30-40% of a commercial building's peak-season electricity can come from its chiller plant. Yet its sequencing logic often dates to commissioning day and remains largely unchanged. That combination makes plant sequencing a high-impact target for commercial building energy management.
How Commissioning-Day Sequencing Runs
A commercial chiller plant generally includes two or more centrifugal or scroll chillers, cooling towers for heat rejection, and chilled water piping to air handling units throughout the building. At commissioning, a controls engineer writes the sequence that starts the second (or third) chiller and shares load among the units in operation.
That sequence usually relies on threshold rules. If chilled water return temperature passes X degrees, start the second chiller. If leaving chilled water stays above setpoint for Y minutes, add a stage. If building load falls below Z percent, stop the lead chiller and run the lag unit. These rules are a reasonable starting point because they are written before the engineer sees how the building loads and unloads across seasons and occupancy patterns.
The rules then stay fixed. Years later, the building has two years of load history, occupancy has changed, equipment has aged, and efficiency curves may differ from those of new equipment. Still, the commissioning thresholds run unchanged. The second chiller starts at 56F return temperature whether or not that is efficient under current conditions. A fixed load split may also ignore where each unit now operates most efficiently.
Where Sequencing Wastes Energy
Chiller part-load efficiency varies with equipment type and age. Most centrifugal chillers perform best at 70-85% of rated load and relatively poorly below 40%. One large chiller at 35% load may use more energy than two smaller units at 55-65% each. The best crossover depends on each plant's actual efficiency curves, which can differ even between chillers in the same plant.
Commissioning thresholds usually stage equipment from temperature readings or estimated load percentages. They do not find current kW/ton efficiency and compare it with the result under another staging choice. That requires live performance data: compressor power, cooling tons, and calculated efficiency, showing whether adding or removing a stage lowers total plant power.
The difference between actual and best staging is clearest in mild-weather shoulder seasons. Moderate load can leave one large chiller operating inefficiently at part load, while another sequence would perform better. Demand charges are also lower then, so the waste often goes unnoticed.
Resetting Chilled Water Temperature
Chilled water supply temperature is the most direct chiller-efficiency control. Colder water requires more compressor work and raises kW. ASHRAE's reset guidance, ASHRAE 90.1, Section 6.5.4.4, recommends raising supply temperature when load allows. Doing that safely requires continuous confirmation that the current load can be served at the higher temperature.
The practical question is: what is the warmest chilled water supply temperature at which every zone holds its setpoint under current conditions? The answer shifts with outdoor air temperature, occupancy, solar load on the envelope, and each zone's thermal lag. A fixed setpoint or weather-adjusted schedule table cannot answer it precisely.
Continuously resetting chilled water supply temperature from actual zone demand can produce 5-12% lower chiller-plant electrical use during shoulder-season months. In Atlanta, with a cooling season lasting eight to nine months, those months represent a meaningful share of annual chiller runtime.
Cooling Towers in the Plant Loop
Plant efficiency also depends on the chiller's condenser water. Cooling tower fan operation determines that temperature and directly affects heat rejection. Colder condenser water improves chiller efficiency, with each degree Fahrenheit reduction in entering condenser water typically improving efficiency by 1.5-2%.
Cooling tower fans can lower condenser water temperature while limiting their own energy use. The balance matters because fans consume substantial electricity, especially at high speed. The best condenser water setpoint trades lower chiller work against added fan energy to minimize total plant power. That balance changes with wet bulb temperature, which determines tower capacity at each fan speed.
Commissioning-day tower controls often use fixed setpoints, such as maintaining condenser water at 85F through fan staging. Machine-learned optimization evaluates chiller efficiency and tower energy together, minimizing total plant kW rather than chiller kW alone. It may use lower condenser water temperatures during humid periods, when wet bulb is high and reaching 85F is harder, and higher temperatures during dry periods, when the tower can provide 78-80F with little fan energy.
The Learning Period's Value
Machine-learned sequencing needs a training period to model equipment efficiency. Over the first 30-60 days, it reads chiller performance across operating conditions and builds an empirical curve for each unit. Those curves may differ from nameplate or design values because of age, fouling, and maintenance history.
The learning period can also expose equipment anomalies. Higher-than-modeled power at a given load may point to condenser tube fouling, refrigerant problems, or mechanical degradation. Finding that signal before an alarm or capacity failure gives maintenance teams earlier notice and can reduce unplanned downtime.
In our early-access building pilots, the main operational value has come from avoiding peak kW events during staging transitions and improving shoulder-season efficiency through continuous chilled water reset. The effect is strongest where commissioning setpoints were built for Atlanta summer design conditions and never revised for mild weather.