The many benefits of distribution automation (DA) - visibility, fault detection and isolation, energy efficiency, and asset management - are creating a “second wave” of smart grid investments and integrations, following the widespread adoption of advanced metering infrastructure (AMI). Currently, in fact, the business case for DA is better than for any other single system in the phased steps of grid modernization.

In those phased steps, typically AMI comes first, followed by DA. In fact, DA relies on AMI’s end-of-line sensors, a.k.a. “smart” or interval meters, to enable its benefits. The hitch in this picture lies in the fact that the fundamental AMI system must accommodate DA functionality.

Utilities that adopted AMI as an end in itself, without a well-considered technology road map, may be rudely surprised to learn that their AMI choices of a few years back were not made with future integrations in mind. We need an industry standard that spells out the architecture of communications infrastructure within the meter so it can send ‘last gasps’ to the outage management system (OMS), voltage data to the distribution management system (DMS) and serve other functions to systems other than AMI. But because there is no such industry standard with respect to AMI systems, the endeavor remains very much a caveat emptor situation.

Some AMI systems simply do not lend themselves to DA integration and may require replacement or a laborious, expensive, inefficient workaround. For utilities that have not yet embarked on an AMI implementation, looking ahead to future systems integration can avoid duplicative efforts and costly mistakes. In fact, a successful DA integration with AMI unlocks the value in both systems. In addition, the creation of a technology road map and adoption of these and other technologies should drive organizational change toward a more holistic approach to smart grid. De-siloing will bring efficiencies and further unlock the value in technology adoption, something that regulators will increasingly demand as they scrutinize cost recovery and rate cases.

The market and the business case
Of the approximately 48,000 distribution substations in the U.S., fewer than half have any sort of automation. Substations with some automation and those without automation typically connect to feeders with no automation or monitoring whatsoever. Today, very few distribution feeders send any kind of real-time information upstream. This creates large areas of, shall we say, “unobservability.” We just don’t know what’s happening on the system.

The imperative to learn what’s happening on the system will only grow stronger, and the need to support a variety of data streams from the field and route them efficiently is only going to grow exponentially in the near future. As more distributed, renewable energy is integrated into the grid, and as the utility copes with two-way power flows, the utility will face new safety and protection challenges. Add to that the additional, two-way data flows that will accompany dynamic pricing and the interaction of that signal with a home energy management system. When the peak price of electricity moves customers to shed load, the utility will want to understand precisely how much load is being shed, individually and in aggregate. Ideally, these data flows, like those for AMI and DA, would use the same communication network: the AMI support infrastructure.

All of these considerations are driving utilities to implement a DMS to manage complexity. But the DMS is only as good as the information coming from the field. These factors explain why distribution automation or distribution optimization, if you will, currently represents the most cost-effective step and the best business case of all smart grid solutions.

The substations and feeders without automation in the U.S. pose an enormous challenge - not coincidentally, a huge addressable market for vendors and a cost-effective route with big operational and organizational payoffs for utilities. The thinking I’ll outline here is designed to assist both parties in making the best products and the most cost-effective investments.

First, break down the walls
The holistic approach to smart grid, in particular, and grid modernization, in general, requires a strong dose of executive leadership due to entrenched interests that have persisted in power utility culture. For instance, AMI implementations typically fall under the purview of a metering group inside the utility, while DA is under a distribution engineering group in operations. The two systems share a need for service-territory-wide communications systems. Too often, a siloed utility builds two systems, side by side, when a single, well-vetted system could be built to serve both purposes. That results in redundant efforts, duplicative expense and two separate data streams that would serve the organization far better if they were integrated.

The simplest way to avoid this misstep is to have executive leadership, sometimes aided by a third party, bring together the metering group and the distribution engineering group to jointly determine their mutual, functional requirements for a common communication network. And don’t stop at accommodating DA functionality, because as I’ve mentioned, demands on that network will only grow with time.

Cooperation leads to a stronger business case for both systems in this example. Indeed, a general rule of thumb for a technology road map and resulting utility investments is to develop them with a horizontal organizational structure that results in cost-effective investments and integration-friendly systems. As this becomes a more widely recognized best practice in the smart grid era, regulators will come to expect this approach and conceivably may base decisions on whether it’s being implemented.

Integrating the acronyms
Many AMI technologies are designed for meter-related data output only - those 15-minute interval readings that flow upstream to the network management system (NMS), which manages the communication network aspect of AMI and also feeds the data to the meter data management system (MDMS). The MDMS stores that data and feeds it to applications, such as generating customer bills, analyzing usage patterns and so forth.

The interval meter’s so-called ‘last gasp’ when an outage occurs isn’t metering information; that signal needs to be routed to the OMS, where it can be analyzed to determine the cause and extent of an outage. Some AMI systems cannot split off that last gasp to the OMS. Similarly, another distribution automation function - voltage data coming back from the end-of-line sensor, in this case the meter, needs to be routed to the DMS to ensure that the utility is achieving the 114v to 126v ANSI standard at the customer premise. That is not easily accomplished with some AMI systems. Note that one doesn’t need the voltage readings from every meter, just those at strategic points at the ends of selected feeders.

