Server immersion cooling submerges whole servers in a dielectric liquid that draws heat off every component at once, and Paul Industries installs the mechanical systems around it: rigging and setting the tanks, piping facility water to their heat exchangers or CDUs, building the heat rejection, filling and filtering the fluid, then testing, starting up and maintaining the facility side. This page is written for owners, engineers and operators planning a tank deployment. It explains the two immersion families, where the heat goes after it leaves the tank, what the floor and the room have to accommodate, how the fluid is handled, and what a maintenance program looks like. For the wider picture of liquid immersion cooling for data centers alongside cold plates and air, start with our data center hub.
What is immersion cooling?
Immersion cooling is a liquid cooling method in which servers sit fully submerged in an electrically non-conductive fluid, so heat moves from processors, memory, power supplies and drives straight into the liquid with no air path at all. The fluid is then cooled by water, and the water carries the heat out of the building. Because every surface on the board is wetted, an immersion cooled server needs no fans, no airflow baffles and no humidity control; OCP guidance for immersion hardware has fan detection disabled so the firmware does not alarm on missing fans.
Most tanks are horizontal open baths: a lidded tub, roughly the footprint of a rack laid on its back, with servers hanging vertically in slots and lifted out from the top. A smaller group of designs uses sealed chassis that slide into a vertical rack, each holding its own fluid. Immersion cooling servers are usually standard boards with a few preparations: fans removed, vented spinning disks replaced with sealed helium drives or solid-state storage, optical transceivers chosen so fluid cannot fill the air gap between ferrules, and labels, cables and thermal interface materials checked against the fluid, because some plasticizers and adhesives leach into it over time.
How does immersion cooling work, from the chip to the heat rejection plant?
Heat travels through three stages: the components heat the dielectric liquid, a liquid-to-water exchanger moves that heat into facility water, and the facility water carries it to dry coolers, cooling towers or chillers outside. Each stage has its own fluid, its own pumps or natural circulation, and its own failure modes, so they are designed and tested as separate circuits.
- Inside the bath. In a forced-flow tank the cool fluid enters low, rises through the server slots and leaves warm from the top toward the exchanger. In a natural-convection bath there are no tank pumps; warm fluid rises, the bath stratifies with the coldest liquid at the bottom, and a submerged coil or cold wall removes the heat.
- At the exchanger. The immersion cooling heat exchanger is normally a brazed or gasketed plate unit, either built into the tank skid with its own fluid pumps and filters or housed in a separate coolant distribution unit serving one or several tanks. On the water side it is the only point where the building touches the system.
- To the plant. Facility water supply and return run from the exchanger back to the central plant, through the same kind of headers, isolation valves, strainers and balancing devices used for any process load. Our data center cooling systems page covers that central plant.
The practical consequence is that the dielectric fluid never leaves the white space and the piping that crosses the hall carries ordinary treated water or glycol. That keeps the plant side familiar to anyone who has installed chilled water or condenser water, while the tank side brings its own requirements for rigging, containment and fluid care.
Single phase vs two phase immersion cooling
The two families differ in whether the fluid boils: in single-phase systems the liquid only warms up and is pumped or convected to an exchanger, while in two-phase systems it boils on the hot components and the vapor is condensed above the bath. That one difference changes the fluid chemistry, the tank construction, the condenser connection and how much attention the fluid inventory needs.
| Factor | Single-phase immersion | Two-phase immersion |
|---|---|---|
| Typical fluids | Synthetic hydrocarbons, mineral oils, natural and synthetic esters | Engineered fluorochemicals with a low boiling point |
| Boiling point | Well above operating temperature; OCP asks for 150 °C or higher at sea level | 45 to 55 °C in the OCP fluid specification, usually near 50 °C |
| Heat transfer at the chip | Sensible heat; relies on flow or convection over heat sinks | Latent heat of boiling; very high heat flux at the component surface |
| Tank sealing | Lidded to keep moisture and dust out; small evaporation | Sealed lid, condenser coil and vapor management to keep fluid in the tank |
| Fluid density | Lighter than water, typically under about 900 kg/m³ | Roughly twice as dense as hydrocarbon fluids |
| Fluid handling | Continuous filtration, periodic sampling, top-off as servers carry fluid out | Filtration to below 1 micron, desiccant and carbon media, tight control of vapor losses |
| Facility water connection | Plate exchanger on the tank skid or in a CDU | Condenser coil above the bath fed directly with facility water |
| Facility water temperature | Warm water; tested installations have run on 40 °C supply | Must stay low enough below the boiling point to condense vapor at full load |
Single phase immersion cooling in practice
Single-phase tanks appeal to many operators because their fluids are easy to contain, low in evaporation and tolerant of an opened lid during service. The design burden sits in flow distribution inside the tank and in the pumps and filters on the skid. For the facility, a single-phase tank behaves like a constant-load process heat exchanger with a fairly narrow temperature rise, which makes it easy to serve from a warm-water loop.
