AI and Clean Energy Workshop: Problems Technology Can Solve

Introduction

In August 2024, Work on Climate hosted an expert-led virtual workshop to help aspiring climate founders interested in AI discuss challenges around clean energy. AI requires a lot of energy to function, but it can also help us come up with solutions to problems with our energy system. 

Through the workshop, participants gained understanding of the most pressing problems related to a clean energy transition and how AI could help. The event encouraged learning and growth for all, and it allowed participants to meet with people who share their interests and concerns. The expert panelists provided invaluable insights for our aspiring founders to carry with them. 

The virtual workshop had themed rooms for more intimate discussions. Those rooms focused independently on separate topics, including: 

• Room 1: Grid (Broadly)
• Room 2: Grid Interconnection and distributed energy resources (DERs)
• Room 3: Enablement – Permitting, policy, rates, regulation 
• Room 4: Energy Storage and Battery Technology 
• Room 5: Solar and Nuclear Generation 
• Room 6: Forecasting – Supply, demand, and prices 
• Room 7: Wind

All of these topics combined helped to create a complete picture of the challenges we’ll face, as well as the opportunities we have with the clean energy transition, and how AI can help. 

In this blog post, we’ll cover the points that emerged from discussions with experts and participants in Rooms 1, 2, 3, and 6, which focused on challenges. The next blog post will focus on rooms 4, 5, and 7, which discussed new technology solutions. 

Room 1: The Energy Grid

This discussion focused on the grid more broadly, guided by expert speakers in the field. 

Those experts included Olya Irzak, co-founder of Diamond List, CEO of Frost Methane Labs, and Advisor of Strong Atomics; Troy Hodges, Data Science Manager of grid analytics at Kevala, inc.; Eric Wallace-Deering, Manager at corporate decarbonization start-up Pilot44; and Dr. Kristin Guilfoyle, Project leader at the National Renewable Energy Laboratory (NREL). They guided the discussion through its context, challenges, stakeholders, and impacts.

The grid infrastructure is aging and reaching capacity limits. 

• Aging infrastructure includes transformers and physical space they occupy 
• To upgrade infrastructure, it costs a lot of money. Companies may face having to increase costs for consumers to cover infrastructure upgrades. 
• Utility companies need to balance reliability with cost.

Grid monitors lack real-time data for grid management. 

• Grid monitors can’t easily make decisions about grid stability and capacity due to a lack of data. 
• The system is reactive rather than proactive – operators should be able to predict and prevent problems. AI is a tool that could leverage existing data to make better predictions in order to increase proactive behavior. 
• This lack of data prevents full integration of renewable energy sources like rooftop solar power that can support greater grid flexibility. 

The power grid is vulnerable to natural disasters. 

• The prevalence of hurricanes, wildfires, and floods put the grid at risk for outages and damage.
• Grid systems aren’t designed to handle these events frequently, leading to power outages. 
• Grids need better predictive tools to help anticipate disaster and allowing for preemptive actions like reinforcing infrastructure. 

The grid can’t manage bidirectional power flow. 

• Current grid power flow is centralized, with power plants directing energy to consumers. 
• Distributed Energy Resources (DERs) are technologies like solar power and home battery systems, and these DERs allow energy to flow back into the grid. 
• The grid lacks the infrastructure to handle this, leading to inefficiency and an inability to integrate renewable energy. 

The room’s attendees concluded that grid systems worldwide must be transformed to adapt to growing power requirements. AI could help manage real-time grid operations and predict vulnerabilities to disasters. However, the physical infrastructure, both aging and lacking bidirectional power flow, requires collaboration between utilities, grid operators, regulators, and consumers.  

Room 2: Interconnectedness and Distributed Energy Resources

This session focused on the grid’s interconnection with renewable energy, and distributed energy resources, or DERs, which include rooftop solar, electrical vehicles, and battery storage that all decentralize the grid. 

Speakers were experts in energy, and included Zach Birnholz, climate action instructor at Terra.do; Archy de Berker, co-founder of Axle Energy; Dan Chapman, supporting advancement of technology into energy infrastructure at Harvest Venture Builders; and Lyon Lay, working with Voltus to decarbonize the grid. 

Grids lack granular data for flexibility. 

• The grid needs fine-tuned real-time data to manage all of the DERs available today, and without data, grid operators can’t fully integrate DERs into their operations. 
• Energy created by DERs is currently underutilized because of the lack of data. 
• Companies including Voltus and entities like New York State Energy Research and Development Authority (NYSERDA) are working on projects to support DER integration, some involving AI to manage data. 

DERs use different “communication protocols” for transmitting data. 

• Grid operators must create case-by-case solutions due to a lack of standardization on how DERs communicate. 
• The lack of unification slows down scaling of DERs. 
• Manufacturers and grid operators must collaborate to standardize protocols. 

Renewable energy needs to interconnect with the grid quickly, but there are bottlenecks. 

