Waterside economizer and free cooling field guide

What a waterside economizer is

A waterside economizer is a way to make the plant's chilled water using the cooling tower alone, with the chiller compressor off or unloaded, whenever the outdoor air is cold enough to let the tower produce cold water. The name people use for it on the floor is free cooling, because the expensive part of the plant, the compressor, stops running and the tower and the pumps do the work for the price of fan and pump power alone.

The setup that does this on most plants is a plate-and-frame heat exchanger sitting between the condenser-water loop and the chilled-water loop. On a cold day the tower makes condenser water cold enough that, across the heat exchanger, it can cool the building's chilled water down to the supply temperature the coils need, and the chiller never has to run. The cooling-tower guide covers how the tower rejects heat by evaporation, and the chilled-water pumping guide covers how the water moves out to the coils. This guide is about the third piece: the arrangement and the controls that let the tower do the cooling job the chiller normally does.

The one idea to hold onto from the start is that nothing here makes cold that was not already available outside. The economizer harvests cold weather. When the air is cold and dry enough, the cooling tower can make water cold enough to cool the load, and the whole job of the design and the controls is to use that free cold every hour it is there and hand back to the chiller the moment it is not.

How does free cooling work?

Free cooling works by using the cooling tower to reject the building's heat directly, skipping the refrigeration cycle that normally lifts that heat from the chilled-water temperature up to the condenser-water temperature. In normal operation the chiller's compressor does that lift, and the compressor is the biggest electricity user in the plant. Free cooling deletes the lift when the weather makes it unnecessary.

Here is the path the heat takes. The building load warms the chilled water at the coils. That warm chilled water gives its heat to the condenser water across the plate heat exchanger. The condenser water carries the heat to the cooling tower, and the tower evaporates it off to the atmosphere. The same loops, the same pumps, the same tower are all still working. The only thing that stopped is the compressor, and with it the largest block of energy on the bill.

This only works when the tower can make condenser water cold enough. The tower can drive its water down close to the outdoor wet-bulb but never below it, so on a cold, dry day with a low wet-bulb the tower makes genuinely cold water, and on a hot, humid day it cannot. That is why free cooling is a winter, shoulder-season, and night strategy in most climates, and a year-round strategy only where the wet-bulb stays low. The hours follow the wet-bulb.

Why free cooling is worth the trouble

The compressor is where the money goes. A water-cooled chiller spends most of its energy running the compressor to lift heat from the chilled-water side to the condenser side, and on a large plant that is hundreds of kilowatts whenever the machine runs. Shut the compressor off and replace it with a heat exchanger, two pumps already running, and a tower already running, and the plant's cooling energy drops to a fraction of what it was.

The catch, and the reason free cooling is not free to design, is that it only pays where there are real cold hours and a load to use them. A building that needs no cooling in winter gets nothing from a winter economizer. The places it pays hardest are the ones with a cooling load that does not go away when the weather turns cold: data centers, telecom rooms, labs, indoor process cooling, and any interior zone that rejects heat year-round. Those loads plus a cold climate are where the economizer runs the most hours and saves the most.

On the right plant the savings are not marginal. In a cold climate a process or data center load can run on free cooling for thousands of hours a year, and those are hours the compressor would otherwise have run flat out. That is the case the economizer is built for. In a hot, humid climate with a load that disappears in winter, the same equipment may run a few dozen hours a year and never pay back, which is exactly why the climate and the load drive the decision, not the brochure.

What is the difference between a waterside and an airside economizer?

A waterside economizer makes free cooling with the cooling tower and the water loops, with no outside air brought into the building. An airside economizer makes free cooling by opening outdoor-air dampers at the air handler and bringing cold outside air directly into the supply air, so the mechanical cooling backs off. Both chase the same prize, free cooling from cold weather, but they harvest it on different sides of the system.

The split that decides which one fits is whether you are cooling a water plant or an air system, and whether the building can take outside air. A central chilled-water plant serving coils all over a building is the natural home for a waterside economizer, because the cold is delivered as cold water through the same distribution that already exists, and no outside air has to reach the spaces. An airside economizer needs the air handler to be sized and ducted for a large outside-air flow, plus relief air, and it brings in whatever the outside air carries with it: humidity, smoke, salt, dust.

