Hybrid Energy System Applications to Data Centers
Location
Ada, Ohio
Start Date
9-12-2025 2:30 PM
End Date
9-12-2025 2:40 PM
Description
The increasing demand for artificial intelligence, advanced computation, and data storage has resulted in an increase in the number and size of data centers. According to International Energy Agency (IEA), data centers consumed 415 TW of electricity in 2024 (1.5% of global electricity demand), which resulted in the production of 1% of the global greenhouse gas (GHG) emissions. It is projected that by 2030, the electricity demand by data centers will double, representing just under 3% of the global GHG emissions. Additionally, concerns over data centers’ water consumption are increasing. A large number of these data centers, which use water both directly (on-site for cooling) and indirectly (off-site for electricity generation or manufacturing), are located in water-stressed regions, such as California, Arizona, and Texas. Hybrid energy systems (HESs) are particularly well-positioned to alleviate concerns with GHG emissions and water consumption, while enabling reliable and uninterrupted power to data centers. HESs combine multiple generation (such as natural gas, wind, nuclear, etc.) and storage systems (such as thermal, battery electric, etc.) and enable participation in tertiary markets (such as hydrogen generation and water desalination). The presence of multiple generation technologies enables a built-in redundancy, which is an important design consideration for data centers. As opposed to conventional systems where backup generators are often idle and used only when power failures occur, HESs enable the concurrent operation of multiple generators and storage systems at various locations when economically justified (based on price signals). Far from simply adding an independent subsystem, effective HES designs must ensure that interactions among energy technologies are synergistic, thereby economically benefiting the entire system.
To ensure economic viability and synergistic interaction among HES subsystems, a promising approach is the control co-design (CCD) of generation, storage, and/or conversion subsystems. CCD is system-level optimization strategy in which both the physical attributes (such as capacity) and control (such as dispatch, charge, and discharge) are concurrently optimized to enhance system performance. A techno-economic assessment through a net present value (NPV) objective function, along with subsystems’ basic dynamics enable performance evaluations of various HES architectures.
This study employs a CCD approach to evaluate the performance of several HES configurations, with increasing levels of complexity, consisting of small modular reactors, thermal storage systems, battery-electric storage systems, and water desalination for data center applications. Using simplified linear dynamics, the CCD problem can be solved efficiently for the entire lifetime of the HES (approximately 30 years), with hourly mesh. The impact of various policies on the design and performance of the proposed HESs, such as carbon tax and tiered water rates, along with data center-specific considerations such as on-demand available back up power, and projected demand and costs reflecting frequent modernization will shed light on potential solutions for the future of energy landscape both within the United States and globally.
Recommended Citation
Baker, Ashley, "Hybrid Energy System Applications to Data Centers" (2025). College of Engineering Student Research Colloquium. 15.
https://digitalcommons.onu.edu/eng_student_research_colloquium/2025/Presentations/15
Hybrid Energy System Applications to Data Centers
Ada, Ohio
The increasing demand for artificial intelligence, advanced computation, and data storage has resulted in an increase in the number and size of data centers. According to International Energy Agency (IEA), data centers consumed 415 TW of electricity in 2024 (1.5% of global electricity demand), which resulted in the production of 1% of the global greenhouse gas (GHG) emissions. It is projected that by 2030, the electricity demand by data centers will double, representing just under 3% of the global GHG emissions. Additionally, concerns over data centers’ water consumption are increasing. A large number of these data centers, which use water both directly (on-site for cooling) and indirectly (off-site for electricity generation or manufacturing), are located in water-stressed regions, such as California, Arizona, and Texas. Hybrid energy systems (HESs) are particularly well-positioned to alleviate concerns with GHG emissions and water consumption, while enabling reliable and uninterrupted power to data centers. HESs combine multiple generation (such as natural gas, wind, nuclear, etc.) and storage systems (such as thermal, battery electric, etc.) and enable participation in tertiary markets (such as hydrogen generation and water desalination). The presence of multiple generation technologies enables a built-in redundancy, which is an important design consideration for data centers. As opposed to conventional systems where backup generators are often idle and used only when power failures occur, HESs enable the concurrent operation of multiple generators and storage systems at various locations when economically justified (based on price signals). Far from simply adding an independent subsystem, effective HES designs must ensure that interactions among energy technologies are synergistic, thereby economically benefiting the entire system.
To ensure economic viability and synergistic interaction among HES subsystems, a promising approach is the control co-design (CCD) of generation, storage, and/or conversion subsystems. CCD is system-level optimization strategy in which both the physical attributes (such as capacity) and control (such as dispatch, charge, and discharge) are concurrently optimized to enhance system performance. A techno-economic assessment through a net present value (NPV) objective function, along with subsystems’ basic dynamics enable performance evaluations of various HES architectures.
This study employs a CCD approach to evaluate the performance of several HES configurations, with increasing levels of complexity, consisting of small modular reactors, thermal storage systems, battery-electric storage systems, and water desalination for data center applications. Using simplified linear dynamics, the CCD problem can be solved efficiently for the entire lifetime of the HES (approximately 30 years), with hourly mesh. The impact of various policies on the design and performance of the proposed HESs, such as carbon tax and tiered water rates, along with data center-specific considerations such as on-demand available back up power, and projected demand and costs reflecting frequent modernization will shed light on potential solutions for the future of energy landscape both within the United States and globally.