• Currently, most renewable projects must go through lengthy reliability studies to ensure that adding to the grid will not destabilize the system, but this process is extremely slow.
• We can’t afford to delay interconnection of renewable projects because we need to move away from fossil fuels as soon as possible. 
• Streamlining the process requires intense collaboration between energy developers and regulators. 

 DER owners don’t currently benefit from the potential of dynamic grid pricing (adjusting electricity prices based on real-time supply and demand). 

• People may be incentivized to invest in DER if they could adjust to real-time price signals and be compensated for providing flexibility in the system through their DERs.
• Dynamic pricing encourages grid flexibility by allowing DER owners to be compensated for their grid contributions.
• Grids should be more resilient to fluctuations in renewable energy generation through this model. 

The discussion highlighted the potential transformation that interconnection and integration of DERs can enact on the energy grid if we can manage to create opportunities and remove barriers. 

Room 3: Enablement

In this discussion, expert speakers and participants focused on the bureaucratic barriers that prevent efficient energy deployment. Those may include regulations, permitting, and other policy frameworks, which all contribute to clean energy enablement. As described previously, clean energy projects are frequently held up by bottlenecks that increase costs and slow the process of decarbonization. 

Enablement experts included Crystal Soo, an energy consultant at Atrium Economics; Kristin Landry, consultant and engineer; Keith Benes, Senior Fellow at the U.S. Department of Energy; and Tom Konrad, a hedge fund manager, writer, and coach specializing in renewable energy. 

Delays in permitting and regulatory approval at the local level create inefficiencies. 

• Local communities often oppose renewable projects like wind farms because of NIMBYism (Not in My Backyard) – a phenomenon in which people don’t approve of transformative projects coming to their own communities, despite supporting them in theory. 
• Processes are often unclear and vary greatly depending on location.
• Organizations like Lawrence Berkeley National Laboratory conduct research on combatting NIMBYISM. 

Regulators don’t always give a reason for rejecting innovative development projects.

• No centralized database tracks projects that are rejected. 
• Without knowing what went wrong, developers may repeat the same issues. 
• A central database will save time and money, making the whole process more efficient. 

Required interconnection studies create delays. 

• The studies assessing adding new projects to the grid are expensive and time-consuming.
• Projects can be cancelled if they trigger a need for grid updates. 
• This process perpetuates reliance on fossil fuels for even longer.

Lack of automation creates delays in project management. 

• Rapid transformation of energy means managing dozens of interconnection projects. 
• Current project management tools can’t handle the load, leading to errors. 

It’s unclear who is responsible for the cost of upgrading the grid. 

• Project developers often bear the full cost, which is unfair and disincentivizes developers. 
• Cost-sharing mechanisms have not been implemented. 

Renewable energy developers face a swath of difficult regulatory problems based on outdated processes. Speakers agree that to move forward, we must rethink the whole process and make it more streamlined, given the total transformation required to move away from fossil fuels. 

Room 6: Forecasting supply, demand, and prices

This discussion focused on forecasting supply, demand, and prices of energy. It included expert speakers from clean energy companies and energy experts. 

Speakers included Sean Kelly, Co-Founder and CEO of Amperon; Tom Walkinshaw, an energy professional working on large-scale clean energy projects; and Spencer Kuzmier, Co-Founder and Managing Partner at Cosine Energy. 

Grid operators must rely on forecasting data from energy utilities and government agencies to make decisions on managing resources, planning for spikes, and optimizing operations. 

• Current forecasting has limitations because it doesn’t account for changes involving renewable energy.
• The forecasts also rely on historical data which is no longer relevant, especially as electrification grows. 
• Optimizing solar, wind, and battery storage requires better forecasting models. 

Renewable energy output is unpredictable, which causes price volatility. 

• Participants in the energy market need to be able to plan and trade effectively. 
• Price volatility disincentivizes investment in energy projects. 

Uncertainty is hard to communicate well.

• It’s challenging for stakeholders to understand renewable energy forecasts as a probability distribution rather than deterministic predictions. 
• The lack of certainty increases reliance on fossil fuels, since they will often be used when there is any uncertainty at all. 
• Forecasts need to better explain what’s really happening to all stakeholders. 

Current AI models predicting grid events may be based on old data.

• AI Forecasting uses historical data, which may not be the best option as the grid landscape changes. 
• Past trends may no longer predict future events. 
• Physics-based AI models could incorporate real-time grid behavior in their predictions. 

As the energy landscape transforms, everything about the status quo will change. Participants reflected on how tradition may be out the window and we may be lost without improving our forecasts for energy supply, energy demand, and renewable output. 

Rooms 1, 2, 3, and 6 delved deep into reflection on the challenges we face with our energy grid. These participants and experts learned from one another that total transformation is the way forward – we know the problems, and we have the solutions. Now all there is left to do is act, and we can use new tools like AI to help us.