New York Hospital Deploys Beacon to Reduce Operating Costs

Nationally-ranked hospital uses machine learning to optimize its cooling system.

Learn more

$180k

Annual Savings

2500 tons

CO2e Eliminated

Overview

A cornerstone of New York’s largest healthcare provider and private employer, this hospital is a 450-bed nonprofit research and academic medical center. It consists of ten buildings comprising an entire city block, and its facilities total a combined 780,000 square feet.

THE Challenge

Reducing operating costs for the condenser water loop

As a nonprofit, the hospital needs to maximize its financial resources to offer the best patient care and enable advances in medical research. However, operating a hospital campus is extremely costly due to 24/7 operations and the need to comply with rigid standards for patient safety and comfort. Healthcare facilities consume nearly 10% of the total energy used by commercial buildings in the U.S.

The hospital’s condenser water loop alone costs $1.57 million annually in electricity and water. Tagup approached their energy management team to present our artificial intelligence software as a solution to reduce costs and carbon emissions.

LOCATION

New York City, NY

CONTROL SYSTEM

Schneider

TOTAL SQUARE FOOTAGE

780,000

HVAC EQUIPMENT

Carrier Chillers, BAC Cooling Towers

THE SOLUTION

Beacon machine learning

Tagup and the hospital’s team discussed their existing cooling system equipment and controls to get a baseline of current costs and operating strategies. We discovered that the buildings run on a series of three condenser water loops, and their established control strategy primarily used fixed setpoints.

Tagup integrated its Beacon machine learning software with the hospital’s building management system in a pilot program targeting one of the three condenser loops. This pilot proved potential energy and water savings that could be realized under an ML-based control scheme.

‍

We’ve seen a meaningful reduction in our system costs. The setpoint recommendations are easy to interpret and implement. We’re looking forward to seeing how the system improves further over time.

Assistant Director of Engineering

‍

Beacon uses machine learning to dynamically adapt equipment operation and maximize the efficiency of cooling systems. Its algorithms predict how the system would behave if operated differently based on responses to changes in weather and load. After an accurate picture of performance has been captured, Beacon recommends operational changes that would improve efficiency, or even use closed-loop control to automatically adjust setpoints.

‍

THE SUCCESS

Uncovering immediate value

Beacon used machine learning to recommend a new setpoint control strategy targeting the hospital’s condenser water supply. This approach takes into account external factors and flexibly balances chiller and tower settings based on measured consumption values. For the one loop alone, Beacon identified savings of $39-44k per year. That success has expanded our deployment to include the other two condenser water loops, where savings jump to an estimated $180k annually, or 11-12% of condenser water loop costs.

As a result of these efficiencies, the hospital will reduce its carbon emissions by nearly 2,500 CO2e metric tons. This is a significant step toward compliance with New York’s Local Law 97, which requires buildings to meet minimum energy performance standards by 2024, and cut greenhouse gas emissions 40% by 2030.

Of note, Beacon is still continuously learning and optimizing the hospital’s setpoints in real-time, based on real world conditions. The machine learning software continues to improve as it ingests more operating data.

In addition to realizing energy savings, it reduces equipment maintenance and replacement costs by running the system at peak efficiency.

‍