Demand for electricity is expected to soar as decarbonisation efforts in industry and transport creates greater demand. This has prompted concerns that current infrastructure is not ready for the increased loads, especially considering the demand from a rapid switch towards electric vehicles.
Researchers from Stanford University believe that many infrastructure upgrades may be unnecessary if proper software solutions are utilised instead.
A coordinated approach led by data-driven networks would not only improve the reliability of the electric grid, but also help to reduce its peak load, they said. If implemented widely, they believe it would save utility companies and their customers billions of dollars in grid infrastructure upgrades.
The study shows that, under status quo local control of demands and resources, four-in-five existing transformers will fail by 2050. Those transformers will need to be replaced, as could other infrastructure such as voltage regulators and possibly even electricity distribution lines.
"Under centralised control of the demands and resources across the grid, the number of transformer failures declines to just one in four," said Ram Rajagopal, civil engineering professor at Stanford and co-senior author of the study. "Such dramatic reductions in future distribution grid upgrades can accelerate the rate of electrification adoption."
Electricity demand is projected to balloon over the next several decades in part because electricity is widely considered more climate-friendly than fossil fuels and because more people around the world will gain access to electrified heating of homes and water, air conditioning and cooking.
Meanwhile, distributed energy resources like rooftop solar panels and residential and commercial batteries are expected to increase alongside smart appliances like electric vehicle chargers, electric space and water heaters, and air conditioners, whose power usage can be controlled to reduce consumer costs and total power system demand.
"The team decided to look at how much coordination of distributed energy resources and smart appliances can help grid reliability," said electrical engineering professor Abbas El Gamal, co-senior author of the study.
"We found that the benefits of coordination can be very significant using nothing but software. No new infrastructure. No replacements of distribution lines. It can all be done using existing computer clouds."
The team used models of distribution networks of varying sizes, mixes of homes and businesses, and from a range of climates across the US. They then applied recent projections of increases in electrification and distributed energy resources and their adoption scenarios up to the year 2050.
Thomas Navidi, lead student author of the paper, said: "One key finding that caught our attention was how valuable coordinating thermal loads is to our findings. It was very significant, even more so than battery storage and flexible scheduling of EV charging."
These loads include everything that uses electricity to raise or lower temperatures, from cooling or heating a home to making aluminium.
In addition to improved reliability, the team also found that coordination can reduce peak load on the distribution grid by some 17 per cent, which has the added benefit of reducing electricity costs during extreme climate conditions.
Next steps include developing a coordination scheme that would work across large numbers of homes and businesses, rolling out a pilot, and then – if the pilot works – developing a broader rollout that would include incentives to consumers to install the necessary software.
Last winter, the UK's National Grid introduced the Demand Flexibility Service, a system that warns consumers when the electricity grid is nearing peak capacity and gives them incentives to lower their energy requirements until demand drops.
