There are many definitions of what constitutes a smart grid and many visions for how a more intelligent grid will enable the future energy economy. While these visions vary, they also have many characteristics in common—characteristics such as distributed intelligence, adaptive self-healing, and multi-way communications across the entire energy delivery chain. Some approaches to smart-grid design involve deployment of high-performance line sensors, while others rely upon use of the advanced meter infrastructure (AMI) to provide visibility into distribution grid state. While smart meters don’t provide support for the highest-performance smart-grid functions, they can provide significant capability when the AMI system is properly designed to support the evolution to a smart grid. Proper design for this purpose implies modifications to several aspects of the AMI system, including types of meters, specifications for the meter communication system, and design of the meter data management solution.
The question, then, is how far can smart metering take us toward realizing a smart grid, and what factors should a utility consider in the design of a smart meter system if the eventual goal includes smart grid?
A smart grid uses sensing, embedded processing, digital communications and software to integrate grid-derived information into utility processes and systems, thus making it observable (able to measure the states of all grid elements), controllable (able to affect the state of any grid element) and automated (able to adapt, self-adjust and correct). A smart grid supports the three main pillars of utility function: 1) delivery of reliable, high-quality, sustainable energy: 2) asset utilization optimization and asset lifecycle management; and 3) advanced customer services and choice enablement.
Given this definition, an AMI system can supply some, but perhaps not all of the capability around grid observability. Since residential meters are located on the low-voltage grids, they do provide some visibility to quantities such as voltage, but do not give us visibility into voltage and current phasors. And, while residential meters may have the capability of capturing some details of transient phenomena such as voltage sags, they may not capture proper parametric descriptions of these events. They also provide limited visibility into real and reactive power flows on the main feeders, something that becomes increasingly important as the penetration of distributed energy resources increases. Aggregation or roll-up strategies for computing power distributions from premise meter data have been suggested, but most AMI systems do not support data transfer at rates necessary to make this a viable alternative to sensing at the medium-voltage feeder level and most residential meters do not measure reactive power, although many commercial and industrial (C&I) meters do. New developments in metering, such as feeder meters, offer potential solutions to these limitations.
If a utility is going to start with smart metering to build its smart grid, how much capability will the smart meter system provide and will it make the grid smart enough? And how much capability will be needed to get there? Different combinations of meters and sensors produce corresponding levels of grid intelligence enablement (see Figure 1). The choice of what combination of technology is right for a utility depends on the degree of intelligence desired.
When using AMI as a smart-grid entry point, it’s important to understand and plan for the communication system underpinning the smart grid, as well as the architecture of the data-handling system. Doing so will enable these components to be better able to support smart-grid functionality when an organization is ready. These requirements will change the economics of the AMI deployment because the utility will be building in future capability. And, planning for these requirements may mean that a utility finds itself implementing some capabilities or capacities that are not strictly required for the meter-only system. Yet, such planning is imperative; incorporating the capacity to support new capabilities down the line may prove too costly to implement.
When thinking about building a smart grid from a meter system, a utility needs to think about specifying the communication system in order to accommodate the speed requirements of a smart grid, as opposed to just the residential meter system. Meter systems generally don’t require high performance if they are to be used for the meter-to-cash process only. If the goal is to use the meter communication system as the basis for a smart-grid implementation, then it’s wise to consider the range of smart-grid functionality to be supported, the associated technical requirements and performance factors—such as bandwidth, latency, burst/flood response and average response time as a function of number of active nodes. Otherwise the utility runs the risk of either stranding an expensive asset, being forced to build additional communication assets that could have been avoided, or effectively being cut off from some types of smart-grid capabilities. If the meter communication network also will be used to carry control messages for grid devices, then significant cyber-security concerns must be addressed as well.
Another issue that should be considered at the AMI planning stage if the goal is to also support smart-grid capability is the system connectivity model. If the utility wishes to use meters as sensors, it will need a complete connectivity model to provide context for interpretation of grid sensor data. This model includes meter-to-feeder phase connectivity (via the distribution transformer). The problem with using AMI for smart-grid support is that utilities may not have good connectivity models. In fact, utility connectivity databases may be only 50 to 80 percent accurate, and many have no meter-to-phase information at all. To use the meter system as a sensor network, this information must be corrected (or obtained if it never existed) and kept up to date.
Capturing meter-to-phase information can be extremely expensive, especially if it must be done after meters already are rolled out. It’s much better, therefore, to capture this information during smart-meter deployment. Once established, the connectivity data must be kept accurate, which typically involves additional new utility processes because meter connectivity can change over time—when transformers are switched from one phase to another, for example.
