Steve Rogerson talks with Jason Ghidella from Mathworks about how digital twins can help stabilise electricity grids.

Nobody likes power cuts. However, the utilities that supply our electricity are finding it harder to maintain a stable supply as more renewable sources come online, from giant wind farms to rooftop solar panels that let householders sell back to the supplier. On top of that, there is a growing electric vehicle charging network, with individual cars acting as power banks that can also supply the grid with electricity.

As such, utilities are finding it more difficult not just to maintain the power flow but even to keep track of what is going on. Planning future additions to the grid can be mind-blowing. However, the answer could lie, as some utilities have already found out, in digital twins. These virtual models can help predict how the grid will respond to changing requirements both on the supply and demand side, and how it will react to freak incidents such as severe weather.

To build such a digital twin might itself seem daunting, but this can be made easier using a technique known as model-based design that lets engineering teams adopt a system-level approach using executable models – these are the digital twins of the individual parts – that unify all aspects of the grid into one development environment.

This can reduce the risks of integrating assets such as smart grids, substations and control centres. At the same time, model-based design can strengthen collaboration across different disciplines to help utilities manage aging infrastructure and deploy technologies faster and more reliably.

Organisations using this approach are seeing significant returns: up to 60% cost reductions, improved system resilience and faster deployment cycles, all while maintaining compliance and grid stability.

One company that can help with model-based design is Mathworks, and I caught up with the firm’s principal product manager Jason Ghidella this week to find out what problems utilities are facing how digital twins can help.

“There is an exploding complexity,” Jason told me. “And renewables create a whole new range of details you need to deal with. They need to build complex software that works on their devices. Requirements are spread over different teams. And then when you get them all back together, you find they don’t work properly.”

Model-based design, he said, was the answer Everything can be done in the models and this lets the engineers see quickly what can be done and what can’t.

A good example of how the Mathworks approach can help can be found at Eversource Energy (www.eversource.com), New England’s largest energy utility.

It provided what is known as probabilistic load flow (PLF) capabilities to enhance system planning for Eversource. The rapid rise of electric vehicles, heat pumps and solar panels added uncertainty to its electric distribution planning, so Eversource built a PLF system that helped it process and prioritise millions of grid scenarios by likelihood and risk.

Eversource found traditional model scenarios inadequate for future planning. Instead, the company integrated PLF automation into its power system analysis to simulate scenarios to improve its distribution modelling capabilities and support investments in enhancing data analytics.

To do this needs accurate representations of each element within the system. These virtual models can include the technical specification of a solar panel or EV charger, the design of the turbines on a wind farm, all done to the fidelity needed to understand how they will operate when put together. Each model, or digital twin, needs that level of accuracy. Once these digital twins are in place, model-based design lets engineers build larger digital twins of the complete system, letting them move assets, take them on or off line, replace them with different versions, and so on to see the effects on the whole system.

“We think of digital twins as a virtual representation,” said Jason. “They can be designed to work at each stage of a product lifecycle through the design process, making sure elements don’t overload or whatever. And they can be used once the system is operating for debugging and optimising. We use simulation and verification to get rid of all the bugs.”

For utilities, this can help them ensure anything plugged into the grid will work properly, from renewables to microgrids.

“You can simulate the effect on the whole system,” said Jason. “Without a digital twin, you have to test everything in the field and that can be incredibly expensive and dangerous.”

I asked Jason how this could help a utility, say, add a wind farm to its grid. From the start, he said, this could involve making virtual models of every element of the wind farm or use existing models of the elements within in it such as the turbines, energy levels, technical specifications and so on, plus the effects of different wind speeds.

“I can even use it to size the wind farm,” said Jason, “like to see how big the turbines need to be.”

You then have to look at how integrating this wind farm into the grid will affect stability.

“This is the important aspect,” he said. “You don’t want to make the grid unstable, so you go through all the scenarios, such as what happens if there is a big storm, what are the effects if a turbine goes offline for maintenance and so on. You have a digital representation of the grid and the wind farm and you can see what all the effects are.”

Simulation, he said, was important, but mostly this is just simulation of the grid and not everything that interacts with it.

“We provide a flexible environment to model the complete system and one that will function at the right level of fidelity,” he said. “We use AI and machine-learning components to model the system. This can then plug into your product development to test before you build prototypes.”

And if all that sounds a bit too much, Mathworks (www.mathworks.com) also has consultants to hold your hand and guide you through the process.

“We can take you from the initial steps to deploying a full environment,” said Jason.

As I said at the start, power cuts are annoying, but with more utilities adopting model-based design and using digital twins to simulate their grid and predict the impact of external and internal events, hopefully the number and severity of such black outs will reduce significantly. You can leave the lights on.