Harnessing the Potential of AI in Industrial Refrigeration Systems
The Regents of the University of California, Santa Barbara
Recipient
Santa Barbara, CA
Recipient Location
21st
Senate District
37th
Assembly District
Active
Project Status
Project Update
During the 2025 calendar year, the project achieved three significant milestones. First, CrossnoKaye completed the full installation and commissioning of the ATLAS base-control system at Lineage Logistics' Mira Loma facility, including the deployment of power metering equipment that provides real-time facility energy monitoring through the ATLAS cloud platform. This installation established the operational foundation for all subsequent algorithm development and testing activities.
Second, CrossnoKaye and the Mira Loma site team demonstrated early AI-based load shifting and energy optimization capabilities. The team operationalized the One Button Shutdown feature, enabling automated partial and full refrigeration system curtailments without operator intervention. Testing confirmed the facility's ability to perform daily pre-cooling cycles followed by partial system shutdowns ranging from 1 to 5 hours, with analysis of temperature dynamics and usage/demand arbitrage supporting a recommended 3-hour peak-period curtailment during summer months and curtailments of 5 hours or longer as conditions permit during cooler months.
Third, the UCSB research team published two peer-reviewed academic papers advancing the project's AI algorithm development. The first, "The Price of Simplicity: Analyzing Decoupled Policies for Multi-Location Inventory Control," was accepted and presented at the IEEE Conference on Decision and Control in December 2025, providing a mathematical framework that quantifies the performance advantage of centralized control strategies over conventional distributed approaches. The second, "Cooling Under Convexity: An Inventory Control Perspective on Industrial Refrigeration," has been accepted to appear at the American Control Conference in 2026, directly advancing the project's goals for scalable dynamic adjustment of suction pressure in industrial refrigeration systems.
The Issue
California plays a critical role in the cold storage market, with nearly 400 million cubic feet of storage space supporting national and international food supply chains. Growing demand for cold storage—driven by the rise of online grocery sales (projected to grow 11.7% annually) and the need for temperature-controlled pharmaceutical storage—has increased the urgency to improve energy efficiency in these facilities.
While energy reduction and thermal load shifting have shown great potential, their widespread adoption remains limited due to several key challenges:
1) Outdated Infrastructure: The average U.S. cold storage facility is 34 years old, relying on antiquated systems to manage billions of dollars in products.
2) Lack of Scalable Algorithmic Solutions: Load shifting is critical for state-of-the-art operations but is largely confined to facilities with the expertise and resources to implement it. The operational burden of maintaining safe, efficient, and scalable load-shifting capabilities hinders broader adoption.
These challenges are compounded by highly variable operational requirements across facilities, aging infrastructure, complex rate structures, limited visibility of new solutions, compliance variations, cybersecurity risks, and a lack of skilled labor. Addressing these barriers is essential for achieving scalable, energy-efficient operations in California’s cold storage sector.
Project Innovation
Over the past five years, our partner CrossnoKaye has developed the ATLAS Platform (TRL 8), an intelligent cloud control software that transforms outdated industrial facilities into smart ones by moving control architecture from the physical plant to the cloud. ATLAS enables remote control, external data integration, 24/7 monitoring, and offers economic benefits such as energy visualization, rate optimization, and load management.
While ATLAS has proven effective for energy reduction and load shifting, significant hurdles remain for scaling these technologies across cold storage facilities. These challenges include adapting control algorithms to varying operational conditions, forecasts, rate structures, and storage availability—complex tasks that are impractical for manual implementation or traditional decision-making software, often increasing costs and errors.
Fortunately, recent advances in AI, reinforcement learning, and optimal control can address these issues. While such technologies have transformed many sectors, they have yet to be fully applied to refrigeration and heavy industry processes. This project, led by experts from UCSB and CrossnoKaye, aims to fill these gaps by deploying scalable AI-driven energy efficiency and load shifting solutions for industrial refrigeration, with the support of Lineage Logistics, a leader in cold storage.
The project will focus on two main technology areas:
1) Energy Reduction: AI-based control algorithms that optimize facility operations to minimize total energy expenditure.
