Collaborative Policy Initiatives for Accelerating AI Adoption in the Energy Industry

Collaborative Policy Initiatives for Accelerating AI Adoption in the Energy Industry

As pioneers in digital solutions, the energy industry has a unique opportunity to leverage Artificial Intelligence (AI) in the energy transition in the coming decades and significantly reduce its carbon footprint in response to climate change implications. A significant body of research suggests that AI has the potential to revolutionize various areas, such as improving power grid efficiency, maximizing the harvesting of renewable energy sources such as wind, solar and geothermal, and minimizing fuel waste. However, a lack of collaboration between the industry and policymakers may hinder the growth and implementation of AI applications to catalyze that transition.

Policymakers should encourage collaboration among various stakeholders to ensure a smooth transition powered by AI toward environmentally friendly energy sources and autonomous operations. Here are four key issues that can benefit from strong policy support and may otherwise impede the digital transformation of the energy industry.

Modernizing Legacy Assets and Infrastructure and the Role of Policy in Facilitating AI Integration

AI and the Digital Twin paradigm can significantly enhance the planning and operation of energy facilities. While many modern energy infrastructures are designed and built with digital frameworks in mind, a substantial portion of the energy infrastructure in numerous countries is decades past its design lifetime and in deteriorating condition, leading to less than optimal efficiency. Many of these assets lack the necessary requirements for modern Internet of Things (IoT) hardware, preventing the application of advanced AI or data-driven solutions to address common risks such as corrosion, erosion, fracture damage, leaks and spills.

Aging oil and gas facilities, such as wells, storage vessels and pipelines, often have a certain level of leakage or emission. Abandoned (orphaned, plugged or inactive) oil and gas wells in particular pose a significant yet often overlooked environmental risk. According to the United States Environmental Protection Agency (EPA), there are over three and half million abandoned oil and gas wells in the United States alone with potential emission of methane, a highly potent greenhouse gas with a global warming potential many times greater than carbon dioxide.

Digital solutions can play a crucial role in monitoring and mitigating the environmental risks associated with these abandoned wells. One approach involves using remote sensing technology, such as satellite imagery or drone-based monitoring systems, to identify and track methane emissions from abandoned wells. This data can be combined with AI-driven geospatial analytics to pinpoint the location of significant leaks, enabling more targeted and efficient remediation efforts.

Another approach, IoT can be utilized to continuously monitor abandoned wells by installing low-cost, low-power sensors at well sites. Operators and regulators can gather real-time data on potential leaks and other indicators of environmental risk. AI algorithms can then analyze this data, correct for wind or temperature effects, and predict and prevent potential issues before they become critical health or environmental risks.

Another area in the energy industry that suffers from legacy assets is the aging fleet of coal-fired power plants. These facilities often face operational inefficiencies and increased emissions due to outdated equipment and systems. Digitalization, including data-driven predictive maintenance and real-time monitoring of operations, could significantly enhance the efficiency, safety and environmental performance of these plants.

To facilitate and accelerate the digitalization process, policymakers should introduce targeted policies and programs that provide financial and technical assistance to power plant operators. For example, they could establish a public-private partnership program to fund the development and deployment of digital solutions specifically tailored to the needs of aging power plants. This program could include grants, low-interest loans or tax incentives for operators who commit to modernizing their facilities with digital monitoring tools and AI-enabled optimization technologies.

Unlocking the Power of Data by Policy Frameworks in the Energy Sector

Data-driven solutions depend heavily on the availability of large quantities of high-quality, high-resolution data. Providing access to good quality data for government agencies, research institutes, scholars, students and technology startups can accelerate the development of state-of-the-art solutions. Despite this need, many oil and gas companies and nationally owned energy entities consider their data confidential, often sharing energy generation data reluctantly, if at all.