Two phase immersion cooling in practice
Boiling pulls heat away from the hottest parts, which keeps component temperatures close to the fluid’s boiling point even as chip power rises. The condenser coil is fed with facility water and is the only water connection, but its supply temperature has to leave a working margin below that boiling point, so two-phase systems ask for cooler water than single-phase tanks do. Every time a lid opens, some vapor can escape, which is why these tanks use sealed lids, a freeboard zone above the liquid and vapor recovery.
Immersion cooling fluid: types, properties and handling rules
The liquid in the bath has to be a good electrical insulator, carry heat well, stay chemically stable for years and remain compatible with every material on the board. The OCP Base Specification for Immersion Fluids sets the targets most suppliers now publish against, and the main chemistries are synthetic hydrocarbons, natural and synthetic esters, and fluorochemicals.
- Hydrocarbons. Synthetic hydrocarbon blends dominate single-phase tanks. Mineral oil immersion cooling appears in some installations, while refined synthetics are chosen where oxidation stability and low odor matter most.
- Esters. Natural and synthetic esters offer high flash points and good biodegradability, and some are chosen where environmental handling rules are strict.
- Fluorochemicals. Engineered fluorinated fluids are used for two-phase tanks because they boil at the right temperature and are non-flammable. Many fall under the broad PFAS definition used by regulators, any molecule containing a CF2 or CF3 group, and the OCP fluid specification flags the restrictions moving through the EU and several US states as a supply risk owners should weigh before buying a large inventory.
Selecting a dielectric fluid for immersion cooling also means checking the tank maker’s approved list and the server maker’s warranty conditions. OCP has published warranty guidance precisely because fluid choice can affect hardware coverage, and a fluid outside both approvals creates an argument nobody wants after a failure.
Fire, safety and environmental points
The OCP specification asks single-phase fluids for a closed-cup flash point, tested to ASTM D93, of 150 °C or higher at sea level and 155 °C where the site sits around 1,500 meters. Under NFPA 30 any liquid with a flash point at or above 93 °C (200 °F) is a Class IIIB combustible liquid, the lowest-hazard class, so hydrocarbon immersion fluids are combustible but hard to ignite, and the authority having jurisdiction decides how the room is classified and what suppression it needs. Ozone depletion potential is required to be zero and global warming potential low.
OCP’s immersion requirements add the housekeeping rules that matter on site: containment for at least the full volume of the largest tank, through leak trays or a double wall; enough absorbent on hand for a spill of that whole volume; a well-ventilated room as directed by the fluid’s safety data sheet; fluid kept out of floor drains and sewers; disposal through a proper waste route; and at least one person trained in spill response present during any service. Static charge can build in moving dielectric liquid, so tanks, pumps and transfer hoses are tied into the grounding system. Fluid on the floor is also a slip hazard, which is why drip trays, absorbent mats and a defined drain-down spot for wet servers belong in the layout.
Server immersion cooling facility water: temperatures and heat rejection
Immersion tanks are among the most tolerant loads for warm facility water, and in single-phase systems the supply can often run in the upper ASHRAE classes, W32, W40 or W45, so dry coolers or towers handle most or all of the year without chillers. A published single-phase study ran reliably at 40 °C facility supply, with pump power under five percent of the IT load. The actual limit for any tank is the one its manufacturer guarantees at full IT load, and immersion cooling temperature targets differ between fluids, tank designs and the chips inside.
OCP’s interface rules give the piping designer firm boundaries. The tank’s facility water connection has to tolerate anything from a partial vacuum of 50 kPa absolute to 1,000 kPa gauge, accept plain water or glycol mixtures up to 50 percent, and be marked blue for supply and red for return. Threaded or flanged connections are preferred to quick disconnects on facility water, and a side-stream strainer removing particles above 500 microns is recommended. Where metals meet at the interface, the electrochemical potential difference should stay within 0.15 V, or a dielectric break or other galvanic measure is needed.
| Heat rejection | Where it suits immersion | What to plan for |
|---|---|---|
| Dry coolers on a glycol loop | Single-phase tanks on warm supply in most climates; zero water consumption | Peak summer approach, glycol concentration, coil cleaning; see our dry cooler guide |
| Cooling towers with an isolating plate exchanger | Hot, humid sites, large halls, and two-phase condensers needing cooler water | Makeup water, treatment and Legionella control; see data center cooling towers |
| Chillers, often as trim | Two-phase tanks with tight condenser margins, or halls sharing chilled water with air-cooled rooms | Chilled water reset, bypass control, part-load efficiency; see data center chillers |
| Heat reuse loop | Warm return water for building or district heating | Matching return temperature to the heating demand, backup rejection when demand drops |
Water quality on the facility side follows the ASHRAE limits for facility water systems, and the plate exchanger in each tank or CDU fouls just like any other if the loop carries scale, corrosion products or biological growth. Where towers are involved, cooling tower water treatment protects those exchangers as much as the tower fill. Because a dry cooler loop uses no evaporation at all, immersion halls are often part of the conversation about data center water usage.
