Chiller types and selection: centrifugal, screw, scroll

What a chiller does, and the two axes that pick one

A chiller is a machine that makes chilled water. It runs a refrigeration cycle to pull heat out of a water loop, sends that cold water to coils in air handlers or to a process, and dumps the heat it collected somewhere outside the building. The cold water is the product. Everything else is in service of making it at the temperature, the flow, and the cost the building needs.

Selecting a chiller comes down to two independent choices, and people tangle them together at their peril. The first is the compressor, the part that does the work of moving and compressing refrigerant: centrifugal, screw, scroll, reciprocating, or no mechanical compressor at all in an absorption machine driven by heat. The second is how the chiller rejects the heat it collected, either to outdoor air with fans (air-cooled) or to water through a cooling tower (water-cooled).

Those two axes are nearly orthogonal. You can get a screw chiller air-cooled or water-cooled. You can get a centrifugal almost always water-cooled. So the right way to read a chiller selection is as a pair: the compressor type and the heat-rejection method, chosen against the load, the site, and the operating cost. How that chiller fits into a full plant, and how it stacks against direct-expansion cooling, are their own decisions covered in the chiller-plant and chilled-water-vs-DX guides.

What are the types of chillers?

Chillers are typed by their compressor first. The five families are centrifugal, screw, scroll, reciprocating, and absorption, and they sort roughly by capacity from small to large with absorption sitting across the range on its own logic. The second type axis is heat rejection, air-cooled or water-cooled, which is a separate choice layered on top of the compressor.

Centrifugal uses a spinning impeller to add velocity to the refrigerant, then converts that velocity to pressure. It is a dynamic compressor, and it owns the large-tonnage chilled-water plant. Screw and scroll are positive-displacement machines that trap a pocket of gas and squeeze it: screw with a pair of meshing rotors in the mid to large range, scroll with two interleaved spirals in the small to mid range, usually in modular banks. Reciprocating uses pistons and is mostly legacy at this point. Absorption replaces the electric compressor with a heat-driven chemical cycle.

The table below is the thirty-second version. Treat the capacity bands as typical, because every manufacturer publishes its own line and the boundaries overlap heavily.

Centrifugal chillers: the large-tonnage standard

A centrifugal chiller compresses refrigerant with a spinning impeller rather than by trapping and squeezing a pocket of gas. The impeller flings refrigerant outward, adding velocity, and a diffuser around it turns that velocity into pressure rise. There are no pistons and no meshing rotors in the gas path, which is why a centrifugal runs smoothly and quietly and why it scales to capacities the displacement machines cannot reach efficiently.

This is the machine you find under the big chilled-water plants: high-rise offices, hospitals, universities, large data centers. At full load a water-cooled centrifugal is the most efficient compressor type per ton, and the gap widens as tonnage grows. The trade-off is that a centrifugal is a high-flow, dynamic machine, so it behaves differently at the low end than a displacement compressor does.

Surge is the failure mode that defines the type. When the load drops too far or the lift gets too high, the impeller can no longer hold pressure against the condenser, and flow reverses momentarily, then snaps forward again. You hear it as a rumble or a bark, and repeated surge hammers the bearings and thrust faces. Centrifugals manage it with inlet guide vanes and, on variable-speed machines, with speed, but the operator and the commissioning agent both watch the surge line. The chiller-plant guide covers surge as an operating and acceptance concern in detail.

Magnetic-bearing oil-free centrifugal chillers

Magnetic-bearing centrifugals levitate the compressor shaft on a magnetic field instead of riding it on oil-lubricated bearings. No bearings touching means almost no friction and, the part owners care about most, no oil system. The oil pump, the oil heater, the oil filters, and the slow loss of efficiency as oil migrates into the heat exchangers all go away.

Two things follow. Part-load efficiency is unusually strong, because the variable-speed permanent-magnet motor and frictionless shaft lose very little at reduced speed, which is exactly where a chiller spends most of its life. Manufacturers publish AHRI-certified IPLV figures for these machines that reach into the low 0.3s kW per ton and below on water-cooled lines, well under a conventional high-efficiency screw. Maintenance also drops, since there is no oil to sample, change, or chase out of the tubes.

The cost is first cost and the smaller frame's tonnage ceiling per compressor, which the manufacturers address by running several compressors in one machine. For a plant that runs long hours at part load, which is most of them, the operating-cost case is usually the one that decides it. Confirm the certified IPLV at your conditions, not just the headline number, because the rating point and the actual lift change the result.