Waterside wins where outside air is a problem. Data centers, labs, and any space with tight humidity control or contamination concern often cannot flood the room with raw outside air, so they take their free cooling on the water side where the outdoor air never enters the conditioned space. The airside economizer is simpler and cheaper where it fits, but it cannot serve a chilled-water plant, and that distinction, not a preference, is what picks one over the other.

When can a cooling tower provide free cooling?

A cooling tower can provide free cooling when the outdoor wet-bulb temperature is low enough that the tower can make condenser water cold enough to cool the chilled-water loop across the heat exchanger. The wet-bulb is the driver, not the dry-bulb temperature you read on a thermometer, because the tower cools by evaporation and the wet-bulb is the floor evaporation can reach.

Work the chain of temperatures backward and the requirement is clear. The chilled water has to leave the heat exchanger at, say, the supply setpoint. The condenser water has to be colder than that by the heat exchanger's approach. And the tower water has to be colder than the wet-bulb by the tower's approach. Stack those approaches up and the outdoor wet-bulb has to sit several degrees below the chilled-water supply temperature before full free cooling is possible. The tighter the approaches you designed and paid for, the higher the wet-bulb at which free cooling still works, and the more hours you get.

A common design reference point, and the one written into the energy standard, is full free cooling capability at an outdoor wet-bulb around 45 degrees F. But treat that as the design anchor, not the on-off line, because the real changeover wet-bulb depends on your chilled-water setpoint, your heat exchanger and tower approaches, and how much load you are carrying. The colder and drier the climate, the more hours the wet-bulb spends below that line, which is why the same equipment is a workhorse in a northern climate and a curiosity in a hot, humid one.

The plate heat exchanger, the indirect type

The most common waterside economizer is the indirect type, built around a plate-and-frame heat exchanger between the condenser-water loop and the chilled-water loop. The cold tower water runs through one set of channels, the building's chilled water runs through the alternating channels, and heat crosses the thin plates from the warm chilled water into the cold tower water. The two streams never mix. That separation is the whole point of going indirect.

A plate-and-frame unit is a stack of corrugated stainless plates with gaskets between them, clamped in a frame, and the two fluids flow countercurrent through alternating gaps. The corrugations turn the flow turbulent and pack a lot of heat-transfer surface into a small box, which is why a plate exchanger can make a close approach in a footprint that fits a mechanical room. Plates can be added or pulled to change capacity, and the unit can be opened for cleaning, which matters because one side of it sees open tower water.

Keeping the dirty tower water out of the chilled-water loop is the reason indirect free cooling dominates. Open cooling-tower water is full of dissolved minerals, airborne dirt, and biological growth, and you do not want any of that in the clean chilled-water loop, the coils, or the chiller's evaporator. The plate exchanger lets the tower water stay on its own side, fouling only the surfaces you can isolate and clean, while the chilled-water side stays as clean as the rest of the closed loop. It costs an approach penalty, because heat now crosses a plate wall instead of going straight to the tower, and that penalty is the price of a clean loop.

Direct free cooling and the strainer cycle

Direct free cooling skips the heat exchanger and routes the cold tower water straight into the chilled-water loop. With no plate wall in the way there is no approach penalty across an exchanger, so direct free cooling is the most thermally efficient arrangement and it makes the most free-cooling hours for a given wet-bulb. The old name for it is the strainer cycle, because a strainer or filter on the line is all that stands between the open tower water and the coils.

That is also exactly why it is the least common choice. Putting open tower water into the chilled-water loop means the dirt, minerals, and biological growth from the tower now circulate through the coils and the rest of the system, fouling everything the strainer does not catch. The filtration has to be serious and it still does not make tower water as clean as a closed loop. The fouling risk is real enough that most engineers accept the heat exchanger penalty of the indirect type to keep the loop clean.

