Central Chilled-Water Utility Plants: Design, Control, and Commissioning

A central plant is more than a large chiller room

A central chilled-water utility plant produces cooling for one building, a campus, or a district and distributes that cooling through a hydronic network. Hospitals, universities, government installations, data centers, laboratories, and large commercial campuses use central plants because shared equipment can provide scalable capacity, centralized maintenance, operational redundancy, and opportunities for heat recovery, economizer cooling, thermal storage, and coordinated energy management.

Consolidation creates leverage, but it also couples the plant to every connected building. A temperature, flow, pressure, valve, or sequence problem at remote loads can change plant operation. The useful performance boundary therefore extends from heat production and rejection through distribution and into the coils or process loads that create the return-water condition. Optimizing only the chiller can increase pump, tower, or airside energy—or leave occupants without adequate cooling.

Define the complete system and its operating boundaries

A water-cooled plant normally includes one or more chillers, chilled-water pumps, condenser-water pumps, cooling towers, expansion and air-management equipment, water treatment, strainers, isolation and control valves, electrical distribution, metering, and a building automation system. Plate-and-frame heat exchangers may provide hydraulic isolation, waterside economizer operation, or separation between fluids. Air-cooled chillers eliminate the condenser-water loop and evaporative towers, but their performance, ambient operating limits, sound, low-ambient controls, and redundancy still require project-specific evaluation.

The metering boundary must be stated whenever capacity or efficiency is reported. Chiller kW/ton covers the chiller package within the defined electrical and thermal measurement points. Plant kW/ton should use synchronized cooling output and the electrical input of every included component—commonly chillers, chilled-water pumps, condenser-water pumps, cooling-tower fans, and other defined auxiliaries. Mixing boundaries or using unsynchronized trend samples produces a convincing number that may not describe an actual operating condition.

• Identify which pumps, tower cells, heaters, treatment equipment, controls, and auxiliaries are included in each performance metric.
• Document fluid type and concentration, because density, specific heat, viscosity, freeze protection, and heat-transfer performance change with solution properties.
• Separate rated chiller performance from installed plant performance and from the delivered performance seen at remote buildings.
• Coordinate refrigerant, pressure-vessel, electrical, water-treatment, environmental, and life-safety requirements with the applicable codes and manufacturer instructions.

Pumping architecture establishes the hydraulic rules

In a primary-secondary plant, the production loop and distribution loop are hydraulically separated by a common pipe or decoupler. Primary pumps establish flow through the operating chillers; secondary pumps respond to distribution demand. The arrangement can protect chiller flow and simplify some staging decisions, but decoupler flow direction and temperature mixing must be understood. Excess primary flow returns cold supply water to the plant. Excess secondary flow pulls warm return water toward the supply side, raising the secondary supply temperature and signaling that production and distribution are out of balance.

Variable-primary-flow plants use variable-speed pumps to serve both operating chillers and the distribution system. They can reduce equipment count and pumping energy, but the controls must respect each chiller's allowable flow range, rate of flow change, proof requirements, and minimum evaporator flow. A bypass valve, staged pump logic, or another engineered method may be required at low load. The bypass is not a substitute for verified flow measurement or a stable transition sequence.

Other plants use variable primary-variable secondary pumping, tertiary building pumps, distributed pumping, or heat exchangers to manage long distribution systems and buildings with different pressure requirements. No arrangement is automatically superior. Selection and operation should follow the actual load profile, distribution pressure, chiller limits, redundancy objectives, maintainability, and total life-cycle performance.

Flow and delta-T determine delivered cooling capacity

For water near typical HVAC conditions, cooling capacity is commonly screened as Q = 500 × gpm × ΔT in Btu/h, or tons ≈ gpm × ΔT / 24. The 500 factor is an approximation based on water density and specific heat; accurate work should use measured mass flow and fluid properties at the evaluated condition. Flow and supply/return temperatures must represent the same boundary and time interval.

Low delta-T occurs when return water is colder than expected for the delivered load. The plant must then move more water to provide the same cooling. That can increase pump energy, consume distribution capacity, force additional pumps or chillers online, and prevent a campus from delivering its nominal connected tonnage. Low delta-T is a system symptom—not a diagnosis. Causes can include excessive coil flow, poor valve authority, leaking or three-way valves, bypasses, fouled or incorrectly selected coils, airflow problems, sensor error, unstable controls, and loads operating outside their design conditions.

Corrective work begins with simultaneous measurements and a hydraulic map. Verify flow and temperature sensors, identify bypass and decoupler behavior, compare valve command with actual flow, inspect coil and airside conditions, and determine whether the problem is local, building-wide, or plant-wide. Artificially raising return temperature or imposing a flow limit without fixing the load-side cause can hide the symptom while reducing cooling performance.

Staging must follow available capacity and stable transitions

Chiller staging should be based on an engineered indication of plant load and available operating capacity, with appropriate persistence, deadband, and transition logic. A thermal-load calculation based on validated flow and temperature difference can be useful, but it must be reconciled with chiller status, operating limits, minimum loading, power, valve position, pump capacity, and the effect of bringing another machine online. Flow alone is not cooling load.

