Artificial intelligence has been making headlines recently due to its effectiveness Rapidly increasing energy demandespecially the escalation Electricity use in data centers Which enables the training and deployment of state-of-the-art generative AI models. But it’s not all bad news, as some AI tools have the potential to reduce some forms of energy consumption and enable cleaner grids.
One promising application is using AI to optimize the power grid, which would improve efficiency, increase resilience to extreme weather, and enable the integration of more renewable energy. To learn more, MIT News Talk to Priya Dontethe Silverman Family Career Development Professor in the Department of Electrical Engineering and Computer Science (EECS) at the Massachusetts Institute of Technology (MIT) and a principal investigator at the Laboratory for Information and Decision Systems (LIDS), whose work focuses on applying machine learning to improve the power grid.
Q: Why does the power grid need improvement in the first place?
A: We need to maintain a careful balance between the amount of energy that is put into the grid and the amount that comes out at each moment in time. But on the demand side, we have some uncertainty. Energy companies do not require customers to pre-register how much energy they will use in advance, so some estimation and forecasting must be made.
Then, on the supply side, there is usually some variation in costs and fuel availability that grid managers must respond to. This is becoming a bigger problem due to the incorporation of energy from time-variable renewable sources, such as solar and wind, where weather uncertainty can have a significant impact on the amount of energy available. Then, at the same time, depending on how the power is flowing in the grid, there is some power lost through resistive heat on the power lines. So, as a network operator, how can you ensure that everything is working all the time? This is where optimization comes in.
Q: How can AI be most useful in improving the power grid?
A: One way AI can be useful is by using a combination of historical data and real-time data to make more accurate predictions about how much renewable energy will be available at a given time. This could lead to a cleaner energy grid by allowing us to handle and use these resources better.
AI can also help address complex optimization problems that power grid operators must solve to balance supply and demand in a way that also reduces costs. These optimization problems are used to determine which power generators should produce power, how much they should produce, and when they should produce it, as well as when the batteries should be charged and discharged, and whether we can benefit from flexibility in power loads. These optimization problems are computationally very expensive, which forces operators to use approximations to be able to solve them in the possible amount of time. But these rough estimates are often wrong, and as we integrate more renewable energy into the grid, it is thrown away further. AI can help by providing more accurate approximations in a faster manner, which can be deployed in real-time to help grid operators manage the grid responsively and proactively.
AI can also be useful in planning next-generation power grids. Planning for power grids requires the use of massive simulation models, and AI can play a big role in making those models work more efficiently. Technology can also help with predictive maintenance by detecting where anomalous behavior on the grid is likely to occur, reducing the inefficiencies that come from power outages. On a larger scale, AI could also be applied to speed up experiments aimed at creating better batteries, which would allow more energy from renewable sources to be integrated into the grid.
Q: How should we think about the pros and cons of AI, from an energy sector perspective?
A: One important thing to remember is that AI refers to a heterogeneous set of technologies. There are different types and sizes of forms used, and different ways to use the forms. If you use a model that is trained on a smaller amount of data with fewer parameters, it will consume much less power than a large, general-purpose model.
In the context of the energy sector, there are a lot of places where, if you use application-specific AI models for their intended applications, the cost-benefit trade-off works to your advantage. In these cases, applications enable benefits from a sustainability perspective – such as integrating more renewable energy sources into the grid and supporting decarbonisation strategies.
Overall, it’s important to consider whether the types of investments we make in AI actually align with the benefits we want from AI. On a societal level, I think the answer to this question now is “no.” There is a lot of development and expansion of a particular subset of AI technologies, and these are not the technologies that will have the greatest benefits across energy and climate applications. I’m not saying that these technologies are useless, but they are incredibly resource-intensive, while they are not responsible for the lion’s share of the benefits that can be felt in the energy sector.
I’m passionate about developing AI algorithms that respect the physical constraints of the power grid so we can deploy them credibly. This is a difficult problem to solve. If an LLM says something that’s a little bit incorrect, as humans, we can usually correct that in our heads. But if you make the same mistake when improving the power grid, it could lead to widespread power outages. We need to build models differently, but this also provides an opportunity to leverage our knowledge of how the physics of the power grid works.
More broadly, I believe it is crucial that those of us in the technical community push our efforts toward promoting a more democratic system for developing and deploying AI, and do so in a way that aligns with the needs of real-world applications.
(Tags for translation)Priya Donte
