This project uses extremum-seeking control (ESC) to demonstrate model-free, sensor free, algorithm-driven, real-time optimization (RTO) for packaged heating, ventilation, and air-conditioning (HVAC) systems. This project deploys RTO within a wide range of representative building types across the U.S. Army Reserve (USAR) to facilitate easy transition throughout Department of Defense (DoD) installations. The RTO improves energy efficiency, system resilience, and reliability by overcoming labor and cost hurdles of traditional manual approaches to optimizing building systems due to high costs, knowledge gaps, and lack of dedicated staff. In contrast to existing RTO methods, the ESC enabled RTO is cost-effective, easy to deploy, and sustains long-term benefits. The benefits of ESC enabled RTO will be showcased on two aspects of packaged HVAC operation where there is a large potential for energy savings through improved life-cycle management and more resilient operation: (1) sensor-free economizer operation; (2) minimization of reheat in systems connected to variable-air-volume boxes. Pacific Northwest National Laboratory, in partnership with USAR, will build off its successful ESC enabled RTO pilot at two USAR buildings to ensure the success of this larger scale deployment.

ESC is a cutting-edge yet simple and cost-effective technology for RTO of building energy systems. The key technology demonstrated by this project is a class of model-free ECS driven RTO algorithms with capabilities for end-user monitoring and adjustment. ESC can be deployed across different problem types with minimal configuration and setup, and it does not require mathematical models of the system, training data, or a learning period. The key differentiator of ESC is that it can optimize any cost function that can be measured or calculated without having to describe it mathematically. ESC determines the gradient for the optimization through direct perturbation of the system instead of relying on output from a large number of sensors. The algorithm then automatically moves the system in the most favorable direction to reach and maintain the optimal operating point over time. ESC enabled RTO allows building control to move beyond setpoint regulation to direct performance enhancement, such as maximizing energy efficiency and minimizing operating costs. ESC starts optimizing out-of-the-box via the principle of “perturb and observe,” where the system being optimized is subjected to a small change that is used to automatically determine the best direction to move the control variables.

Approximately 33% of all DoD’s total energy use is attributable to buildings. DoD has more than 100,000 packaged HVAC systems installed across nearly 300,000 buildings. However, most of these HVAC systems are identified as operating at low efficiency levels due to broken systems and components, poorly configured and tuned controllers, and a lack of dedicated labor and expertise to optimize the operations. The project team will apply the ESC-based RTO to the challenge of economizer control and reheat minimization, where optimal operation has the potential to significantly reduce cooling energy (by more than 20%). The technology also minimizes or eliminates the need for sensor measurements, increasing system reliability and robustness. The technology can potentially reduce labor costs up to 50% for system maintenance and optimization. This project is expected to have a high return on investment with minimal Environmental Security Technology Certification Program project costs to prepare and maintain the deployed solutions and reduced deployment labor leveraging USAR Enterprise Building Control System sustainment resources and remote access and deployment capabilities. (Anticipated Project Completion - 2026)