An AMI system is the glue between the meter and the utility. Functionality in the meter needs to be matched to functionality in the supporting systems, the ‘infrastructure’ in ‘advanced metering infrastructure.’ That means the communication network, among other things. Thus, an AMI system needs a certain flexibility to integrate properly with DA functions, such as routing meters’ last gasps to the OMS and steering voltage information to the DMS.

For utilities that have installed AMI, this underscores the need to evaluate the underlying systems with DA integration in mind. A utility may have had the foresight to develop a carefully thought-through road map and be in a good position to reap the benefits of DA. If that foresight was lacking, the consequences can be laborious and expensive. It’s technically true that AMI data can be routed through the NMS and the MDMS to reach the OMS and DMS, but that’s a cumbersome route that challenges bandwidth and latency. A well-architected system would avoid that scenario.

Further, as meters gain functionality, they may well be upgraded or swapped out for more advanced ones. What a utility wants to avoid is ripping and replacing the underlying infrastructure - again, the ‘I’ in ‘AMI.’

The AMI system needs to have enough flexibility to support the metering information going to the NMS and MDMS, but also support other data outputs on the smart meter and be able to route that to other systems. I’ve mentioned routing last gasps to the OMS and sending voltage data to the DMS, but as time goes on, deriving value from more functionality in the meter requires having the functional-ity to route those data streams to other systems and destinations over a common communication infrastructure.

Vetting the DMS
The DMS contains the network model manager, which is a critical piece of software. Utilities would be wise to look closely at this functionality during the procurement process. The DMS must interface with the utility’s geographic information system (GIS), so it is imperative to know whether the particular DMS in question will, in fact, integrate easily with the utility’s particular GIS. The DMS must know what data to pull from the GIS, how that information is stored and how to retrieve the needed data for building the network model.

A DMS that works well with a GIS is important because, as the data in the GIS changes, incremental updates inform the network model in the DMS and keep it up to date. The OMS also has a network model for outage analysis that depends on the GIS, as well.

The network model
Think of the network model as two major sets of information. One is the power system connectivity information, which includes the electrical characteristics of grid assets. For instance, that includes transformers, the model of each transformer and its connection information - is it YY, is it ΔΔ, is it ΔY grounded? Power system connectivity information also includes the branches, nodes and capacitor banks connected on the distribution feeders to ground.

The second set of information in the network model is the real-time information about the network, the operational information - the voltage, current, real and reactive power flows, statuses of switches and circuit breakers, and so forth.

DA functions, up close and personal
When we talk about DA, we’re talking about three primary functions: improving reliability with fault detection, isolation and restoration (FDIR) for optimal feeder reconfiguration; reducing losses with VAR control; and managing load or demand with voltage control. (Voltage is directly proportional to load, so when we control voltage, we control load. VAR is a reactive power, directly proportional to losses.)

Today, with DA, the utility can combine voltage and VAR control with integrated volt/ VAR control (IVVC). In fact, a DMS optimizes these applications. But to do so, the DMS requires real-time information, knowledge of what’s happening on the distribution system downstream of the substation.

To assess whether an AMI system will support DA functionality, one needs to weigh the response re-quirements of the DA applications. Three metrics must be assessed: speed, bandwidth and latency. For instance, FDIR requires a 2-3 second response for rapid switching. (Those are SCADA-level speeds.) Capacitor controls require about 30-60 seconds. Many AMI systems are designed to support only 15-minute interval reads, yet intelligent electronic devices often need to send megabytes of data upstream at one time, requiring speed, bandwidth and low latency.

Here are some questions to ask yourself: Can your utility countenance delay in operational com-mands being enacted? Are hundreds of milliseconds of latency tolerable? There are other considerations: Cybersecurity practices such as encryption affect the performance of data communications, increasing latency. Seeking the 200 millisecond latency one is accustomed to while adding ‘overhead’ in the form of security measures may not be realistic.

When the utility adds sensors at both substations and feeders, much more information heads upstream. That has an impact on the system’s ability to meet the response requirements of DA applications, in terms of speed, bandwidth and latency.

Avoiding the abyss, and stranded assets
The utility needs to ask hard questions of its vendors to avoid the downside described here - making a short-sighted investment in AMI.

Is there a migration path with your vendor? What’s on that path? An easy ‘board swap?’ A more difficult, more expensive ‘box swap?’ Maybe there’s no swap; maybe it just doesn’t exist. If there’s no path forward, will that result in a stranded asset? In SCADA procurement, for instance, if a vendor said it would be supported by a top-end X-brand server in the family of servers, then if you need greater computing capacity - and that’s a given - the system has no way to grow if it’s based on the current top-end equipment.

Thinking through your technology road map with a good understanding of succeeding systems’ functional requirements will lead to better results and more cost-effective investments. Hopefully, this exposition of the technology challenge in integrating DA with AMI and the importance of the road map contributes to better choices for stakeholders.

About the Author

John D. McDonald is an IEEE Smart Grid technical expert, as well as Smart Grid Business Development Leader – North America, Global Smart Grid Strategy Group, at GE Grid Solutions. He is an IEEE Fellow, past president of the IEEE Power & Energy Society (PES), an IEEE PES distinguished lecturer, board chair of the Smart Grid Consumer Collaborative and CIGRE US National Committee (USNC) VP, Technical Activities, among other affiliations.