Direct free cooling shows up where the efficiency gain is worth the water-quality fight and the system can tolerate it, or on simpler plants where the operator has made peace with the cleaning. For most commercial and institutional plants, and for anything with sensitive coils or a chiller evaporator to protect, the indirect plate exchanger is the safer call and the strainer cycle stays on the shelf.

Refrigerant-migration free cooling in the chiller

Some chillers carry their own free-cooling mode built into the machine, and it works by letting the refrigerant move without the compressor. When the condenser is colder than the evaporator, refrigerant naturally boils off in the warm evaporator, travels to the cold condenser, gives up its heat, and drifts back, all driven by the temperature and pressure difference rather than by the compressor. The trade calls this refrigerant migration, the thermo-cycle, or thermosiphon free cooling.

The appeal is that it needs no external heat exchanger and no extra valving in the water loops. The free cooling happens inside the chiller's own refrigerant circuit, with the compressor idle, when the condenser water from the tower gets cold enough. Some packaged and modular chillers are sold specifically with this feature, and on those machines free cooling is an integral mode rather than a separate plant addition.

The capacity available this way is modest compared with a full plate-exchanger economizer, because the natural refrigerant movement can only carry so much heat without a compressor pushing it, and it needs a real temperature difference between the loops to get going. It is a good fit on chillers that offer it and on plants where adding a separate heat exchanger and the valving is not practical. Where it exists it is the simplest free cooling there is, because the cooling moves through equipment the plant already has.

What is the difference between integrated and non-integrated free cooling?

Non-integrated free cooling is either-or: the plant runs on the chiller, or it runs on the economizer, and it switches between them. Integrated free cooling runs the economizer and the chiller together, with the economizer pre-cooling the chilled water and the chiller picking up only the remaining load. Integrated is the better design and it is what the energy standard now requires, because it captures the partial-free-cooling hours that non-integrated throws away.

The difference shows up in the shoulder hours. On a day when the wet-bulb is low but not low enough for the tower to carry the whole load alone, a non-integrated plant cannot use the economizer at all, so it runs the chiller at full lift while perfectly good free cooling sits unused. An integrated plant routes the chilled-water return through the heat exchanger first, lets the tower knock off whatever it can, and hands the chiller a load that is already partly cooled, so the compressor does less work. Those partial hours are far more numerous than the full-free-cooling hours in most climates, so integration is where most of the annual savings actually live.

Plumbing the integrated arrangement means putting the heat exchanger in series with the chiller on the chilled-water return, with a control valve that diverts the return through the exchanger when free cooling is available, and a sequence that ramps the chiller down as the exchanger ramps up. Non-integrated is simpler to control because it is a clean switchover, but the missed partial hours are a steep price, and on any plant of size the integrated design is the one worth building.

Sizing the plate heat exchanger and the approach

The plate heat exchanger is sized around an approach: how many degrees warmer the chilled water leaving the exchanger is than the cold tower water entering it. A close approach, on the order of 2 to 4 degrees F on a well-selected economizer exchanger, means the tower water barely has to beat the chilled-water setpoint, so free cooling stays available at a higher wet-bulb and the plant gets more hours. The cost of that close approach is a bigger exchanger with more plates and more surface.

This is the lever the design either pulls or skips. Specify a tight exchanger approach and you buy hours every shoulder day for the life of the plant. Accept a loose approach to save first cost and you have quietly shortened the free-cooling season, because the wet-bulb now has to drop further before the tower can make water cold enough to do anything. The exact approach is a selection at the design wet-bulb and flow, set in the project specification and the manufacturer's selection, not a number off a guide.

The energy standard also limits how much pressure drop the economizer can add. A water-to-water economizer heat exchanger has to keep its water-side pressure drop low, commonly cited at under 15 ft of head, or the design has to arrange the piping so the exchanger's drop does not load the pumps when the plant is running in normal mechanical-cooling mode. The practical move is isolation valves around the exchanger so it is in the flow path only when free cooling is running, and out of it the rest of the time. Size the exchanger for the approach, then make sure it is not robbing pump energy the eleven months it sits idle.

When does the plant switch to free cooling?