There are, however, some examples of more transparent data reporting practices. These include the Marcellus Shale Energy and Environment Laboratory in the United States, the NLOG Dutch Oil and Gas portal in the Netherlands, and the Volve field database in Norway. While these initiatives allow researchers to access and train data-driven models on real-world data, they still cover a limited range of data types.

These examples can serve as a model for other countries or companies that control vast energy resources but are restricted by strict data privacy regulations that prevent data sharing. International or multinational organizations, such as the Organization of Petroleum Exporting Countries (OPEC) and the Organization for Economic Co-operation and Development (OECD), can spearhead more transparent data-sharing practices by engaging various stakeholders and creating collaborative platforms. These platforms would enable AI research while maintaining data privacy and security, ultimately benefiting the broader energy industry.

In the renewable energy sector, particularly solar and wind power, accurate weather forecasts and resource availability predictions are crucial for optimizing power generation and grid integration. Sharing data on weather patterns, power output and equipment performance can enable the development of AI models that improve the efficiency and reliability of renewable energy systems.

A successful example of data sharing in this context is the National Renewable Energy Laboratory’s (NREL) Wind Integration National Dataset (WIND) Toolkit in the United States, which can be used as a role model for different countries. This freely available service provides high-resolution wind resource data, which researchers and industry professionals can use to develop advanced forecasting models, optimize wind farm layouts, and improve grid integration strategies.

Government policy can play a crucial role in supporting secure data-sharing initiatives and standardization of data collection and reporting in the energy sector. Policymakers can create open data platforms with specific standard formats that centralize various data types, such as solar irradiance, wind speed, oil and gas production, energy transmission and equipment performance metrics, making them accessible to researchers, technology developers and industry stakeholders. Additionally, governments can embed data-sharing requirements as part of the energy project licensing and permitting process, ensuring that valuable data is collected and made available for further innovation and development.

Building a Talent Pipeline Through University-Industry Partnerships Strengthen AI Expertise in the Energy Sector

A common challenge faced by energy companies is the lack of in-house data science expertise, which often leads them to outsource AI services. This gap between AI knowledge and domain-specific industrial knowledge negatively impacts the development and implementation of digital solutions. To address this issue, the energy industry should pursue a long-term goal of embedding AI solutions within its current processes rather than outsourcing one-off projects. This can be achieved by forming interdisciplinary teams that include both data scientists and domain experts.

To support this goal, government entities such as the U.S. Department of Energy, or equivalent organizations in other countries, can empower national laboratories to increase AI-related university-industry partnership programs and allocate more resources to universities and research institutes with computer science and AI programs. Additionally, government-funded universities should be encouraged to develop internship frameworks within industry-university partnerships. This will facilitate greater student exposure to the energy industry and provide companies with access to a larger pool of talent from local universities.

An alarming trend in recent years is that many energy sector companies struggle to attract and hire new graduates from computer science, data science and machine learning (ML) engineering programs. This is partly due to the industry’s negative image among younger generations and, to a greater extent, the allure of employment opportunities at Big Tech companies such as Amazon, Google, Meta and Tesla.

Policymakers should make immediate efforts to reverse this trend by collaborating with the energy industry and educational institutions to create a win-win strategy. This would involve providing energy companies with greater access to computer science graduates while creating better employment opportunities for students. Such efforts can help the energy sector develop a more diverse workforce and cultivate in-house data-driven and AI teams, ultimately creating a sustainable digital culture within the industry.

Demystifying AI by Promoting Public Education and Addressing Industrial Security Concerns

Since the early days of computers, fear has surrounded the potential negative or harmful aspects of AI, with concerns ranging from paranoid to legitimate. Some factions in policymaking might refuse to support the reliability of AI systems due to these concerns, potentially radicalizing the political landscape and posing a severe risk to government-backed research and technology adoption in the industry.

To address these misconceptions and alleviate concerns about safety and security risks associated with AI, close collaboration among policymakers, AI experts and the energy industry is essential. One example of advanced AI research that contributes to securing digital solutions is the development of adversarial training techniques, which enhance the robustness of AI models by exposing them to carefully crafted adversarial examples during the training phase, making the models more resistant to potential cyberattacks or manipulation.

