Controlling Heat: Chillers, Air, and Water
Cooling is the discipline the building management system exists to serve, and in a heat-dense facility it is unforgiving. Every watt drawn by the workload becomes a watt of heat the mechanical plant must move from the equipment to the outside air. The BMS is the controller wrapped around that transfer. To read a facility's cooling is to read how it fights entropy — and where it is winning or losing.
Begin with the chain of heat. Inside the space, air absorbs heat from the equipment. That warm air is drawn into cooling units — computer-room air handlers or air conditioners in a raised-floor world, or fan-coils and air handlers more generally — where it passes over cold coils and gives up its heat to a chilled-water loop or a refrigerant. The chilled water carries the heat to the chiller plant, where large compressors or a refrigeration cycle lift it to a higher temperature. From there a condenser-water loop and cooling towers, or dry coolers, reject the heat finally to the atmosphere. The BMS controls every stage of this relay: fan speeds, valve positions, chilled-water temperature, chiller staging, tower operation. It is one long feedback loop from the hot aisle to the sky.
Setpoints are where control philosophy becomes visible. The supervisor holds target values — supply-air temperature, chilled-water temperature, humidity, differential pressure — and the loops drive the plant to meet them. The temptation of the past was to run cold and wet: very low supply temperatures and tight humidity, on the theory that colder is safer. The modern understanding is that this is expensive and often unnecessary; running the space warmer, within the ranges equipment tolerates, cuts the work the chillers must do. A facility's chosen setpoints are therefore a direct readout of its efficiency posture. Conservative, cold setpoints signal an operator prioritizing margin over energy; warmer setpoints within a considered band signal one that has tuned the plant against real equipment tolerances.
Humidity deserves its own attention because it cuts both ways. Air too dry invites static discharge; air too humid invites condensation and corrosion. The BMS holds humidity in a band, and doing so can consume real energy — humidifying dry air, or dehumidifying by overcooling and reheating. Wide, well-justified humidity bands are a sign of a facility that has questioned inherited assumptions; narrow bands maintained at cost are a sign of caution or of standards never revisited.
The largest efficiency lever the BMS operates is economization — using the outside environment to cool for free when conditions allow. Airside economization brings filtered outdoor air directly into the space when it is cool and dry enough, idling the compressors. Waterside economization uses cool ambient conditions to reject heat through the towers or a heat exchanger without running the chillers, or with them lightly loaded. A facility in a suitable climate that has invested in economization can run large parts of the year on dramatically reduced mechanical cooling. Whether and how aggressively a facility economizes — visible in the BMS through the changeover logic and the hours spent in each mode — is one of the clearest signals of how seriously it pursues efficiency, and how well it exploits its location.
Staging and sequencing are where the BMS shows its craft. Cooling capacity is not one big switch; it is many units that must be brought on and off in the right order and number to match a load that shifts through the day. Good sequencing runs the minimum equipment needed, loads each unit near its efficient point rather than running many units lightly, and rotates duty so wear spreads evenly and no single unit ages ahead of the rest. Poor sequencing runs too much plant against too little load, or hunts as units fight one another. The staging logic in the supervisor is a fingerprint of engineering quality.
Resilience lives in the same system. Critical cooling is built with redundancy — spare units, redundant loops, and the ability to lose a component without losing the space. The BMS orchestrates the failover: sensing a failed unit and bringing standby capacity online fast enough that temperature never breaches its limit. The tightest constraint is the speed of that response, because unlike a server the thermal environment has momentum, and a cooling loss shows up as a rising temperature that will not wait. How a facility handles a cooling failure — how quickly standby engages, how far temperature excurses, how cleanly it recovers — is written plainly in the trend logs afterward.
Read cooling, then, as a stack of decisions the BMS enforces: how cold and how wet to run, how much to lean on the outdoors, how to stage the plant, and how to fail over. Each decision trades energy against margin, and each is legible at the supervisor. A facility's cooling strategy is not a private engineering matter — it is a public statement, recorded continuously, about how well it understands the relationship between its heat, its climate, and its equipment.