Artificial Intelligence Will Help Power America’s Clean Electricity Grids

Most countries will not meet zero-goals by 2050 based on current trajectories. It’s an unfortunate situation that needs to be taken seriously. While there are many factors delaying decarbonization efforts, the lack of clean energy-powered electric grids is perhaps the most significant barrier for nations.

The United States is not exempt—despite the recent influx of federal funds and legislation. Our electric grid infrastructure remains outdated, disparate, and managed by a complex network of regional operators. What’s more, just one-fifth of all U.S. electricity is generated from renewable energy sources.

However, that won’t be the case forever. Renewable energy sources continue to become more readily available. This presents American grid operators with a challenge but also opportunity.

Clean Energy Conundrum

Thanks to renewable energy, electric utilities are making headway with their climate goals, but they will be dealt unprecedented challenges from these less-dispatchable energy sources. Gone are the days of traditional unidirectional energy flow from centralized electric power plants. With the emergence of disparate clean energy sources, grid operators now must contend with bidirectional flows pouring into and out of the grid. For example, residential and commercial solar panels can produce surplus energy that flows back into the grid.

This brings us to the next challenge—prosumers. American citizens and companies have heeded the call for cleaner lifestyles. The U.S. Energy Information Administration estimated that residential solar panels generated 37 million megawatt-hours, accounting for 18% of all solar energy, in 2022. Globally, we can expect the number of grid-interactive commercial and industrial buildings to grow to 203 million by 2031, mainly utilizing demand response, solar power generation, and battery energy storage technologies.

Although it’s heartening to see Main Street, Wall Street, and Pennsylvania Avenue align on the journey to energy transition, the growing support of clean energy has led to an influx of clean energy production being fed into the grid. Residential, commercial, and industrial renewable energy sources, like solar panels, wind farms, and electric vehicles (EVs), will likely add more congestion to the already challenged existing grid. This could result in asset deterioration, higher transportation losses, blackouts, and voltage fluctuations, among other concerns. On the flip side, clean energy sources are intermittent in nature and therefore less predictable. Solar panels in the northeast may overwhelm the grid during the sunnier summer months, but during a stormy winter season, there may not be enough energy produced from solar to adequately power the electric grid. This is why the integration of battery storage is an important part of the energy portfolio.

Lastly, the increased use of renewable energy has exacerbated a long-standing, industry-wide issue—supply and demand tracking. Electric supply and demand data has always been essential, but tracking this data is particularly important as the nation turns to clean energy sources. With fluctuating bidirectional energy flows from self-reliant prosumers and decentralized sources, grid operators must have real-time analytics to ensure they have the right amount of power going to the right customers at the right time. In other words, supply and demand data is needed to quite literally keep the lights on.

Power of AI

To be more specific, data and analytics are needed to enable a key technological pillar within the American grid modernization agenda—artificial intelligence (AI). And it’s not just electric supply and demand data that’s needed. Grid operators must monitor and analyze data from connected technologies, like Internet of Things (IoT) sensors, and storage technologies, like the aforementioned batteries. They will also need to track everything from weather patterns and geospatial metrics to EV charging stations and historical generation, consumption, and distribution data—according to Capgemini’s World Energy Markets Observatory report.

This data will then collectively enable AI solutions like forecasting models that can predict energy input and output as well as consumption. AI is, and will continue to be, present in more physical applications as well, like AI chips in smart meters. By predicting when the grid may be overstressed or underutilized, grid operators can prevent blackout scenarios, meet consumer expectations, and ensure the grid remains in balance.

How to Make AI’s Predictive Potential a Reality

So, how can electric utilities harness the power of AI and deploy these use cases? Some may argue grid operators already are exploring AI applications, and they’d be right. AI is not a new technological development, but since the latest technology will be applied to rapidly advancing infrastructure powered by decentralized, intermittent energy sources, utilities must rethink their AI use cases, and development and deployment tactics.

Here are a few steps that should be on every company’s AI checklist:

  • Enhance data quality.
  • Break data silos.
  • Expand data-sharing ecosystem across the private and public sectors.
  • Adopt modern data and AI platforms in hybrid cloud environments.
  • Prioritize use cases using a data-driven process tracking against factors such as regulatory requirements and customer expectations.
  • Continually monitor AI model performance and track against business goals.

Getting Back on Track

The U.S.’s decarbonization ambitions depend on the success of our clean electric grids. We have almost 30 years to reach net-zero emissions and billions of dollars in funding to help get us there, but that won’t matter if the time and money are not strategically utilized to enable renewable energy sources and modernize the grid.

While AI, and the data and analytics that enable it, is not the end-all-be-all of the national grid modernization program, it is a significant technological pillar that should not be overlooked. Utility leaders would be wise to take another look at their AI capabilities and plans, and consider how they can enhance their roster of AI solutions and improve their deployment and enablement tactics.

Claire Gotham is vice president of Utilities & Renewables at Capgemini Americas.