Policymakers should create various task forces to work with the industry, carefully study the AI life cycle within the industry, and assess the vulnerabilities of AI models at different stages, such as data acquisition chain of custody, model training, model deployment, and integration in sensitive infrastructure, and model performance. Engaging in a thorough analysis of potential AI risks will help identify areas where improvements are needed to protect sensitive data and critical infrastructure. Policy support for accelerated development in active research areas, such as Human-in-the-Loop AI systems, Explainable AI (XAI), and Hybrid AI (combining physics and domain-specific knowledge with machine learning), can help shed light on some unknown and invisible dimensions of AI. These research areas aim to make AI systems more transparent, understandable and controllable by humans. Addressing the “black box” approach in many advanced machine learning algorithms can reduce public concerns and avoid potential radical opinion shifts against AI.

Once potentially harmful aspects of AI are identified, a framework can be established to minimize or eliminate risks associated with autonomous edge control systems, unmanned vehicles, and AI-enabled decision support tools in critical infrastructures, such as power plants, grid network management centers, and oil and gas facilities. This framework should include guidelines for secure data handling, robust AI model development and testing, and ongoing monitoring and evaluation of AI systems in operation.

One sensitive aspect of the potential role of AI in improving energy security and maintaining a balanced energy mix is the area where advanced AI and digital solutions can have a significant impact on nuclear energy safety. AI applications in predictive maintenance can be used to monitor and predict the behavior of nuclear reactors, providing real-time information about the reactor’s health and detecting some anomalies that might indicate potential safety risks. AI can also be used to assess the vulnerability of nuclear plants against some natural disaster scenarios such as earthquakes or tsunamis. By enhancing the safety and efficiency of nuclear energy, human-in-the-loop AI can address some public concerns and help maintain the nuclear electricity share in the global energy mix and support sustainable energy goals.

As many countries face challenges related to climate change, the energy industry will be at the forefront of utilizing new technology like AI to achieve net zero emission targets. The successful adoption and implementation of AI in the energy industry depend on collaboration among the energy industry, AI experts and policymakers. By focusing on four key areas – modernizing legacy assets and infrastructure, enabling data collection and sharing, improving university and industry partnerships, and addressing public concerns about AI safety and security – we can ensure that AI becomes a powerful tool in our transition toward a more sustainable energy future.

AI-enabled solutions can make a significant impact on the way we produce, store, transmit and consume energy. Policymakers must work closely with the energy industry and AI experts to create a collaborative environment that allows for the rapid and effective adoption of AI technologies. By doing so, we can tackle the pressing challenges of the energy transition and make substantial progress in reducing the carbon footprint of the energy sector, ultimately contributing to a cleaner, more sustainable world.

Author Profile
Product Owner - Beyond Limits

Cyrus Ashayeri serves as a product owner at Beyond Limits, where he focuses on the intersection of artificial intelligence (AI) and the energy sector. Beyond Limits is an industrial grade, AI company that utilizes a hybrid AI approach, combining data-centric techniques with embedded human knowledge to affirm trust in software-driven decisions, manage operational risk, and drive profitability. Ashayeri’s experience spanning national oil corporations, security think tanks, and academia lends him an interdisciplinary perspective on energy systems and infrastructure.

Holding a PhD in Petroleum Engineering from the University of Southern California (USC), Ashayeri specializes in data-centric modeling of hydrocarbon resources. He has been instrumental in numerous energy initiatives, including a notable decarbonization project for a leading international oil company’s operation in California. As a published author of technical works, he has extensively explored the potential of machine learning applications in oil and gas production modeling, and the role of policy in technological integration in energy transition. As an invited guest lecturer for Engineering Diplomacy classes at USC, Ashayeri continues to share insights from the energy sector’s move toward a more sustainable future with the next generation of engineers.

Related posts