Screw chillers: the mid-range workhorse

A screw chiller compresses refrigerant between a pair of meshing helical rotors that trap gas and progressively squeeze it as the rotors turn. It is a positive-displacement machine, so unlike a centrifugal it does not surge, and it holds capacity and pressure well across a wide load range. That tolerance is why the screw is the default choice through the mid to large range where a centrifugal is oversized and a scroll bank is getting unwieldy.

Screws come air-cooled and water-cooled, which is part of why they show up everywhere. The air-cooled screw is a common packaged outdoor chiller for mid-size commercial buildings that do not want a cooling tower. The water-cooled screw competes with the smaller centrifugals on tonnage but wins where the load profile is rough, the lift is high, or surge-free operation matters more than the last bit of full-load efficiency.

Screws use a slide valve or variable-speed drive to unload, and good part-load efficiency is one of their selling points. The maintenance item that defines the type is the oil and the bearings: a screw is an oil-flooded machine, the oil seals the rotor clearances and carries heat, so oil analysis and oil and filter changes are part of owning one. Let the oil degrade or the charge run low and you lose efficiency first and the rotors later.

Scroll chillers: the small-to-mid modular machine

A scroll compressor nests one fixed spiral inside one orbiting spiral, and as the moving scroll orbits it pushes gas inward through shrinking pockets until it is compressed at the center. Scrolls are compact, quiet, and have few moving parts, which makes them reliable and cheap to build. The catch is that a single scroll is small, so scroll chillers reach useful tonnage by ganging several compressors together.

That gives the scroll chiller its real strength, which is modularity. A scroll-array or modular chiller stacks multiple small refrigerant circuits in one frame, or even multiple shippable modules piped together, and stages compressors on and off to follow the load. Lose one compressor and the others keep running, so redundancy is built into the architecture rather than bolted on. For small to mid buildings, and for jobs where the equipment has to come up a freight elevator in pieces, the modular scroll is hard to beat.

Where scroll falls behind is large-tonnage efficiency. A bank of small compressors will not match a single large centrifugal at design load. But because the bank stages, its part-load behavior is good, and most buildings live at part load. Maintenance is light: scrolls are usually sealed or semi-hermetic, so there is no open oil management like a screw, and a failed module is replaced rather than rebuilt in place.

Reciprocating chillers: the legacy piston type

Reciprocating chillers compress refrigerant with pistons in cylinders, the same way an old automotive compressor or an air compressor works. They were the standard small and mid chiller for decades and you still find them running in older plants. New selections rarely land here, because the scroll has taken the small end and the screw has taken the mid, both with fewer moving parts and better part-load efficiency.

Per ton, a reciprocating machine is generally the least efficient of the mechanical compressors, and it has the most wearing parts: valves, rings, rod bearings. If you inherit one, the questions are whether parts are still available, what the kW per ton actually is at your conditions, and whether a scroll or screw replacement pays for itself in energy and maintenance. Often it does, which is why the population keeps shrinking.

What is an absorption chiller?

An absorption chiller makes chilled water without a mechanical compressor. Instead of an electric motor driving an impeller or rotors, it is driven by heat, commonly steam, hot water, or a direct gas burner, using a lithium-bromide and water cycle. Water is the refrigerant. Lithium bromide is the absorbent that pulls water vapor out of the evaporator, and applied heat boils the water back out of the solution to run the cycle. The only meaningful electric load is small pumps.

The reason to choose one is the energy source, not efficiency. An absorption chiller's coefficient of performance is far below an electric machine's: a single-effect unit runs a thermal COP around 0.7, and a double-effect, driven by higher-temperature steam or hot water, reaches roughly 1.4, against electric chiller COPs several times higher. So absorption only pencils where the heat is cheap or already there: waste heat from a generator or process, a combined-heat-and-power plant, a campus steam loop in summer, or a site where gas is cheap and electricity is expensive or supply-constrained.

Single-effect machines run on lower-grade heat, roughly low-pressure steam or hot water in the 200 to 240 degrees F range, while double-effect machines want higher-temperature steam or hot water and return the better COP. Absorption units are physically large for their tonnage and need careful water-side and crystallization control, so they are a deliberate engineering choice, not a default. Verify the available heat source temperature and quantity against the manufacturer's data before committing, because the machine is only as good as the heat feeding it.