The plant switches to free cooling when the cold water the tower can make, given the current wet-bulb, beats what the chilled-water loop needs by enough to be worth the changeover. The usual enabling logic compares the cooling-tower entering wet-bulb, plus the tower approach, plus the heat exchanger approach, plus a small offset, against the chilled-water return temperature. When that sum drops below the return temperature, free cooling can carry load and the sequence brings it on.

The offset and a deadband keep the plant from short-cycling around the changeover point. Switch on the instant the math crosses and a plant near the line will flip between modes every time a cloud passes, hammering valves and chiller starts. A real sequence enables free cooling a degree or two on the favorable side of breakeven, holds it through a deadband, and only drops back to mechanical cooling when the wet-bulb has clearly risen past the point where the tower can keep up. On an integrated plant the handoff is gradual: the economizer takes what it can, the chiller fills the gap, and as the wet-bulb falls the chiller unloads toward off.

Raising the chilled-water setpoint at changeover is the move that buys the most hours, because warmer chilled water is easier for the tower to make. If the load will tolerate a few degrees warmer supply, resetting the setpoint up extends the wet-bulb range over which free cooling works. The controls own this changeover, and the building automation system and the sequence of operations are where it lives. The single most common reason a plant with an economizer never sees the savings is that this changeover was never written, never commissioned, or written so timidly that free cooling almost never engages.

The sequence of operation

The sequence of operation is the written logic the building automation system follows to run free cooling, and it is the part that decides whether the hardware ever earns its cost. A good sequence covers the enable condition, the staging of the heat exchanger and pumps, the position of the diverting and isolation valves, the chiller bypass or unload, and the conditions to fall back to mechanical cooling, all keyed to the wet-bulb.

Walk a typical integrated sequence. The wet-bulb falls past the enable point, so the controls open the isolation valves to put the heat exchanger in the flow path, start or confirm the condenser-water and chilled-water pumps, and run the tower fans to drive the condenser water as cold as the wet-bulb allows. The chilled-water return is diverted through the exchanger, the tower pre-cools it, and the chiller is allowed to unload as the supply setpoint is met by the exchanger alone. When the exchanger can carry the whole load, the chiller compressor stops. When the wet-bulb climbs back, the sequence reverses in order, bringing the chiller back before it drops the economizer so the loop never loses cooling.

Two things separate a sequence that works from one that looks fine on paper. First, it has to use the wet-bulb, measured or computed from dry-bulb and humidity, not the dry-bulb alone, because dry-bulb does not tell you what the tower can make. Second, it has to reset the chilled-water setpoint upward where the load allows, to stretch the hours. The controls integration here is real work, and it overlaps the chilled-water plant sequence the pumping guide describes, so the free-cooling logic and the staging logic have to be written and tested together, not bolted on after.

The cooling tower in free-cooling mode

Free cooling leans on the tower harder than normal operation does, because now the tower is the only thing making cold, not just rejecting the chiller's heat. To carry the load alone it has to make colder water, which often means running more tower cells and more fan power than the same load would need with the chiller running. A plant designed for free cooling is frequently designed with extra tower capacity for exactly this reason, so the tower can reach the cold water temperatures free cooling depends on.

The cooling-tower guide covers how the tower makes its approach to the wet-bulb and how capacity control with variable-speed fans holds a target water temperature. In free-cooling mode the target is colder than the mechanical-cooling condenser-water setpoint, so the fans run harder and the control has to be allowed to chase that colder water rather than holding the warm minimum the chiller normally wants. The chiller's minimum condenser-water temperature limit does not apply when the chiller is off, so the controls have to switch targets cleanly between the two modes.

The hard part is that free cooling happens in cold weather, which is precisely when a tower is most at risk of freezing. Running the tower hard to make cold water in a hard freeze is the situation the next section is about. A tower picked and operated for free cooling has to be a tower picked and operated for cold-weather running, and the two requirements come as a pair.

Freeze protection in cold weather

Free cooling runs the tower in the coldest weather of the year, so freeze protection is not an add-on, it is part of the design. Water standing in the tower basin can freeze, drift can ice over the fill, and ice loading can collapse fill and break louvers. A plant that runs free cooling without a freeze plan finds the plan the hard way, as a split basin and a flooded mechanical room on the first deep cold snap.

