Baltimore Aircoil Company is one of the leading manufacturers of cooling solutions for HVAC, industrial, and refrigeration equipment in the United States. With over 80 years of experience developing customized cooling solutions, they quickly recognized the need for improvement in system performance at their global headquarters.
Well aware that even the most advanced building automation systems (BAS) have limited optimization capabilities, BAC sought to develop AI-powered control techniques that pushed beyond the simple onboard electronics for cooling towers.
The end goal was to lower their system’s operating costs, and they partnered with Tagup to make it happen.
BAC’s chilled water system consists of two condenser water pumps (Patterson Pumps, 7.5 HP at 540 gpm), a centrifugal chiller (150 HP York MaxE), and a cooling tower (BAC SC3000); all managed by a Schneider Building Operation Workstation.
Aligning with conventional cooling system wisdom, the system was initially commissioned using a fixed setpoint control strategy to strike a reasonable balance among the energy needs of each component in their cooling system. This traditional approach is meant to assure stable operation and reduce maintenance requirements. It ensured average system performance, but made optimization impossible as it failed to account for external factors that affect cooling system performance:
- Component interactions are complex; changes in cooling tower fan speed or condenser pump speed impact the rate of evaporation and chiller power consumption.
- Weather, cooling load, and the price of electricity and water can vary day-to-day and throughout the day, which affects system operating costs and performance.
- Even the most thoroughly commissioned plants are vulnerable to drift over time –equipment degrades, leading to increases in operating costs.
To reach peak efficiency for their cooling system and peak comfort at their headquarters, BAC needed a new, dynamic control strategy that accounted for the existing hardware, utility rates, building load demands, and changes in weather.
With Beacon, BAC’s dynamic cooling problem gained an equally dynamic solution.
Like many cooling systems, BAC HQ lacked the instrumentation needed to support optimization. Tagup was able to implement inexpensive, non-intrusive IoT sensors to collect the basic data points needed, and Beacon was deployed within the cooling loop. After gathering data from cooling system sensors, the solution trained a machine learning model to predict how changes to operational parameters would impact system performance.
Factoring in external conditions like weather forecasts and energy and water costs, Beacon began to implement its performance recommendations, connecting with the cooling tower fan controller and adjusting the temperature setpoint for water leaving the tower.
Beacon’s real-time system modeling and machine learning-powered performance simulations offered a great opportunity for BAC to diagnose their cooling system’s performance and optimize it autonomously.
Over the evaluation period, cost savings totaled 10% compared to baseline operation, with a high daily savings of 15%. Building and weather data collected over an extended period showed that Beacon can reduce annual operating costs by 10-15% and deliver peak savings of as much as 30% under certain load and environmental conditions (see Figure 1).
Take part in the future of Industrial AI.
Schedule a call for a no-risk estimate of savings potential at your site(s).