MIT News
Q: Why does the facility grid must be optimized in the primary place?
A: We want to take care of an actual balance between the quantity of power that’s put into the grid and the quantity that comes out at every moment in time. But on the demand side, we now have some uncertainty. Power firms don’t ask customers to pre-register the quantity of energy they’re going to use ahead of time, so some estimation and prediction have to be done.
Then, on the provision side, there is usually some variation in costs and fuel availability that grid managers must be aware of. That has develop into a fair larger issue due to the combination of energy from time-varying renewable sources, like solar and wind, where uncertainty within the weather can have a significant impact on how much power is offered. Then, at the identical time, depending on how power is flowing within the grid, there’s some power lost through resistive heat on the facility lines. So, as a grid operator, how do you make certain all that’s working on a regular basis? That’s where optimization is available in.
Q: How can AI be most useful in power grid optimization?
A: A method AI might be helpful is to make use of a mix of historical and real-time data to make more precise predictions about how much renewable energy will likely be available at a certain time. This could lead on to a cleaner power grid by allowing us to handle and higher utilize these resources.
AI could also help tackle the 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 find out which power generators should produce power, how much they need to produce, and after they should produce it, in addition to when batteries needs to be charged and discharged, and whether we are able to leverage flexibility in power loads. These optimization problems are so computationally expensive that operators use approximations so that they can solve them in a feasible period of time. But these approximations are sometimes improper, and once we integrate more renewable energy into the grid, they’re thrown off even farther. AI can assist by providing more accurate approximations in a faster manner, which might be deployed in real-time to assist grid operators responsively and proactively manage the grid.
AI is also useful within the planning of next-generation power grids. Planning for power grids requires one to make use of huge simulation models, so AI can play an enormous role in running those models more efficiently. The technology also can help with predictive maintenance by detecting where anomalous behavior on the grid is prone to occur, reducing inefficiencies that come from outages. More broadly, AI is also applied to speed up experimentation aimed toward creating higher batteries, which might allow the combination of more energy from renewable sources into the grid.
Q: How should we expect in regards to the pros and cons of AI, from an energy sector perspective?
A: One necessary thing to recollect is that AI refers to a heterogeneous set of technologies. There are differing types and sizes of models which might be used, and other ways that models are used. For those who are using a model that’s trained on a smaller amount of information with a smaller variety of parameters, that’s going to eat much less energy than a big, general-purpose model.
Within the context of the energy sector, there are quite a lot of places where, should you use these application-specific AI models for the applications they’re intended for, the cost-benefit tradeoff works out in your favor. In these cases, the applications are enabling advantages from a sustainability perspective — like incorporating more renewables into the grid and supporting decarbonization strategies.
Overall, it’s necessary to take into consideration whether the sorts of investments we’re making into AI are literally matched with the advantages we wish from AI. On a societal level, I feel the reply to that query straight away is “no.” There’s quite a lot of development and expansion of a selected subset of AI technologies, and these will not be the technologies that can have the largest advantages across energy and climate applications. I’m not saying these technologies are useless, but they’re incredibly resource-intensive, while also not being answerable for the lion’s share of the advantages that could possibly be felt within the energy sector.
I’m excited to develop AI algorithms that respect the physical constraints of the facility grid in order that we are able to credibly deploy them. It is a hard problem to resolve. If an LLM says something that’s barely incorrect, as humans, we are able to often correct for that in our heads. But should you make the identical magnitude of a mistake when you’re optimizing an influence grid, that could cause a large-scale blackout. We want to construct models in another way, but this also provides a possibility to profit from our knowledge of how the physics of the facility grid works.
And more broadly, I feel it’s critical that those of us within the technical community put our efforts toward fostering a more democratized system of AI development and deployment, and that it’s done in a way that’s aligned with the needs of on-the-ground applications.