The defenses come in layers. A basin heater, usually an electric immersion heater on a thermostat, keeps the standing basin water above freezing, but it protects only the basin water, not the fill or the exposed piping. A remote indoor sump keeps the water inside where it cannot freeze. Heat trace and insulation protect exposed pipe runs. On the closed side, running glycol in the chilled-water or the economizer loop lets that loop tolerate freezing temperatures, which is common on northern plants and reduces the risk of freezing the free-cooling coils or piping. The cooling-tower guide covers the basin, the sump, and the cold-weather strategy in more depth.

The detail that bites is that glycol cuts heat transfer and changes the approach. Add glycol for freeze protection and the same exchanger and tower make slightly warmer chilled water for the same wet-bulb, which trims the free-cooling hours a little. That is a fair trade against a frozen plant, but it has to be in the design numbers, not discovered after. Plan the freeze protection with the economizer, because the season you most want free cooling is the season the freeze can kill the equipment that delivers it.

Filtration and water quality

The economizer lives or dies on water quality, because the tower water is dirty by nature and free cooling pushes more of it through the heat exchanger more of the time. Open tower water scrubs dirt out of the air and concentrates minerals as it evaporates, so without filtration the plate exchanger fouls, the approach widens, and the free-cooling hours you paid for slip away as the fouled exchanger needs a colder wet-bulb to do the same job.

Side-stream filtration is the usual answer on the condenser-water side: a fraction of the flow is continuously pulled off, filtered, and returned, which keeps the suspended solids down across the whole loop. That filtration plus the water treatment program the cooling-tower guide describes, the scale, corrosion, and biological control, is what keeps the tower side of the exchanger clean enough to make its approach. On a direct, strainer-cycle plant the filtration burden is heavier still, because the same tower water reaches the coils.

The plate exchanger has to be opened and cleaned on a cycle, and that interval depends on the water and the filtration, not a fixed schedule. A drifting approach across the exchanger is the early signal that it is fouling, the same way a drifting tower approach signals a fouled tower. Watch the exchanger approach over a season and you see the fouling coming while it is still a cleaning job, not a lost-capacity problem in the middle of a cold snap when you need every hour.

Data center free cooling and PUE

Data centers are the strongest case for waterside free cooling, because they combine a cooling load that runs every hour of the year with an energy bill where cooling is a large, visible share. The computing load does not follow the weather, so on every cold hour there is a real load waiting to be cooled for free, and a cold-climate data center can run on free cooling for thousands of hours a year. That is the application the integrated economizer was built for.

The savings show up as PUE, power usage effectiveness, the ratio of total facility power to the power delivered to the computing equipment. Cooling is one of the biggest non-computing loads, so shutting the chiller compressor off for thousands of hours pulls PUE down hard. Operators describe free cooling as the single largest mechanical lever on PUE in a suitable climate, and the integrated arrangement matters even more here than elsewhere because the load is always there to soak up the partial-free-cooling hours.

Two design choices stretch the data center economizer further. Running a higher chilled-water supply temperature, which modern IT equipment tolerates better than older gear did, means the tower can make usable cold water at a higher wet-bulb, so more hours qualify for free cooling. A wider design delta-T cuts the flow and the pump energy. Both choices push more of the year into the free-cooling column, which is why high-temperature chilled water and free cooling are designed together on these plants rather than separately.

Chilled-water reset to buy more hours

Raising the chilled-water supply temperature is the cheapest way to add free-cooling hours, because warmer chilled water is easier for the tower to make. Every degree you can let the supply setpoint climb is a degree of wet-bulb margin you hand back to the economizer, so the changeover happens at a higher wet-bulb and the season gets longer. The chilled-water pumping guide covers how the loop and the coils respond to setpoint changes, and the limit is what the coils can still do at the warmer supply.