What is the difference between air-cooled and water-cooled chillers?

The difference is where the chiller dumps its heat. An air-cooled chiller rejects heat straight to outdoor air with a condenser coil and fans, all packaged in one unit that sits outside. A water-cooled chiller rejects heat into a condenser-water loop, which carries it to a cooling tower where it evaporates off to atmosphere. That single difference drives efficiency, footprint, cost, and what it takes to maintain the plant.

Water-cooled is more efficient, and the reason is physics, not marketing. A cooling tower rejects heat by evaporation, so it can drive the condensing temperature down toward the outdoor wet-bulb, which is well below the dry-bulb the air-cooled coil is stuck with. Lower condensing temperature means less lift for the compressor and less work per ton. Water-cooled plants commonly run well under 0.7 kW per ton at design, while air-cooled units typically sit around 1.0 to 1.4 kW per ton, with the exact figures set by the equipment and the conditions, so use the manufacturer's AHRI-certified ratings.

Air-cooled wins on simplicity and cost. No tower, no condenser pumps, no water treatment, no Legionella program, no freeze protection on a tower loop. For small and mid buildings, and anywhere water is scarce or the operating staff is thin, the simpler air-cooled machine often makes more sense than the few points of efficiency a tower would buy. The crossover, and the tower's own care and feeding, get the full treatment in the chilled-water-vs-DX comparison and the cooling-tower material.

Which chiller is most efficient?

The most efficient chiller, at scale, is a water-cooled centrifugal, especially a variable-speed or magnetic-bearing machine. At large tonnage it converts the least electrical energy into the most cooling, which is why it anchors big chilled-water plants. But efficiency is not one number, and the type that wins at 1,000 tons is not the one you would put on a 60-ton building.

Two metrics matter. The kW per ton is the electrical power drawn for each ton of cooling produced, lower is better, and it is how the trade talks. COP, the coefficient of performance, is the dimensionless ratio of cooling output to electrical input, higher is better, related to kW per ton by the constant 3.516, so a chiller at 0.6 kW per ton is running a COP near 5.9. EER shows up too, mostly on smaller equipment. They all describe the same thing from different directions.

Type ranks roughly like this at full load: water-cooled centrifugal best, then water-cooled screw, then air-cooled screw and scroll, with reciprocating behind them and absorption lowest on a thermal-COP basis because it is not playing the same game. That ranking is a starting point, not a spec. The number that counts is the manufacturer's AHRI 550/590-certified rating at your design conditions, because lift, water temperatures, and tonnage all move it.

Why does part-load efficiency matter more than full load?

Chillers almost never run at full load. The plant is sized for the hottest design hour, and that hour happens a handful of times a year, so the machine spends the overwhelming majority of its life somewhere between 30 and 80 percent. That is why part-load efficiency, not the full-load nameplate, usually decides the operating cost.

AHRI 550/590 captures this with the Integrated Part-Load Value, the IPLV, a single number that blends the chiller's efficiency at 100, 75, 50, and 25 percent load. The weighting is heavy in the middle: the standard formula puts only 1 percent of the weight at full load and 42, 45, and 12 percent at the three lower points, which reflects how a typical commercial building actually runs. When the design conditions differ from the standard rating point, the same method produces an NPLV, a non-standard part-load value computed at the project's own water temperatures.

Variable-speed drives are what make part-load efficiency strong. A VFD slows the compressor as the load falls, and on a centrifugal it also helps hold off surge; on a screw it beats throttling with a slide valve alone. Pair a variable-speed chiller with lower condenser-water temperature when the weather allows, and the part-load numbers improve further. Read and compare the IPLV or NPLV, not the full-load figure, when two machines look close, because the part-load number is the one the building will actually live with.

Capacity ranges by chiller type

Tonnage is the first filter in a selection, and it knocks out whole families fast. Below a hundred tons or so you are in scroll and small reciprocating territory. Through the mid-hundreds you are looking at screws and the smallest centrifugals. Above roughly 500 to 1,000 tons the centrifugal takes over and stays there to the top of the range. Absorption spans much of the band but is chosen for its heat source, not its size.

The bands below are typical and overlap on purpose. One ton of refrigeration is 12,000 BTU per hour, the heat to melt a ton of ice in a day, so a 500-ton chiller moves 6,000,000 BTU per hour. Match the family to the load, then confirm the exact model against the manufacturer's selection software at your conditions, because a single line can stretch past these edges.

