Modest upgrades to the next generation of smart meters will allow them to run sophisticated Artificial Intelligence (AI) algorithms. By deploying AI to the grid edge, homes can become more efficient and flexible, and we can cost effectively accelerate the energy transition.
The Current Situation: epically expensive, labour intensive, not very effective
Today’s grid cannot cope with a huge increase in demand from electric cars and heat pumps. Upgrading all that copper will tank already feeble business cases and balancing a grid based on renewables will be extremely expensive. Storage won’t come to the rescue in any helpful timeframe. Wholesale power costs will continue to rise and price spikes will punch through retail margins. The appalling situation in the Ukraine is a harsh reminder of how frail our energy security is.
Demand Side Flexibility (DSF) has an important role but today it operates for industry alone. That finite load won’t be enough to balance fluctuations elsewhere. To corral enough demand, DSF must extend to the home but existing options are expensive. In the EU, we don’t have €800bn to install 200m home batteries. Even if we did, we don’t have the skilled workforce to deliver such a magnificent project. EV chargers are important but they too often sit behind walled gardens and can’t contribute to peak hour flexibility if they only run at night.
Developing distributed and connected energy resources in the home are vital parts of the solution, but as presently construed, current efforts will fail to halt 1.5 degrees of warming, and probably 2.0 as well. We will fail because the solutions are expensive, difficult to install, and they neglect the consumer – they forgo the demand in a demand supply equation!
The Way Ahead: software + smart meters
To overcome these limitations, a practical solution must have several features:
- It needs to be highly scalable and dramatically more cost effective than previous offerings.
- It needs to balance the grid, shave and shift the peaks.
- It needs to engage the consumer.
Recent advances in artificial intelligence (AI) are creating the practical solutions we need. We can avoid the expense and intricacies of the current plan with this technology.
More sophisticated software on an enhanced smart meter will provide intelligence directly into the home. Intelligence that engages consumers, optimises for cost and carbon, and reduces the capital and operating costs of the grid. In sum, an intelligence that accelerates the energy transition.
When deployed across a smart meter rollout, AI starts to resolve the intractable issues facing an energy system in transition. AI can help detect device usage in real time. This real time device detection (aka load disaggregation) helps consumers save energy and enables demand side flexibility. By helping householders identify the most impactful loads to shift, real time load disaggregation makes it simple for domestic users to take part in demand peak events. Enabling the aggregation of vast reserves of residential demand. This can then be tapped by ancillary services and wholesale markets. Short- and long-term demand forecasting becomes more accurate with appliance level mapping, saving infrastructure investment and power costs. Anomalies on the grid, such as faulty transformers, vegetation brush, and power outages can also be identified and located.
With intelligence in the smart meter, shifting load is highly cost effective. To make meters intelligent, an investment of circa €20 per device will suffice. With this investment, trials have shown the average home (even accounting for non-participants) can shift 0.3KW during peak hours. This sounds like a small amount but it is deceptive. Across a rollout, it adds up. Over 50m homes, that equates to 16GW for an investment of €1bn. No other solution is able to achieve that return on investment. A similar investment in batteries would shave just 1GW. These numbers are transformative. They change the debate and the cost economics for meter buyers and regulators. They put smart metering at the heart of resolving climate change.
Accelerating the EU’s clean energy transition
Smart meters can be the cornerstone of the flexible grid and become the intelligence that coordinates the home. Whilst first-generation meters have represented a huge stride forward, they have not fulfilled their potential. Early meters have been forced to sacrifice too much capability for cost and have been held back by rules on data sharing. Saving a few euros on a smart meter may put climate goals out of reach. However, by modestly enhancing meter technology, it’s possible to deliver vastly improved flexibility behind the meter and deliver the energy transition at a fraction of the cost.
Sense uses machine learning algorithms to analyse high resolution electricity data, providing a breakdown of domestic electricity consumption to appliance level in real time. Sense’s AI technology helps consumers save energy, enables demand side flexibility and identifies anomalies on the grid.
This blog is an extract from Sense’s latest thought leadership whitepaper – ”A small step for smart meters, a giant leap for climate change” by Michael Jary. To read the full paper, please visit Sense.com
Sense joined ESMIG in January 2022. See their member profile for more information.
This blog is a space for debate where ESMIG members share their thought leadership. All opinions expressed are the authors. The content of this article is not an official position paper endorsed by the association.