Edge Computing May Be the Future of Power Distribution

A more distributed power distribution network will help utilities better navigate the wave of changes resulting from the smart grid. Very few industries will experience the rapid change that power distribution operations have seen in just the last five or 10 years. Traditionally, the power distribution system was a very passive network, as moving power from generators to end users could easily be controlled from a central location in relatively predictable patterns. Even when challenges arose, centralized control facilities could rely on a deep bench of expert human operators with decades of expertise and experience to mitigate those challenges. Today, however, the dynamics of distribution are evolving rapidly, as power generation simultaneously changes. Predictable patterns of generation are shifting to embrace new, more varied solutions, and with that shift comes a need for more granular operation, increasing grid management complexity and cost. Much of this complexity and cost can be navigated more easily with automation, so, ultimately, the world of power distribution is likely headed for more closed-loop control. The most forward-thinking utilities are already exploring how they can incorporate closed-loop operation by considering ways to build more distributed control into their networks through expanded edge control. The journey is just beginning, but it provides a glimpse into a very exciting future.

New Trends Generate New Concerns

Traditionally, the grid consisted of somewhat isolated generation, transmission, and distribution. Power was generated by large, rotating machines, and it was then transmitted over long distances via an established transmission system. Once the power reached its destination, it was stepped down to lower voltages to move across the distribution system. The system was relatively simple, and primarily unidirectional, from generation to use point (Figure 1).

1. Traditionally, the grid has been relatively simple and primarily unidirectional, with energy distribution centralized. Courtesy: Emerson

As a result of this relative simplicity of operation, control of energy distribution has mostly been centralized. Control actions operated over very long time horizons and were relatively predictable. Typically, changes occurred based on the time of day, day of the week, or day of the year. End users used electricity in the morning, then went to work, causing the load to decrease. When they came home again, they turned on the lights and their television, and started cooking, and the load went up. With some allowance for variability due to heating and cooling changes between seasons, and differences on holidays, the system was stable and reliable. As the world has shifted its focus to dramatically increased reliance on renewable energy, however, the existing power dynamics have changed. Today, distributed energy resources (DERs) are being added to the distribution side of the grid at a higher rate than bulk energy generation is being increased. Those DERs are typically renewable generation—most often solar—so they are reliant on the whims of the weather. Not only does generation from those assets, and their associated load on the network, change day to day, it can even change hour to hour. This increased dynamism presents unique control challenges that are difficult to solve through traditional operation.

2. Grid dynamics are changing with increased renewable and distributed energy resources introducing more complexity, and with bidirectional power flows and decentralized distribution. Courtesy: Emerson

Moreover, DERs are not the only source of variability changing the modern grid. As end users add DERs, new energy availability increases the desire for other electrical assets, such as electric vehicles (EV), that will further transform the operation of the distribution network. Today, utilities can easily see thousands of EVs moving through a particular geographic area. Those EVs charge at different times and different rates, so their demand no longer follows the well-defined curves operators are used to managing (Figure 2). Adding further complexity to this changing system is the evolving Federal Energy Regulatory Commission’s Order Number 2222, which allows DERs to participate alongside traditional generation sources in the wholesale power markets. The order is only in its earliest stages and its full impacts are yet to be seen, but aggregators are already responding by bringing DERs together as virtual power plants and bidding them into the power market. To meet these and other impending changes, utilities will need to manage generation on the distribution side, following market signals alongside the inherent intermittency of DER generation.

Navigating Change Is Increasingly Difficult

Currently, most, if not all, voltage control and power flow control is managed in a centralized control facility using its advanced distribution management system. These control centers rely on connections to all the equipment in the field to inform operators of variations in the system. The operators, in turn, react to the effects of those changes, adjusting setpoints and balancing loads. The increased complexity of control for distribution networks not only requires very experienced personnel, but it also requires lots of them. However, this change is occurring just as utilities are being forced to do more with fewer people. Retirements are sweeping the industry, and no control center can count on still having operators who have been around since the foundation was poured. Moreover, hiring is difficult due to a shortage of available experts. In some instances, organizations can train new people, but bringing new personnel up to speed takes years, and modern workers are only staying in one job for an average of two to three years. In many cases, just as they come up to speed, new personnel move on to other opportunities. As a result of this shift in the modern workforce, utilities need a way to scale operations without needing to scale their expert personnel in tandem. The solution is to build an infrastructure that empowers the personnel they already have. Forward-thinking organizations will begin exploring new control hierarchies to provide users with the tools they need to do more by ensuring constant access to the data necessary to make good decisions. As they do this, they will also be incorporating closed-loop automation to lock in best practices and reduce reliance on manual control.