The catch is the load's tolerance. Push the chilled-water setpoint up and the coils have less temperature difference to work with, so a coil that was marginal at design supply may not dehumidify or may not meet a space at the warmer water. The reset has to stay inside what the coils and the latent load can take, which is why chilled-water reset is usually a controlled, conditional move, raised when the load is light and the humidity is manageable, and pulled back when a space needs the colder water.

Done well, a reset that floats the chilled-water setpoint up in cool weather and only drops it when a zone actually needs colder water can add a meaningful block of free-cooling hours over a season. It costs nothing in hardware, only controls work and a coil-by-coil understanding of how warm the supply can go. On a plant with an economizer, the chilled-water reset and the changeover sequence are two halves of the same strategy, and tuning them together is where the extra hours come from.

The energy and PUE savings

The savings from free cooling are the compressor kilowatts you stop spending, multiplied by the hours the economizer runs. With the chiller off, the plant's cooling energy falls to the pumps and the tower fans, which were already running, so the marginal cost of cooling during a free-cooling hour is small. On a plant that runs thousands of free-cooling hours a year, that adds up to a large share of the annual cooling energy.

The honest framing is that the savings are entirely climate- and load-dependent, so any number has to be tied to the project. The same equipment that saves a fortune on a cold-climate data center saves almost nothing on a hot-climate office that needs no winter cooling. The payback is set by the number of qualifying hours, which comes from the local wet-bulb data, the chilled-water setpoint, the approaches you designed, and whether the load is there to use the cold. Run the hours against the local weather before you promise a payback.

Where it does pay, integration is what carries most of it. The full-free-cooling hours, with the chiller completely off, are the headline, but in most climates the partial hours, where the economizer pre-cools and the chiller does less, are more numerous and contribute as much or more over the year. A plant designed non-integrated leaves that money on the table, which is the practical reason the energy standard pushed integration in the first place.

Climate and where free cooling pays

Free cooling pays in cold and dry climates and struggles in hot, humid ones, and the dividing line is how many hours a year the wet-bulb sits below the changeover point. A common screening figure is whether the site sees a low wet-bulb, on the order of below 55 degrees F, for a few thousand hours a year, which is the kind of climate where the economizer runs enough to justify itself.

The reason it is the wet-bulb and not the temperature is that humidity sets the tower's floor. A hot, dry desert day can have a low enough wet-bulb at night for free cooling even when the afternoon is brutal, while a warm, humid coastal night may never drop the wet-bulb far enough. So the map of where free cooling works is a wet-bulb map, not a temperature map, and two cities with the same average temperature can have very different free-cooling hours.

This is why the design starts with the local weather data, the binned wet-bulb hours for the site, run against the changeover point the plant will actually use. That analysis, not a regional rule of thumb, is what tells you whether to spend on the economizer and how big to make the tower and the exchanger. A plant in the right climate with a year-round load is an easy yes. A plant in the wrong climate is a hard look at whether the hours will ever pay back the equipment.

Designing the free-cooling plant

Designing for free cooling means sizing the heat exchanger, the tower, and the loops to make cold water at the design free-cooling wet-bulb, which is a different and colder design point than the peak-cooling design point. The peak design sizes the chiller and the tower for the hot summer day. The free-cooling design sizes the exchanger and confirms the tower for the cold day when the tower carries the whole load alone, and the two design points have to both be satisfied by the same equipment.

The big choices are the approaches and the tower capacity. A tighter heat exchanger approach and a tighter tower approach both raise the wet-bulb at which free cooling still works, buying hours at the cost of a bigger exchanger and more tower. Extra tower cells give the cold-weather capacity to make the colder water free cooling needs. The chilled-water setpoint and the design delta-T set how warm the water can be and how much flow the loops carry, and warmer setpoints and wider delta-T both favor more free-cooling hours.

Hedge every one of these to the project. The exact approaches, the tower selection, the exchanger size, the setpoints, and the freeze protection all come from the project specification, the local wet-bulb data, and the manufacturers' selections at the design conditions, not from a generic number. The design that pays is the one where the free-cooling hours were actually counted against the local weather and the equipment was sized to harvest them, with the freeze protection built in from the start rather than added after the first cold snap exposes the gap.

