Modernising the grid with next-gen edge computing and AI

Virtualisation, edge computing and edge AI innovations are enabling utilities to realise significant benefits, writes Ramkumar Venketeramani, Director, Edge Innovation and Solutions, VMware Software Defined Edge Division, Broadcom.

600GW. That’s how much new solar capacity European electric utilities need to add by 2030 to meet REPowerEU targets for energy independence – and solar is just one part of the picture. Utilities plan to add hundreds of GW of onshore and offshore wind capacity too, and prepare for millions of new electric vehicles (EVs), heat pumps, and other emerging green consumer technologies.

Meanwhile, Europe must also contend with explosive ongoing growth in energy demand, especially for emerging technologies like generative AI (genAI), which consume an ever larger share of the world’s electricity.   

To meet these challenges, the EU is investing trillions of Euros to bring new renewable resources online. But they also must become far more efficient in how they transmit electricity and manage and operate the grid.

This transition is already well under way in distribution stations, where utilities  are reimagining substation technology from the ground up. As they do, many are finding that software innovations from the IT world – like virtualisation, edge computing, and artificial intelligence (AI) – can play a key role in meeting the challenge. And they’re turning to digital transformation leaders like VMware to help them do it.

The looming energy challenge

If it were only a matter of adding more substations, updating the grid would be difficult enough. But navigating new renewable resources coming online, along with explosive growth in EVs and residential solar, demand a radically different approach to power transmission.

Existing grids were designed as one-way infrastructures transmitting energy from power generation units out to consumers. In the emerging world of distributed energy resources, however – where electricity might be generated in hundreds of edge locations, EV charging stations line every highway and millions of residences become net energy producers – that approach won’t work. The grid must now function more like a multi-directional mesh. Additionally, resources like wind and solar are cyclical, creating peaks and troughs in capacity that must be managed intelligently.

These trends create a perfect storm for utilities. In addition to rapidly bringing up new substations, they need to overhaul existing ones, adding tools to monitor and manage operational technology (OT) at the edge. But significant barriers stand in the way:

• Ageing, hardware-centric technology: Most substations still run on legacy electromechanical equipment, some of which is decades old. Even when using newer microprocessors, OT functions like protection, automation and control (PAC) still require dedicated proprietary hardware. As a result, utilities have little flexibility to evolve features and capabilities in existing substations.
• Slow, inefficient operations: Substation operations and maintenance also pose challenges. The bays of ageing equipment packed into distribution stations require huge amounts of complex (often dangerous) wiring, labour and overhead. This requires hands-on expertise at every distributed site, making scaling capacity a slow, expensive proposition.
• Lack of data-driven intelligence: The challenge of efficiently managing multi-directional power flows is not insurmountable. Modern AI and machine learning (ML) algorithms are more than capable of optimising, if not fully automating this process. To apply contemporary data science though, utilities need visibility into substation data. Today, much of it’s locked away in legacy equipment silos, making it difficult to collect statistics about substation health and performance, much less analyse and act on that data locally.

Utilities will be hard-pressed to meet tomorrow’s energy requirements with yesterday’s technology. They must modernize the grid.

Modernising substations

Modern software approaches can unlock the potential of distributed sensor data to scale capacity, optimise operations and apply AI inferencing across the grid. But bringing contemporary IT approaches to legacy OT environments is easier said than done. Most were designed for data centre and enterprise environments – not tens of thousands of distributed substations, many with unreliable connectivity and no onsite personnel.

Today, utilities are making major strides in overcoming these challenges to reimagine substation OT. VMware is working with groups like the Edge for Smart Secondary Substation (E4S) Alliance to make it easier to implement virtualisation, edge computing and edge AI innovations across the grid.

By virtualising PAC and other distributed applications, and consolidating workloads on commercial off-the-shelf servers, utilities can realise significant benefits – including faster deployment timelines and dramatically lower costs.

For new ‘greenfield’ substation deployments, some utilities we work with have cut construction costs in half and reduced ongoing maintenance costs by 85%. By adopting modern software-driven approaches, utilities can also:

• Accelerate deployments by using a smaller, standardised OT footprint across thousands of remote substations
• Replace multiple bays of legacy equipment with a single server, reducing the number of devices to deploy and maintain in the field
• Use next-generation utility applications – like smart-balancing energy flows, smart metering and predictive maintenance – to optimise electricity flows and maximise efficiency
• Unleash innovation by adding best-in-class software capabilities from a rapidly growing ecosystem of vendors, while using the same consistent hardware and operations
• Gain flexibility to update features and capabilities of substation equipment via software updates, instead of having to swap out hardware
• Continually optimise grid operations by collecting and feeding data into analytical engines, AI inferencing models, and advanced distribution management systems.

Unleash edge computing intelligence

Virtualising distributed equipment is just the first step. Meeting tomorrow’s energy requirements requires distributed intelligence too. That means bringing real-time analysis and decision-making to thousands of remote sites, many with unreliable connectivity and little or no onsite support.

Enter VMware Edge Compute Stack.

VMware Edge Compute Stack provides an edge-optimised runtime and orchestration platform for frictionless management of edge applications and infrastructure at scale.

Using the hypervisor trusted by enterprises worldwide, utilities can support the same AI/ML workloads and software capabilities as any other data centre. But with Edge Compute Stack automatically ensuring that each site has the compute resources it needs, they can extend those capabilities to remote substations.

VMware Edge Compute Stack provides:

• Control plane flexibility: Utilities can push control plane functions out to the edge to enable real-time local AI/ML decision-making, while managing thousands of edge devices from a centralised location.
• Zero-touch operations: Rather than pushing configurations out to the edge with a centralised controller (a problem for remote sites with poor connectivity), Edge Compute Stack uses pull-based orchestration. Using infrastructure-as-code techniques, utilities can define an end-state template for substations. Each site then downloads that definition and configures itself to execute that intent.
• OT optimisation: Utilities can use modern GitOps approaches to manage thousands of edge nodes at scale, but in a way that’s optimised for OT environments with harsh operating conditions and unreliable network connectivity.

Now, utilities can collect more data from more sources, and apply AI/ML intelligence to act on it at the edge – accelerating substation deployments and optimising electricity flows for a multi-directional grid. Meanwhile, by analysing substation data and flagging potential issues, utilities can perform proactive maintenance, avoid catastrophic failures, and extend the life of substation equipment.

Learn More

Ready to see how virtualisation, edge computing and edge AI can help your organisation meet the energy challenges of tomorrow? If you would like to meet our team at Enlit Europe 2024, send us an email. For more information, we’ll be presenting at the Summit and Digitalisation Hub sessions, or you can review the following resources:

• Webinar: Modernizing electrical substations for a resilient grid with edge computing 
• Video: Accelerating edge computing for electric utilities with VMware Edge Compute Stack
• Blog: VMware Edge Cloud Orchestrator – The key to centralized edge management for electric utilities
• Blog: UK Power Networks transforms the power grid with VMware software-defined systems

About the author