Регистрация в MyAdvantech

MyAdvantech является индивидуальным порталом для клиентов Advantech. Став членом Advantech, вы можете получать последние новости о продукции, приглашения на вебинар и специальные предложения интернет-магазина.

Зарегистрируйтесь прямо сейчас, чтобы получить быстрый доступ 24/7 к вашей учетной записи.

Electric Grid Modernization Challenges: The Importance of Implementing Edge Intelligence for Active Grid Management


The North American electric grid is paramount to economic infrastructure and it’s the largest distributed electrified infrastructure we have. But, simply put, it’s old. With a changing world, the system is becoming more distributed and more complex, and the flow of electricity is becoming bidirectional. What is the path forward to handle it? How do we modernize this massive infrastructure?

The Urgency of Needed Change

The world of electrical energy has shifted dramatically in the past few decades, and this shift has been outpacing the modernization of the electrical grid itself, which was designed more than 100 years ago. This is causing a growing sense of concern and urgency around modernization of North American electrical infrastructure.

According to Intel’s Prithpal Khajuria, global segment leader for smart infrastructure, the electrical grid has become more bidirectional—which it was not initially designed to handle. The electrical grid was designed to carry electronics from utilities to consumers. Large power plants were built and placed in distributed locations outside of major cities with the intention of energy flowing “downhill” to consumers.

Our use of the system has now changed, becoming more distributed and more complex. The flow of electricity is becoming bidirectional, forcing the industry to rethink the electrical grid relationship. Khajuria explains in the past, it was clear who the producer was and who the consumer was. With a changing landscape those lines have blurred. In some cases, the consumer can be a producer too, which alters the traditional roles of producers and consumers. Further complicating the function of the grid is the growth of Green energy sources, which are often transient in nature compared to the steady, controlled output of traditional power generation plants.

Electrical consumption is also on the rise. Khajuria says our consumption of electricity is expected to quadruple in the coming decade as more infrastructural systems become electrified, such as mobile infrastructure, transportation systems infrastructure, and more. Commerce also keeps heading more and more toward digitalization.

“That’s going to drive the need for more data centers,” says Khajuria. Mike Fahrion, CTO of IIoT Solutions at Advantech, adds that current projections forecast anywhere from a 3% to 10% increase in data center consumption of electricity by 2030—an increase that will have major effects on our current electrical grid.

As the demand continues to goes up and generation continuous to be more distributed, there has to be work toward a system that sustainably generates and delivers that electricity. The electric grid is moving toward a “system of systems” architecture, and it will only be as successful and intelligent as its weakest point.

The Way Forward

To move forward and solve the challenges of a changing electrical grid landscape there are two options. There is always the possibility to build more infrastructure, but this is expensive, multifaceted, and would take a lot of time. The second—and best option—is to lean on new technologies, such as Internet of Things (IoT) technologies and Edge Intelligence.

In order to balance and mitigate challenges from changes in the electrical grid, it is paramount utilities begin to modernize infrastructure and start to participate in active grid management with data-driven, autonomous decisions. This means adding edge intelligence into substations and building a modern infrastructure at the edge.

Adding edge intelligence to the electrical grid means it will transform into a “grid of micro grids on a system of systems” that is more intelligent than ever before, says Khajuria. This built-in intelligence will allow us to have more insight into day-to-day operations and what’s to come—learning how to realistically power the needs of our world. Digitizing the grid and implementing edge intelligence into substations is a greener, more sustainable way to move forward.

The first obstacle in completing this task is coordination of all the major players, who must work in a synchronized manner. There are more than 3,000 electrical utilities that make up the full U.S. infrastructure. This massive, complex infrastructure is responsible for three respective major functions: power generation; power transmission; and power distribution.

To ease the process, Khajuria says to help industry move forward we should look to what IoT technologies are already working, such as virtualization of the substation infrastructure onto open architecture hardware. Considering the infrastructure of a substation, there can be 10 boxes doing 10 different things—each one in a silo. Currently, data distributed from these boxes are not combined or processed together at the control center.

Virtualizing these functions into software that runs on shared, open architecture hardware adds flexibility and overcomes the limitations of data silos. Now, all applications are sitting on one infrastructure in a virtualized environment. All data can be readily shared between applications and systems. This realizes four major benefits for utilities:

  1.  Reduced capital expenses, as the amount of needed hardware is lessened
  2. Manageability of applications reducing field maintenance costs
  3. System security
  4. Ease of operations across the board

According to Khajuria, utilities who implemented a virtualization solution at the edge found they could save up to 70%* in operational and maintenance costs.

*Salt River Project, Microprocessors vs. VPR tests, SRP Labs, Chandler AZ, Fall 2019. This is a preliminary estimate, and is dependent on existing standards/site-specific configurations. You should consult other sources to evaluate accuray. Your costs and results may vary.

A large number of critical and non-critical equipment used in the modern grid already is equipped to leverage the benefits of virtualization. Proven use cases include the following: remote accessibility that decreases operations and maintenance costs, fleet management of deployed assets for rapid installation of software patches and upgrades, and more.

In a virtualized grid, most assets are controlled by specific software that resides inside a virtualization element referred to as a “container” or a virtual machine. These assets feed data upstream into an edge computing platform running a larger number of containers. These then feed further upstream to decentralized edge servers, which access shared storage technology to provide control of the complete grid. This schema requires three key elements: the network, the physical layer (hardware to be virtualized), and software architecture.

How to Use an ADMS at the Edge

Once the virtualized infrastructure is built and all the data is accessible, the next step is to start using that data to extract intelligence, such as causes of failure, predictive maintenance, network anomalies, and more. Edge intelligence in a substation allows for an Advanced Distribution Management System, or ADMS. The substation can now collect data, complete load segregation, gain access to necessary additional information, and can start profiling each feeder and analyze its current state.

Once all the information and behavior is collected and stored, a plan can be built based on advanced analytics. This is when the grid becomes a “system of systems” where each substation becomes intelligent and, ultimately, manages itself and provides needed information to the control center.

For example, looking at a weather forecast and profile of the substation, one can see the amount of localized energy being consumed and the amount of energy being generated from renewables from any given time. For the sake of this example, let’s say from 9 a.m. to 2 p.m., 80% of all energy consumed by consumers is generated by renewables, locally. If this is the case, only a small amount of energy is needed from the utility. In this example, by 3 p.m. local energy production is going to drop down to 20%. If this is the case, then you’ll need 80% from the grid side at this time. The control center knows by 2 p.m. it must start spinning the turbines to start feeding the power. In other words, a pattern is now visible which allows for the practice of active grid management.

With the availability of data at the edge, we are able to use modern analytic technologies to gain a better understanding how systems work in order to take corrective actions before failures occur.