Commissioning the changeover

Commissioning a waterside economizer is mostly about proving the changeover actually happens and that free cooling carries the load it should across the range of wet-bulbs, not just confirming the hardware is installed. The most common finding on a new economizer is that the sequence was never tuned, so free cooling either never engages or engages and immediately drops back, and the savings the design promised never show up.

The checks are specific. Force the plant into free-cooling mode at a favorable wet-bulb and confirm the isolation and diverting valves move, the exchanger comes into the flow path, the chiller unloads and stops, and the supply setpoint is held by the tower alone. Then walk it through the partial-load condition on an integrated plant and confirm the economizer pre-cools while the chiller carries the rest, because that partial mode is where most of the annual hours live and it is the mode most likely to be wrong. Confirm the fallback to mechanical cooling brings the chiller back before the economizer drops, so the loop never loses cooling during the handoff.

Beyond the modes, verify the chilled-water reset works as written, confirm the wet-bulb input the controls use is real and not just dry-bulb, and check that the changeover deadband stops short-cycling. The savings number is worth verifying against the predicted free-cooling hours once a season has run, because the only proof the economizer is paying is the actual hours it ran and the compressor energy it displaced. Skip the part-load and the changeover testing and you have commissioned the one mode the economizer rarely runs in.

Maintenance

The maintenance that keeps free cooling working is short and it is mostly about the heat exchanger, the tower, and the controls. The plate exchanger has to be cleaned on a cycle set by the water quality, because a fouled exchanger widens its approach and quietly shortens the free-cooling season. The tower needs the cleaning, treatment, and freeze-protection upkeep the cooling-tower guide details, and the freeze protection in particular has to be verified before every cold season, not after it fails.

The controls are the part that drifts in ways nobody notices. A changeover setpoint that got bumped to settle a comfort complaint and never put back, a chilled-water reset that was disabled during a troubleshooting visit and left off, a wet-bulb sensor that has drifted out of calibration, any of these can switch off most of the free cooling while the plant looks like it is running fine. The only way to catch it is to trend the free-cooling hours and the changeover behavior season over season and ask why the hours fell if they did. The economizer that was commissioned right and is being watched holds its hours. The one that was set and forgotten slides back toward running the chiller all winter.

Common failures in the field

The failures cluster, and they all end the same way, with the chiller running when the tower could have done the job for free. The first and most common is the changeover that does not work: the sequence was never written, never commissioned, or written so conservatively that free cooling almost never engages, so the equipment sits installed and idle while the compressor runs through every cold month.

The second is non-integrated control on a plant that should have been integrated, missing the partial-free-cooling hours that make up most of the savings in a typical climate. The third is the fouled or undersized plate exchanger, where a widened approach means the wet-bulb has to drop further than designed before free cooling does anything, so the season shrinks. The fourth is the freeze: a tower run hard in cold weather with inadequate basin heat, no glycol, or unprotected piping splits and floods, and now the plant has lost both the chiller and the economizer until it is repaired. The fifth is a heat exchanger and tower approach designed too loose to begin with, which caps the hours no amount of good control can recover.

Every one of these is a design or a commissioning failure, not an equipment defect. The hardware almost always works. What fails is the sequence that was never tuned, the integration that was value-engineered out, the cleaning that was deferred, the freeze plan that was assumed, and the approach that was loosened to save first cost. The economizer that pays is the one where those decisions were made deliberately and proven on the plant.

What to document

An economizer runs on a handful of design points and setpoints, and the record is what lets the next operator tell whether the plant is making its free-cooling hours or has quietly drifted off them. Capture the type, the approaches, the changeover logic, the setpoints, and the freeze protection, because those are the numbers that decide how many hours the plant runs free and the first things to drift.