A host of devices are available today to produce data from a power grid, all of which can be used to help build a smart grid and enable the observability capability. These devices include substation devices (microprocessor relays and associated tranducers), line sensors, smart line devices (switches, reclosers and capacitor bank controllers that can also capture measurements and waveforms), faulted circuit indicators (simple devices for fault detection/location), and various kinds of meters.
Many strategies for how to deploy and use sensing capabilities on a distribution grid are possible. The determination of which sensor types, where to locate them, how they should communicate, and how to manage the data they can produce ultimately depends on requirements for the smart grid that each utility must derive from business needs. Depending on business requirements, quite a range of technical requirements are possible, with significant implications for smart-grid communications in terms of bandwidth, latency, and related characteristics. Consequently, it’s always important to have a clear definition of the specific smart-grid capabilities to be supported before selecting sensors and related infrastructure. Such requirements definition then will make it clear as to how much the meter system can support smart-grid functionality.
If the meter system is to be used as an element of the sensing strategy for a smart grid, then the types and locations of the meters in the AMI system should be included in the process of allocating grid sensors for the smart grid. The sensor allocation process can consider capabilities of residential meters, the more advanced capabilities of C&I meters, and the capabilities of various types of grid sensors to determine number and placement of smart grid measurement points.
The observability of a smart grid through the meter system might be improved by including some number of another type of meter—the feeder meter. These devices attach to medium voltage (MV) feeders, but operate much like C&I meters in terms of communications, interface to meter communications systems, and interface to meter data-collection engines. Such devices have revenue-grade accuracy ratings (0.2 percent), and can provide advanced measurements similar to those typically available from C&I meters. The key here is that they measure parameters on MV feeders, and measure operational data including power flows, energy, and various operational and non-operational parameters at the MV feeder level without having to use communication and data-collection infrastructure separate from that of the meter system.
In this approach, it’s necessary to extract the data generated by these meters and to provide that data to various applications besides the standard meter-to-cash applications. Given this need, the data management and integration architecture to make use of the feeder meters necessarily will include some elements that would not be needed for a standard AMI system. Also, since most of these devices will generate asynchronous event messages, the system must provide the means to process these messages, which often occur in bursts and floods.
Commercial and residential meters have their shortcomings in capturing the array of electrical parameters needed for a smart grid. Feeder meters can help mitigate this weakness. To incorporate these meters into a smart-grid configuration, they can be installed at selected points on MV feeders, and connect to a utility’s standard meter system, communication system or data management system.
Feeder meters can offer additional benefits, one of which is enhancing diversion detection. A utility may create a comprehensive diversion detection solution by combining data from premise meters and feeder meters. Sophisticated data processing of the measurement of power flow into, and out of, a feeder segment bounded by feeder meters, and premise-metered consumption from premises attached to the same segment, can provide diversion detection, diversion amount, and total segment technical losses. By localizing to a segment, diversion investigations can be greatly accelerated. Voltage-drop analysis at the premise meters may further localize the diversion.
Another benefit of feeder meters is as a tool to assist fault detection and location. Smart line sensors can be used to detect, classify, and locate faults, but they are expensive, and not every utility will want to deploy them in large quantities or provide a high–performance communication network for them. Although feeder meters won’t capture information as fast as line sensors or operate at the same high speed (sub-second) rates necessary for grid protection and control functions, they can be used to perform some of the functions of line sensors, such as support for outage management, dispatch, and field service.
Feeder meters also can be used to help in the management of the grid in the presence of distributed generation (DG). DG and storage encompass emerging technologies that allow placement of energy sources at various points on a grid, including at customer premises. The presence of technologies will change distribution grid power flows in ways the grid was not designed to handle. Significant penetration of DG into a power grid introduces new complexities in grid-power flow, volt/VAR regulation, and optimization and power-quality control. As penetration of these technologies on a distribution grid increases, eventually a tipping point is reached where it becomes necessary to apply smart-grid technology to help restore manageability of the distribution grid, primarily by improving observability of the grid. Meters also may be employed at the point of common coupling for distributed generation to provide visibility on DG power flows and impact on power quality.
How far can smart metering take utilities toward the realization of a smart grid? The answer depends on what meters and sensors a utility decides to implement and the degree of intelligence desired. But the options with meters and sensors already possible can be quite sophisticated. The emergence of new meters, sensors, and other technologies is creating opportunities now to rethink traditional approaches to commercial and residential meters and consider more powerful ways to plan for adding smart grid capability—for today and tomorrow—in the continuing journey to high performance.