Updated: Dec 9, 2021
Artificial intelligence (AI) gained prominence as a mainstay of science fiction and regular futuristic predictions about all aspects of various industries, products and more. What for many may evoke images of robots or sentient super computers actually has a rather mundane definition. According to IBM, a leader in the AI space, artificial intelligence is defined as:
The science and engineering of making intelligent machines, especially intelligent computer programs. It is related to the similar task of using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable.
Even this carefully spelled out definition can be hard for many to truly understand, but the truth is that AI already pervades our daily lives, from social media accounts identifying what to show you to real-time optimized routes from Google Maps.
Regarding the energy sector, starting around the turn of the century AI was consistently paired as a concept with the futuristic ‘smart grid’ to describe how utility assets and customer user of electricity would evolve. And without many customers even noticing it, many of these AI applications have already served to revolutionize the power grid across the country today.
The most successful applications of AI and the optimal function of energy providers will both go relatively unnoticed if they’re doing everything correctly, so the following highlights some of the ways the worlds of AI and the grid have crossed paths.
Using AI to Process Mass Data with Unprecedented Speed
While electrons have long been the key asset, or even ‘currency,’ of the power sector, the process of utility digitalization and the integration of the Internet of Things have swiftly elevated data to be arguably just as important as the electricity itself. Artificial intelligence is a key tool in being able to collect, process, and make decisions based on the treasure troves of data newly becoming available to grid operators.
One example of AI processing and utilizing quantities of data that would be impossible otherwise is in detecting problems on the grid in real-time, whether from faulty equipment, cyber threats, or even theft of power from the grid. By regularly reading the data from smart meters across the grid and sensors throughout the transmission and distribution sector, artificial intelligence can learn about normal operations, identify the types of interruptions that will self-correct, and infer when an unexpected pattern arrives in the data that requires intervention. AI raises alarm about potentially issues that previously required manual review and decision-making, so AI availability ensures time isn’t wasted while problems persist anywhere across the infrastructure.
Another functionality that AI enables by processing data constantly comes in the form of keeping the markets always balanced. On a minute-by-minute basis, the rate of power demand and supply are changing across the grid, and keeping these rates as closely aligned as possible is essential to minimizing waste, optimizing operational efficiency, and preventing unexpected blackouts. For example, the use of AI tools to track changes in patterns on the grid was particularly important in the heart of the COVID-19 pandemic as the regular patterns utility operators knew to expect were turned on their head.
Lastly, from a business perspective the power providers can utilize intelligent processing of customer data to best identify new revenue opportunities. Utilizing comprehensive data from customer billing and metering data and seeing what similar customers take advantage of across the utility enterprise, the use of artificial intelligence can make smart suggestions about which customers should be made aware of potential opportunities of which they’d want to take advantage. Rather than providing offerings to customers who won’t want them, AI allows utilities to pick the best offerings for the right customers and increase the chance of adoption of those new programs.
Supplementing Employee Experience with AI Processing Speeds
When AI/digital tools are introduced in new areas of a business, whether in the utility space or otherwise, sometimes existing employees feel a sense of worry. Bringing in AI, they may think will mean that their individual insights, expertise, and experiences will all be shortsightedly replaced with automated tools. But the best way for utilities to introduce AI tools is to do so in areas where the intelligence can be used to supplement and complement the institutional knowledge and intuition that only human workers can provide. In the utility sphere, that coordination between AI and long-time workers takes place in many ways.
For one, the traditional utility method of completing inspections on assets out in the field would be to send crews out to manually visit and check the state of every pole, wire, transformer, and other piece of equipment on the grid. This process has long been costly, time-consuming, and prone to oversights. But with the advent of artificial intelligence, utilities can replace physical visits with reading data from sensors, having machine learning programs look at thousands of images taken from automated drones or even satellite of utility equipment, and other intelligent solutions. These AI methods can flag the hotspots where equipment may need a human visit to really inspect, thus saving the utility lots of time and headaches from sending their hardworking employees across every single piece of equipment.
In that same vein, vegetation management is a critical duty of utilities, especially as dry and hot conditions increase the risk of electricity-sparked wildfires. By using that same drone or satellite imagery, fed into intelligent algorithms, utilities can then identify the areas to have their vegetation management experts review the state of trees around their assets. Rather than having those experts comb through thousands of images or data points, boring them to the point that they might inadvertently overlook a problem area, AI can ensure they only need to review a few dozen areas at a time and decide about where to send key crews.
Lastly, in the wake of major storms and other weather events, knowing where to send power restoration crews first can have massive impact on how swiftly electricity is returned to different areas of the grid. But for decades, utilities were flying blind and simply had to rely upon reports from customers about outages or downed power lines. Today, though, artificial intelligence can be used to identify instantaneously where equipment has been impacted, identify which customers are affected, and even communicate directly with those customers while dispatching utility crews to address the highest priority outages first.
Future AI Opportunities on the Grid
Despite all these tangible ways in which the AI world is already keeping the grid humming today, the fact does remain that we can look ahead to some future applications that are currently in the R&D stage but will soon add their imprint on the grid.
For example, at Stony Brook University a program entitled ‘AI-Grid’ is researching how an AI-enabled grid can be used to keep power systems robust in the face of malicious cyberattacks, faulty equipment, and natural disaster impacts. Not only that, but this AI-backed grid can also identify pain points where disadvantaged communities are being hit with how power costs despite lower reliability, creating a more equitable grid.
Another key trend across the grid is the physical transformation coming from the implementation of distributed energy resources, increased penetration of electric vehicles and on-site energy storage, microgrids, and more. The increased number and scale of nodes consuming, producing, and storing energy across the grid transforms the power industry from the central generation model that had been used for decades, so artificial intelligence is necessary to ensure the real-time management of these different load profiles and will eventually use machine learning to make this process as seamless as possible.
Even at the source of electricity generation, in the power plants themselves, AI is being studied to make that process more efficient to reduce losses and emissions while optimizing performance. By using AI-guided algorithms to time the start-up and shut-down processes of various equipment, anticipating future maintenance needs, and communicating unexpected issues instantly, power generation equipment will be optimized better than ever in the coming years.