, and that number is growing. Using social media, as parents know, often is preferred to making phone calls. An app could reduce outage notification to one click. As generations shift, these trends will become more pronounced.
We see in these trends immediate operational benefits tied to outage detection and power restoration plus the seeds of significant customer interaction and the larger field of general consumer engagement. Educating consumers about the economic and societal benefits of a smarter grid will be the first step in creating a smart grid social network. That network will function much like the smart grid—with open, collaborative, two-way information flow between consumers, the ultimate deciders of smart grid, and utilities, the ultimate providers of smart grid. That interaction will lead to customers’ education in and acceptance of a new utility-customer paradigm in which customers understand and willingly participate in critical utility programs such as demand response. That participation will signal a human smart grid, and it should reflect customers’ understanding of electricity’s role in the macro economy, as well as their own lives and prosperity.
Once consumers in general grasp that energy efficiency, demand response and active energy management have positive implications for their pocketbooks, economic security, energy independence and the environment, they will become the utility’s partners in achieving those goals. As utilities seek to defer new capital investment where possible, wring efficiencies from the grid and move from fossil fuels to more renewable and sustainable sources, they’ll find that an educated, cooperative customer is one of their most valuable resources. Utilities will learn to be customer-savvy, responsive and conscious of opportunities to add value to their commodity and its delivery.
Tying it All Together
Now we’ve got outage data, either through AMI or end-of-feeder sensors, plus social media. How that data is routed and analyzed is critical to speedier power restoration. A well-designed communication network will enable smart meters’ last gasps to trigger near-instant, automated switching. Without AMI, but by combining end-of-line sensor data and social media data, operators play an active role. Outage detection and power restoration will be speedier than relying on customer phone calls, especially if the applications run closed loop and remove the operator from the decision-making process.
Automation in this case has three components: a control center master, field equipment and a communication network. All three factors impact performance and cost.
The advent of public networks, mostly wireless, has provided the tipping point for the distribution automation business case. It’s finally cost-effective to provide a reliable data network across a large geographic area, whether that network is public or private.
Communication network performance is measured by three variables: response time, bandwidth and latency. Utilities’ performance requirements allow different response times for different applications, such as a switch’s needing two seconds, analog information’s needing 15-30 seconds and a capacitor bank can be triggered in 30 seconds. Bandwidth is measured in bits per second: How much data and how quickly can the sequence of events be reported? Latency refers to how much delay is acceptable in transmitting or receiving signals.
The network design must balance overhead on the system with the speed needed for various signals. For instance, cybersecurity and cryptographic requirements will introduce latency into the signal path; a 200-300 millisecond latency can become a 600-800 millisecond latency when cryptography is applied. Industry standard communication protocols (e.g., DNP3) introduce overhead, as well, and the communications system bandwidth might need upgrading to maintain the same update rate at the control center master.
Most utilities will employ a hybrid communication network. At larger substations, utilities might use fiber-optic cable or licensed wireless frequencies. Smaller, peripheral substations require only unlicensed spectrum for wireless radio. Downstream of the substation, the solution likely is a wireless private network, considering response time, bandwidth and latency requirements, as well as cost.
Another challenge is to integrate outage and verification messages from the meters or end-of-line sensors through the substation to the control center and its OMS.
AMI systems typically are designed to support only meters’ interval data, which travels from meter to the head-end system to a meter data management system (MDMS) for storage and analysis. The AMI is not designed for voltage data or the meters’ last gasps, which need to be routed around the AMI’s path and directed into a distribution management system (DMS) or an OMS. The DMS will use voltage data to populate a network model, and an OMS is the proper destination for last-gasp signals.
These separate paths reflect the distinction between operational data and nonoperational data and proper data routing around the AMI system is a nascent functionality that utilities must demand from vendors. Utilities should focus on operational data in this context, but vendors must enable utilities to extract more value from the nonoperational data coming out of IEDs, which will provide value to asset management, maintenance and power-quality efforts.
DMS in Outage Management
Consider the DMS’s role in an integrated system. DMS relies on a network model generated from geographic information system (GIS) data and is populated by substation and feeder intelligent electronic devices and voltage data from end-of-line sensors. A network model manager interfaces with a GIS so it knows what data to pull from the GIS to build a three-phase, unbalanced DMS network model.
A utility needs four applications on a DMS: the aforementioned FDIR and IVVR, optimal feeder reconfiguration (OFR) and distribution power flow (DPF). Protective relays detect a fault, its location and type. Then the FDIR isolates the faulted segment of the feeder and restores power to customers on healthy segments of the feeder using the OFR. An OFR can look ahead to account for switching schedules for routine maintenance to optimize its role. All this should happen in less than five minutes, keeping an event within the MAIFI index for most customers and not impacting SAIDI, SAIFI and CAIDI. Although IVVC is only incidental to outage management, it plays a big role in DMS, optimizing voltage and reactive power for energy efficiency. The DPF is an online tool that allows the operator to simulate the results of switching strategies and thus contributes to energy efficiency by controlling losses and loading on feeder lines. These two related functionalities make an outsized contribution to the business case for integrated distribution system automation. The suite of functionalities available through FDIR and IVVR comprise the best business cases for adding intelligence to the distribution system in the United States.
Advancements in data visualization tools such as dashboards make all the interactions described here graphically clear to the grid operator in the control center and, in circumstances requiring operator action, gives data in the form of actionable intelligence.
With all the aforementioned elements in place plus that crucial two-way communication with the customer, utilities can improve their reliability indices and begin to engage customers in a virtuous cycle that further contributes to speedier outage detection, power restoration and customer satisfaction. Utilities must weigh the larger value of increased reliability and its impact on customers, regulators and other stakeholders.
John McDonald is an IEEE Fellow, an IEEE smart grid technical expert, a past president of the IEEE Power & Energy Society (PES) and past chairman of the IEEE PES Substations Committee. He is director of technical strategy and policy development at GE Energy’s Digital Energy business.