Chiller refrigerants and the low-GWP transition

The refrigerant in a chiller is in the middle of a generational change driven by global-warming-potential limits, and it affects the machine you buy now. The high-pressure machines, most centrifugals and all the positive-displacement types, ran R-134a for years. The low-pressure centrifugals ran R-123. Both are giving way to lower-GWP fluids, and the replacements split by safety class and pressure.

On the R-134a side, R-513A is a common drop-style replacement at roughly half the GWP and the same A1 nonflammable safety class, while R-1234ze(E) takes the GWP near one but carries an A2L mild-flammability rating that brings extra design and code requirements. On the low-pressure R-123 side, R-514A and R-1233zd(E) are the low-GWP successors, both low-pressure HFO-based fluids. Exact charges, pressures, and which refrigerant a given line uses are manufacturer-specific.

What this means in selection is that A2L refrigerants bring refrigerant-detection, ventilation, and machinery-room requirements that an A1 fluid does not, so the building and the mechanical room have to suit the choice. The regulatory phasedown schedule and the adopted edition of the safety and mechanical codes control what is allowed where, and both are moving. Verify the refrigerant, its safety class, and the current code requirements with the manufacturer and the AHJ before you lock a machine in.

The evaporator and condenser

Every chiller has two heat exchangers, and naming them straight keeps the rest of the plant clear. The evaporator is the cold side: refrigerant boils inside it and pulls heat out of the chilled water, which is the product going out to the building. The condenser is the hot side: refrigerant gives up its heat there, either to condenser water headed for the tower on a water-cooled machine, or to air across a coil on an air-cooled one.

On commercial water-cooled chillers both are usually shell-and-tube: water in the tubes, refrigerant in the shell, or the reverse, with thousands of square feet of tube surface packed into the bundle. Smaller and packaged machines often use compact brazed-plate or plate-and-frame exchangers instead. The tube design and water velocity set the approach, the small temperature difference between the refrigerant and the water leaving the exchanger, and a fouled or scaled tube bundle widens that approach and quietly steals efficiency.

Keep the two sides distinct when you talk water loops, because the evaporator carries chilled water and the condenser carries condenser water, and they never mix. The chiller-plant guide covers the heat exchangers from the startup and acceptance side, including how approach is measured and what it tells you about tube cleanliness.

The chilled-water and condenser-water loops

A water-cooled chiller sits between two water loops. The chilled-water loop runs from the evaporator out to the building's coils and back, commonly leaving the chiller near 44 degrees F and returning warmer after it has picked up the building's heat. The condenser-water loop runs from the condenser out to the cooling tower and back, carrying the rejected heat away to evaporate at the tower. An air-cooled chiller has only the chilled-water loop, since it rejects heat to air directly.

Delta-T, the temperature difference between supply and return, is the number that ties flow to capacity. A bigger delta-T moves the same tons with less flow, which means smaller pumps and pipe and less pumping energy, so design delta-T is a real lever. Plants that drift to a low delta-T, the classic low-delta-T syndrome, end up pumping too much water for the cooling delivered, which is a pump-energy and control problem more than a chiller problem.

The chiller selection and the loop design are joined at the hip: the chiller is rated at specific entering and leaving water temperatures and flows, and changing the loop changes the rating. Set the water temperatures and delta-T with the chiller selection, not after it. The chilled-water and chiller-plant guides go deeper on loop design, flow, and the pumping arrangement.

Selecting the chiller: the decision matrix

Selection runs in an order, and taking the steps out of order is how plants end up with the wrong machine. Start with the load and the tonnage, which sets the family. Then decide heat rejection, air-cooled or water-cooled, against the site, the water situation, and the efficiency target. Then weigh the energy source, which only opens the absorption door if cheap heat exists. Then layer in redundancy, space, and budget.

The honest version is that these pull against each other. Water-cooled buys efficiency but costs a tower, pumps, treatment, and operator attention. A single large centrifugal is efficient but a single point of failure, so the plant usually wants two or more machines for staging and redundancy, which nudges toward smaller units. Space and structure can rule out the big water-cooled plant on a tight site and force an air-cooled package on the roof. Budget decides whether the efficient machine that costs more up front is allowed to pay itself back over twenty years.

Run it as a matrix, not a single criterion. The mistake is letting one factor, usually first cost, override a life-cycle answer that everyone can see. Size to the real load with a sound peak calculation, not a padded one, because an oversized chiller runs poorly at part load and costs more to buy and to operate.