Improving Operations at the Edge

One of the most prominent new technologies ushering in the era of the smart grid is grid edge controllers. Grid edge controllers expand control options beyond the control center and into the field, where equipment like voltage control systems, breakers, DERs on the grid, or even EV chargers must be monitored and controlled. The most powerful solutions will likely reside in substations, or even on the DERs themselves, in the former case interfacing with the central distribution control facility, and in the latter case to a controller in a substation. The central control facility will not disappear, however. Rather, many of the functions traditionally performed centrally will now be distributed among grid edge controllers positioned strategically across the distribution system. These controllers will perform high-speed control functions under the supervision of the centralized control facility. For example, consider managing today’s more dynamic performance of voltage regulators and/or reactors in the field from a central location. The data from those devices would need to flow up to the control center, human operators would need to analyze it and make a decision, and then the results of that decision would have to propagate back down to the actuation systems. This process is quite slow for the needs of the modern smart grid, where electrical power flows change near instantaneously. Alternatively, a grid edge controller could be deployed to automatically manage voltage regulators or reactors based on current operating conditions in the field. The controller could identify what the voltage and the power flow need to be, and then dynamically control setpoints on the devices. Such controllers could make these adjustments automatically, then report them to the control center for monitoring by a core group of personnel. Instead of controlling everything centrally, the organization will have created small grids that feed into the main grid. All the coordination, orchestration, and topology visibility would still occur at the central control facility. However, instead of all those control functions running in one huge central control system—and consuming bandwidth to move that data back and forth—the computing power will be distributed across numerous smaller computers spread across the grid, each of which can provide high-speed feedback control.

With Change Comes Benefits

With the addition of grid edge controllers comes the potential for a more granular view of the grid. Much of the data at each substation could be condensed in the edge controllers, with aggregated information, instead of just raw data, sent to the control center. The net effect would be much better visibility into what is happening on the distribution grid, improved ability to control the many connected devices, and much faster speed of reaction (Figure 3).

3. Grid edge control is a prominent new technology that provides a more granular view of the grid with high-speed, localized field control functions that are supervised by a centralized facility. Courtesy: Emerson

A more distributed architecture also increases grid resiliency. When all operation relies on a single control center, utilities have a single point of failure. Even with redundant control facilities—a costly option—operators risk severe outages when something goes wrong. Individual edge controllers distributed across the grid provide individual sectors with some degree of autonomy. In the case of catastrophic failure, such as increasingly common weather events, some sectors could run autonomously to maintain basic operations until control is restored. Additionally, distributed control offers the option of more localized analytics. With edge controllers closer to field installations, teams can run much higher resolution analytics on equipment. Instead of streaming high-resolution data up to control centers, machine learning models can be run right at the source, making decisions and aggregating data—enabling faster control strategies to respond to more short-term events, while simultaneously reducing the required communications bandwidth, and its associated cost.

Preparation Is Possible

Though power distribution has changed dramatically in the last decade, it is increasingly likely that those changes will pale in comparison to the ones coming in the next five to 10 years. Utilities know they will need to make changes not only to compete, but even to operate efficiently in coming years. While not all the necessary changes are obvious today, a trend toward more distributed control seems increasingly likely. Today’s pilot programs using edge control may very well prove to be the foundation for tomorrow’s most critical and effective operations. Fortunately, the technologies exist to begin those pilot programs today and stay ahead of the curve. The companies that do so will likely be the ones to most effectively navigate the smart grid revolution.

—Rick Kephart is vice president of technology for Emerson’s power and water solutions business.