Although a lot about human beings is worth lauding, there is nothing that deserves the honor more than our ability to improve at a consistent clip. This ability, in particular, has really enabled the world to clock some huge milestones, with technology emerging as quite a major member of the group. The reason why we hold technology in such a high regard is, by and large, predicated upon its skill-set, which guided us towards a reality that nobody could have ever imagined otherwise. Nevertheless, if we look beyond the surface for one hot second, it will become abundantly clear how the whole runner was also very much inspired from the way we applied those skills across a real world environment. The latter component, in fact, did a lot to give the creation a spectrum-wide presence, and as a result, initiated a full-blown tech revolution. Of course, the next thing this revolution did was to scale up the human experience through some outright unique avenues, but even after achieving a feat so notable, technology will somehow continue to bring forth the right goods. The same has turned more and more evident in recent times, and assuming one new discovery ends up with the desired impact, it will only put that trend on a higher pedestal moving forward.

The researching team at University of California Santa Cruz has successfully developed an AI-based approach to restore power efficiently in the event of an outage. To understand why this is such a significant development, we must start by acknowledging how, in many communities, infrastructure and its users are totally reliant on a local power-generating utility company for electricity, meaning if and when something goes wrong with that company’s operation, the power goes out. Now, we have reached a point where the world is turning towards smart electricity systems that are decked up with computers and sensors, systems that incorporate multiple local renewable energy sources, such as rooftop solar panels or small wind turbines, to create a backup power reserve. However, in order to make the most of this technology, you need a microgrid, which can be connected to the main power utility source but can also function while disconnected in an “islanding mode”. The new AI approach pledges to optimize the way microgrids pull from these various alternate sources, such as renewables, generators, and batteries. Leveraging deep enforcement learning, the approach is rooted in rewarding the algorithm when it responds successfully to a changing environment. According to certain reports, the method takes into account real-time conditions, while simultaneously using machine learning to find long-term patterns that affect the output of renewables like the varying demand on the grid at a given time, and various weather factors that affect renewable sources. Contextualizing the new approach’s uniqueness is a fact that the technologies we have seen thus far have used a technique called model predictive control (MPC), which has no option but to base all decisions simply on the available conditions at the time of optimization.

“Nowadays, microgrids are really the thing that both people in industry and in academia are focusing on for the future power distribution systems,” said Yu Zhang, an assistant professor of Electrical and Computer Engineering at the University of California Santa Cruz. “Essentially, we want to bring the power generation closer to the demand side in order to get rid of the long transmission lines. This can improve the power quality and reduce the power losses over the lines. In this way, we will make the grid smaller but stronger and more resilient.”.

Explaining the whole mechanism is an example, where if the new system’s prediction says that the sun will shine brightly in an hour, it would automatically use up its supply of solar energy bearing in mind that it will be replenished soon. Of course, the system will have a different strategy for different weathers to ensure the most effective outcome in every case.

The researchers have already put their technology through some initial tests during which they discovered that it significantly outperforms traditional MPC methods. This was especially evident when the forecasts of renewable sources were lower than reality because of their better understanding of all the possible solar profiles throughout any given day. Furthermore, they learnt how the reinforcement learning controller is able to respond much faster than traditional optimization methods in the moment of a power outage. In case these tests don’t present for you a strong enough testimony, then we must acknowledge that the method was also placed first in a recently-concluded global competition which invited participants to use reinforcement learning or similar techniques to operate a power grid.

For the immediate future, the team plans on running their technology on a microgrid in a laboratory setting. In the long term, the hope is to implement it on UC Santa Cruz campus’s energy system to address outage issues currently faced by campus residents.

“We’re modeling a whole bunch of things—solar, wind, small generators, batteries, and we’re also modeling when people’s electricity demand changes. The novelty is that this specific flavor of reinforcement learning, which we call constrained policy optimization (CPO), is being used for the first time,” said Shourya Bose, a Ph.D. student in Zhang’s lab.