Staging and multiple chillers

Most plants run more than one chiller, and for good reason. A single machine has to be sized for the peak, which means it spends most of the year loafing at part load where it is least efficient, and when it trips the building has no cooling. Splitting the plant into two or more chillers fixes both: smaller machines stage on as the load grows, each running closer to its efficient band, and a failed unit leaves the rest carrying the building.

Staging is run by the plant controls under a lead-lag scheme: one chiller leads, others come on as the load rises past staging points, and the lead role rotates to even out runtime. The art is in the staging setpoints and the timing, because staging the next machine on too early wastes energy and too late lets the loop temperature drift. Variable-speed machines complicate the optimum, since two chillers sharing a load at part speed can beat one chiller running hard.

Redundancy is a design decision, not an accident. Whether the plant is N, N+1, or 2N depends on how much a cooling outage costs the building, which is a different answer for an office than for a hospital or a data center. The chiller-plant guide covers staging logic and lead-lag from the controls and commissioning side.

Free cooling and the waterside economizer

When the weather is cold enough, a water-cooled plant can make chilled water with little or no compressor running, which is free cooling. A waterside economizer routes condenser water cooled by the tower through a heat exchanger against the chilled-water loop, so the tower does the cooling the chiller would otherwise do. In a cold climate, or any plant that runs year-round like a data center, the compressor hours and energy this saves over a winter are large.

It only works when the outdoor wet-bulb is low enough for the tower to make condenser water colder than the chilled-water return, so free cooling is climate-dependent and seasonal. The plate-and-frame heat exchanger, the extra valves, and the controls to switch modes are the cost of admission. For a plant with a real cold season and a continuous load, the payback is usually short.

Selecting for free cooling is a plant-level decision that touches the chiller choice, because it favors water-cooled and a control scheme that can hand cooling back and forth between the economizer and the machines. The chiller-plant guide treats the integration as an operating and commissioning topic.

Maintenance by chiller type

The compressor type sets the maintenance burden, and it is worth knowing before you own the machine. A screw is an oil-flooded compressor, so oil analysis, oil changes, and filter changes are routine, and the oil is also a window into the bearings and the refrigerant. Neglect the oil and you lose efficiency first, then rotors.

Centrifugals with conventional oil-lubricated bearings have their own oil system to maintain, plus the bearings and, on geared machines, the gear. Magnetic-bearing centrifugals erase the oil work entirely, which is much of their appeal, leaving the motor, the variable-speed drive, and the controls. Scrolls are usually sealed or semi-hermetic with no field oil management, and a failed module is swapped rather than rebuilt. Reciprocating machines carry the most wearing parts: valves, rings, and rod bearings that need periodic attention.

Two maintenance items cut across every water-cooled type: the tubes and the refrigerant. Condenser tubes scale and foul from the tower water and need cleaning, because a fouled bundle widens the approach and steals efficiency you paid for. The refrigerant charge and any oil in it should be checked, since a low charge or contaminated oil shows up as lost capacity long before it shows up as a failure. Eddy-current tube testing and refrigerant analysis are the tools that catch both before they bite.

The data-center chiller

Data centers changed what a chilled-water plant is asked to do. The load runs every hour of every day with little of the daily and seasonal swing a building has, the cost of an outage is severe, and the density keeps climbing as racks get hotter. That pushes the selection toward water-cooled efficiency, heavy redundancy, and a plant built to run continuously rather than to ride a comfort-cooling profile.

Efficiency matters more here than almost anywhere, because the plant never stops, so a fraction of a kW per ton compounds across 8,760 hours a year. Variable-speed and magnetic-bearing centrifugals are common for that reason, paired with free cooling to cut compressor hours in cold weather. Redundancy is typically N+1 or 2N depending on the facility's tier, and the staging has to ride through a machine failure without dropping the cold aisle.

AI and high-density compute are pushing this further, toward higher chilled-water temperatures, more free-cooling hours, and liquid cooling at the rack that changes the loop temperatures the chiller is selected for. The chiller selection sits inside a larger plant and reliability design, and the chiller-plant and chilled-water guides cover the surrounding system.

Cost and life-cycle

First cost and operating cost rank differently by type, and chasing the cheap purchase is the classic way to buy an expensive machine. A scroll or air-cooled package is cheap to buy and simple to run but burns more energy per ton. A water-cooled centrifugal costs more up front, plus the tower and pumps, and pays it back in efficiency over a service life that commonly runs two decades or more.