Every enable and disable sequence should define isolation-valve position, pump proof, evaporator and condenser flow, tower readiness, chiller permissives, startup timing, loading limits, setpoint coordination, failure response, and restoration. Lead-lag rotation should consider runtime, starts, maintenance availability, equipment size, and actual efficiency—not merely alternate machines on a calendar. Unequal chillers require staging logic that recognizes their individual operating maps and combinations.

• Prove the operating flow path before enabling a chiller and maintain manufacturer-required flow throughout loading, unloading, and shutdown.
• Prevent rapid staging caused by noisy load calculations, unstable differential pressure, or poorly coordinated valve and pump movement.
• Verify that a failed start or lost proof transfers demand safely without trapping flow, defeating safeties, or creating repeated start attempts.
• Remove temporary overrides and confirm automatic operation after every functional test or maintenance intervention.

Reset strategies must optimize the whole plant

Resetting chilled-water supply temperature upward can reduce compressor lift, but it can also increase required water flow, reduce coil latent capacity, increase airside flow or fan energy, and interfere with dehumidification or process requirements. The reset should respond to demonstrated load-side demand—such as representative valve position or reset requests—while preserving critical temperature and humidity conditions. The best setpoint is a system optimum, not necessarily the warmest water the chiller can produce.

Distribution differential pressure should be reset toward the lowest value that allows the hydraulically remote critical loads to maintain control. Static high pressure wastes pump energy and can worsen valve authority problems. Sensor location, valve selection, building interfaces, and the method for identifying critical demand determine whether the reset is stable and effective.

For water-cooled plants, condenser-water control is also a balance. Lower entering condenser-water temperature can improve chiller efficiency, while colder tower water may require more fan or pump energy and must remain within chiller and tower limits. Tower-cell, fan, and condenser-pump staging should minimize total plant energy while maintaining stable flow, water distribution, freeze protection, plume or water constraints, and equipment reliability. A universal sequence such as always run every tower cell or always use the coldest possible water is not a substitute for an engineered operating map.

Economizers, heat recovery, and storage expand the operating map

A waterside economizer can reduce or eliminate compressor operation when outdoor wet-bulb conditions, tower capability, approach temperatures, load requirements, and heat-exchanger performance permit. The sequence must coordinate tower operation, valves, pumps, heat exchangers, freeze protection, transition deadbands, and mechanical cooling. Partial economizer operation may be valuable even when the economizer cannot carry the entire load.

Heat-recovery chillers, absorption machines, or thermal storage can shift the plant's economic and energy optimum. Their value depends on simultaneous heating and cooling loads, available waste heat, utility tariffs, demand charges, source-energy or carbon objectives, storage losses, and operating constraints. These modes should be commissioned as complete energy paths, including their effect on conventional chillers, towers, pumps, and building supply temperatures.

Instrumentation turns performance claims into evidence

ASHRAE Guideline 22-2025 addresses instrumentation and calculation methods for central chilled-water plant thermal load, energy use, and efficiency. A credible measurement system requires suitable flow, temperature, and electrical instruments; documented accuracy; correct sensor placement; synchronized acquisition; and calculations that preserve the defined boundary. Small temperature differences make capacity calculations especially sensitive to sensor bias, so matched temperature measurements and field verification matter.

Rated full-load and part-load chiller performance under AHRI test conditions supports equipment comparison, but it does not predict every installed operating point. Field efficiency changes with load, chilled-water and condenser-water temperatures, fouling, fluid properties, flow, auxiliaries, and control behavior. Operators should review plant kW/ton against load and weather, not treat one annual average or one favorable operating point as a complete performance assessment.

• Trend simultaneous chilled-water flow, supply and return temperature, calculated load, chiller power, pump power, tower-fan power, and relevant ambient conditions.
• Track plant kW/ton, chiller kW/ton, delta-T, flow per ton, differential pressure, equipment starts, runtime, and economizer hours with explicit calculation boundaries.
• Validate virtual points and totals against the underlying sensors, meter scaling, timestamps, and equipment status before using them for acceptance or savings claims.
• Compare efficiency at similar load and weather conditions to distinguish normal operating variation from performance drift.

Commission the plant as one interacting system

Component startup verifies that individual equipment can operate. Plant commissioning proves that the assembled system meets the Owner's Project Requirements and approved sequence under the conditions that matter. Readiness includes calibrated sensors, complete TAB and hydronic balancing information, verified valve and drive operation, current chiller and tower data, water treatment, electrical and controls checkout, trend setup, safeties, alarms, and documented test prerequisites.

Functional testing should exercise plant enable and shutdown, individual and multiple-chiller operation, pump and tower staging, setpoint resets, minimum-flow protection, decoupler or bypass behavior, lead-lag rotation, economizer transitions, alarms, safeties, power or communication interruptions where required, failed starts, and recovery. Results should connect command to physical response, measured capacity, stability, efficiency, alarm indication, and final restoration.

One design-day test cannot represent a plant's full operating map. Seasonal and deferred testing, monitored trends, and periodic recommissioning are needed to verify low-load operation, high-load capacity, wet-bulb-dependent tower performance, winter economizer modes, and changes in campus demand. The turnover record should leave operators with current sequences, setpoints, control narratives, network and piping diagrams, equipment curves, meter information, backups, alarm instructions, maintenance requirements, and a repeatable method for evaluating performance.