2) Load Shifting: Scalable AI control solutions that safely shift refrigeration load and leverage thermal energy storage to align energy use with lower-cost periods or high renewable energy availability, reducing costs and carbon footprints.
While ATLAS has proven effective for energy reduction and load shifting, significant hurdles remain for scaling these technologies across cold storage facilities. These challenges include adapting control algorithms to varying operational conditions, forecasts, rate structures, and storage availability—complex tasks that are impractical for manual implementation or traditional decision-making software, often increasing costs and errors.
Fortunately, recent advances in AI, reinforcement learning, and optimal control can address these issues. While such technologies have transformed many sectors, they have yet to be fully applied to refrigeration and heavy industry processes. This project, led by experts from UCSB and CrossnoKaye, aims to fill these gaps by deploying scalable AI-driven energy efficiency and load shifting solutions for industrial refrigeration, with the support of Lineage Logistics, a leader in cold storage.
The project will focus on two main technology areas:
1) Energy Reduction: AI-based control algorithms that optimize facility operations to minimize total energy expenditure.
2) Load Shifting: Scalable AI control solutions that safely shift refrigeration load and leverage thermal energy storage to align energy use with lower-cost periods or high renewable energy availability, reducing costs and carbon footprints.
Project Goals
AI-Based Algorithms for Scalable and Robust Load Shifting
AI-Based Algorithms for Scalable Dynamic Adjustment of Suction and Discharge Pressure in Conjunction with Load Shifting
AI-Based Algorithms for Scalable and Robust Dynamic Compressor Sequencing in Conjunction with Load Shifting
AI-guided Recommendation Systems for Integration of Thermal Energy Storage and Behind-the-Meter Generation Technologies
Project Benefits
The goals of this project are to deploy new AI-based algorithms for load shifting in industrial refrigeration facilities that can reduce energy expenditures and greenhouse gas emissions by over 20%.
Consumer Appeal
Consumer Appeal
Provides industrial cold storage operators with a commercially available, automated AI-based energy management system that reduces operational complexity and delivers measurable cost savings with minimal staff burden.
Affordability
Affordability
Simulation analysis projects over 27% reduction in annual energy expenditures at the Mira Loma demonstration facility, with an estimated annual return on investment of ~$808,000 against an installation cost that pays back in approximately 3.5 years.
Economic Development
Economic Development
Advances scalable AI control technologies applicable across California's cold storage sector, supporting broader adoption of energy efficiency solutions across the state's nearly 400 cold storage facilities.
Environmental Sustainability
Environmental Sustainability
Demonstrated potential to reduce CO2 emissions by over 25% at the demonstration site, with the greatest impact occurring during peak summer months when grid emissions are highest.
Reliability
Reliability
Enables grid-responsive load shifting and demand reduction at industrial scale, with the Mira Loma facility averaging over 1.6 MW of demand, representing a meaningful and controllable resource for grid stability.
Safety
Safety
Automated control strategies maintain food safety temperature requirements throughout all load-shifting and curtailment events, ensuring compliance with FDA and USDA regulations.
Energy Security
Energy Security
Cloud-based monitoring and control through the ATLAS platform provides continuous 24/7 oversight of facility operations, reducing reliance on manual processes that are prone to human error.
Key Project Members
Jason R Marden
Professor, Department of Electrical and Computer Engineering
University of California, Santa Barbara
Mahnoosh Alizadeh
Professor, Department of Electrical and Computer Engineering
University of California, Santa Barbara
Corrin Terrones
Administrator
University of California, Santa Barbara
Jesse Crossno
Chief Product Officer
CrossnoKaye
Alex Woolf
Vice President, Research and Development
Lineage Logistics
Subrecipients
Crossno & Kaye Inc.
Match Partners
Tech Partner
Lineage Logistics, LLC
Tech Partner
Crossno &
Kaye Inc.
Upcoming Events
We do not have any upcoming events. Check back soon for updates!
Past Events
Browse past events to access presentation slides and recordings.