Run the life-cycle, not the sticker. Energy is usually the largest number over a chiller's life, so a few tenths of a kW per ton across thousands of operating hours dwarfs a difference in purchase price. Maintenance, refrigerant, water, and treatment all belong in the total, as does the cost of an outage for the kind of building that cannot tolerate one.

The blunt version: the cheapest chiller to buy is rarely the cheapest chiller to own, and the building lives with the operating cost long after everyone has forgotten the purchase order. Put the efficiency premium against the years it runs, and let that decide, not the line item that lands first.

What to document

Once the plant underperforms or the machine is due for replacement, the basis-of-design is the only thing that explains why this chiller was chosen over the others. Capture the type, the capacity, the heat-rejection method, the design water temperatures and flows, the certified efficiency, the refrigerant, and the reasoning that picked it over the alternatives.

The table below is the short list of what belongs in the basis-of-design and the equipment record, so the next engineer sees not just what was installed but why.

Common mistakes

• Picking air-cooled to dodge a tower when the building needed water-cooled efficiency, or forcing a tower onto a small building that would have been fine air-cooled.
• Oversizing the chiller off a padded peak, so it runs poorly at part load and costs more to buy and operate.
• Buying one big chiller with no staging or redundancy, so part-load efficiency is poor and a single trip drops the building.
• Comparing machines on full-load kW per ton and ignoring the IPLV or NPLV the building will actually run at.
• Putting in an absorption chiller without a genuinely cheap heat source, then paying for its low COP every hour.
• Choosing the wrong family for the tonnage, like a scroll bank stretched past its efficient range or a centrifugal undersized into surge.
• Letting first cost override a life-cycle analysis that pointed at the more efficient machine.

Field selection checklist

Standards and references

AHRI Standard 550/590 is the rating standard for water-chilling packages, and it is where the efficiency numbers come from. It defines the standard rating conditions, the full-load kW per ton, and the IPLV and NPLV part-load methods, so two chillers rated to it can be compared on the same basis. Use the certified rating at conditions matching your design, because a number at the standard point may not be your number.

ASHRAE Standard 90.1 sets minimum chiller efficiency by type and size in its tables, both full-load and IPLV, and the adopted energy code commonly references it. ASHRAE Standard 15 and the mechanical code govern refrigerant safety, machinery rooms, and the detection and ventilation that A2L refrigerants trigger. ASHRAE 34 classifies the refrigerants and their safety groups. The refrigerant phasedown itself is set by federal regulation on a moving schedule.

Capacity, efficiency, and refrigerant details are properties of the specific machine, so hedge them to the manufacturer's AHRI-certified data and the equipment submittal rather than to a rule of thumb. The adopted code editions and local amendments control what is required and allowed, and they change between cycles, so confirm them with the AHJ before a selection is final.

Units, terms, and conversions

Chiller capacity and efficiency get written several ways across a drawing set, a submittal, and a manufacturer sheet, so the same machine can read differently depending on the document.

Capacity is given in tons of refrigeration, where one ton is 12,000 BTU per hour, or in kW of cooling, where one ton is about 3.516 kW. Efficiency is kW per ton (electrical input per ton of cooling, lower is better) or COP (dimensionless ratio, higher is better), related by COP equals 3.516 divided by kW per ton. EER appears on smaller equipment. IPLV and NPLV are the part-load efficiency figures from AHRI 550/590.

Ton of refrigeration

12,000 BTU per hour of cooling, about 3.516 kW, the heat to melt a ton of ice in a day

kW per ton

Electrical power drawn per ton of cooling produced, lower is more efficient

COP

Coefficient of performance, cooling output divided by power input, equals 3.516 / kW per ton

IPLV / NPLV

Integrated and non-standard part-load value, weighted efficiency at 100/75/50/25 percent per AHRI 550/590

Lift

The pressure or temperature difference the compressor works across, condenser minus evaporator

Surge

Flow reversal in a centrifugal at low load or high lift, heard as a rumble and hard on bearings

Approach

The small temperature difference between the refrigerant and the water leaving a heat exchanger

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• What is a factory authorized chiller startup?
• What is the difference between DX and chilled water cooling?
• What does a compressor do in an AC or refrigeration system?
• What is centrifugal chiller surge?
• What is a rack unit (U)?