Common mistakes

• Installing the economizer with no working changeover sequence, so free cooling never engages and the chiller runs all winter.
• Building a non-integrated plant where an integrated one belonged, missing the partial-free-cooling hours that carry most of the savings.
• Sizing the plate heat exchanger with a loose approach or undersized, so the wet-bulb has to drop further than designed before free cooling does anything.
• Running the cold-weather tower and loop with no real freeze protection, then finding the split basin and flooded room after the first hard freeze.
• Holding the chilled-water setpoint fixed instead of resetting it up to extend the free-cooling hours when the load allows.
• Driving the changeover off dry-bulb temperature instead of the wet-bulb, so the controls misjudge what the tower can actually make.
• Ignoring the climate and wet-bulb reality and expecting free-cooling savings where the hours never qualify.
• Letting the plate exchanger foul on poor filtration, widening its approach and quietly shortening the free-cooling season.

Field checklist

Standards and references

ASHRAE Standard 90.1, the energy standard, is where the economizer requirements live for most commercial work. It requires fluid economizers on covered systems to be capable of meeting up to 100 percent of the expected cooling load at outdoor air no warmer than about 50 degrees F dry-bulb or 45 degrees F wet-bulb, and it requires the economizer to be integrated with the mechanical cooling so it can provide partial cooling even when the chiller is still needed for the rest. It also limits the economizer heat exchanger's water-side pressure drop, commonly cited at under 15 ft, or requires the piping to keep that drop off the pumps in non-economizer mode. The exact thresholds and the list of covered systems shift between editions, so confirm them against the edition the jurisdiction has adopted and any local amendments.

The cooling tower's thermal performance traces to the Cooling Technology Institute test and certification methods, and its design wet-bulb and the systems guidance come from ASHRAE, as the cooling-tower guide covers. For data centers, the ASHRAE TC 9.9 thermal guidelines set the chilled-water and supply conditions the equipment can take, which is what makes the higher chilled-water setpoints and the long free-cooling hours possible, and Uptime Institute and similar references cover the redundancy the plant has to hold while it economizes.

The specifics that govern any real plant are the manufacturers' selections and the project documents. The chiller's free-cooling mode and minimum conditions come from the chiller manufacturer, the heat exchanger approach and pressure drop from the exchanger selection at the design conditions, and the tower approach and capacity from the tower selection at the design wet-bulb. The hours, the savings, and the approaches in this guide are typical design references, not fixed numbers. The local wet-bulb data, the project specification, and the sequence of operations control the real design. Cite the standard that governs the point, and let the manufacturers' data and the spec govern the equipment.

Units and terms

Free-cooling work mixes temperature, flow, and energy units, and the same idea reads differently across a controls screen, a manufacturer's selection, and an energy model.

Temperatures are in degrees F on most US selections and degrees C on metric ones, and the wet-bulb and the dry-bulb are both temperatures, so always note which one a number is. Water flow is gallons per minute, gpm, or cubic meters per hour on metric jobs. Cooling load shows up in tons or in Btu per hour, where a ton is 12,000 Btu per hour. Approach is a temperature difference in degrees. Free-cooling hours are the annual count of hours the economizer can run, read off the binned wet-bulb data for the site. PUE, power usage effectiveness, is the data center ratio of total facility power to computing power, and the lower it is the better the plant is doing.

Waterside economizer / free cooling

Making chilled water with the cooling tower alone, chiller off or unloaded, when the wet-bulb is low enough

Wet-bulb

The lowest temperature evaporation can reach, the floor the tower works against and the driver of free-cooling hours

Approach

The temperature difference across the heat exchanger or between the tower water and the wet-bulb; tighter buys more hours

Plate-and-frame heat exchanger

Stacked corrugated plates that pass heat between the tower water and the chilled-water loop without mixing them

Integrated economizer

Free cooling and the chiller running together, the economizer pre-cooling so the chiller does less

Strainer cycle

Direct free cooling routing tower water straight into the chilled-water loop through a filter, most efficient but fouling-prone

Changeover

The wet-bulb-based control point at which the plant enables or drops free cooling, with an offset and deadband

PUE

Power usage effectiveness, total data center power divided by computing power; free cooling lowers it

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• How does a cooling tower work?
• What is primary-secondary chilled water pumping?
• What is free cooling in a data center?
• How does an evaporative cooler work?
• What is a psychrometric chart used for?