Floor loading, tank layout and service clearances
A filled immersion tank concentrates far more weight on its footprint than a loaded air-cooled rack, so a structural engineer has to confirm the slab or raised floor before layout is fixed. Manufacturers state three static loads in kilonewtons per square meter: the empty tank, the tank filled with fluid, and the filled tank with a full complement of IT. Hydrocarbon fluids are lighter than water, but fluorinated two-phase fluids are roughly twice as dense, which moves the structural answer considerably for the same tank volume.
Layout is driven by servicing from above. Each tank needs clear space on its service side for a technician and a server lift or overhead hoist, height for the lid to open fully, and a drip-drain station where pulled servers can hang until the fluid runs off. Aisles have to fit the tank on its delivery skid, and the path from the loading dock to the hall has to fit it too, including door widths, elevator ratings and any ramps. Facility water and power typically come up through the floor or drop from overhead on the opposite side from the service aisle, so pipe never crosses the lift path. Leak trays or a curbed area under the tank row satisfy the containment requirement and give a place to route condensate and spill drains away from electrical gear.
Paul rigs and sets tanks with millwright and rigging crews, levels them on their feet or plinths, and anchors them where seismic rules apply; our machinery moving teams plan the route in from the truck.
Installing an immersion cooling tank, step by step
A tank installation runs from structural sign-off to a loaded heat test, and each step has a record that goes into the turnover package. The order matters because fluid should never go into a tank whose water side has not been flushed and proven tight.
| Step | Work | Record produced |
|---|---|---|
| 1. Route and placement | Confirm floor loading, delivery path and final positions; set containment trays | Structural sign-off, layout drawing |
| 2. Rig and set | Move tanks or skids into the hall, level, anchor, bond and ground | Placement and level check, grounding check |
| 3. Facility water piping | Branch from the headers with isolation valves, strainers, flow and temperature instruments, vents and drains | Weld or joint records, valve tags |
| 4. Flush and clean | Flush the new branches with the tank exchangers bypassed until the water runs clean | Flush log with filter and turbidity results |
| 5. Pressure test | Hydrostatic test of the site piping; exchanger isolated or held within its own rating | Signed test record |
| 6. Connect and fill water side | Remove bypasses, connect exchangers or condensers, fill, vent and treat the loop | Water quality baseline |
| 7. Fluid fill | Transfer dielectric fluid through a filtered pump from sealed containers into a clean, dry tank | Batch numbers, volume, baseline fluid sample |
| 8. Circulate and leak check | Run tank pumps and filters, check every fitting and gasket on both sides | Leak walk-down record |
| 9. Startup and load test | Start with manufacturer support, load with heaters or IT, verify flows and temperatures | Startup report, alarm setpoints |
Site piping serving the tanks is tested to its code, usually ASME B31.3 or B31.9 depending on the owner’s specification, and ASHRAE TC 9.9 recommends B31.3 practice for site-installed liquid cooling piping; see hydrostatic testing. Paul works through these steps with the commissioning team, and our data center commissioning page explains how the mechanical records feed the higher test levels.
Retrofitting immersion into an existing hall vs designing for it
An existing hall can take immersion tanks when the floor can carry them, a facility water source can reach them, and the room can provide containment and service space; a new hall can be built around all three from the start. Retrofits usually begin with a structural check of the slab and a survey of the nearest chilled water or condenser water mains. Many sites tap the existing chilled water through a plate exchanger and a dedicated secondary loop so the tanks can run warmer than the air-cooled rooms; our plate heat exchanger page covers how those exchangers are selected and installed.
A purpose-built liquid immersion cooling data center can drop raised floor in favor of a slab rated for the tanks, run warm-water mains sized for the final load from day one, lay out the hall around lift paths rather than hot and cold aisles, and choose heat rejection that rarely needs compressors. The air system shrinks to what people, network gear and the remaining air-cooled equipment need. Either way, the piping is installed with valved, capped branches for future tanks, kept short so they do not become stagnant dead legs.
Immersion cooling vs direct to chip cooling
Immersion submerges the entire server in dielectric fluid, while direct to chip cooling sends treated water or a water-glycol mix through cold plates on the processors and leaves the rest of the server air-cooled. For the building, immersion means heavier, horizontal equipment and fewer, shorter pipe runs to each tank; cold plates mean ordinary racks and much more distribution piping, manifolds and fine filtration in the white space.
The two are often combined in one facility on a shared facility water system. Our data center liquid cooling page covers cold plates, CDUs and rear doors in depth, so this page stays with the tank side.
Dielectric fluid handling, filtration and condition monitoring
Immersion fluid stays in service for years when it is kept dry, filtered and tested, and the fluid supplier’s monitoring plan sets which tests run and how often. OCP expects suppliers to state those tests and intervals, typically drawn from dielectric strength, acid number, water content, particle count, flash point, viscosity, density, color and oxidation stability.
The OCP fluid specification gives the limits most programs use as triggers. Dielectric strength should stay above 6 kV per millimeter for the life of the fluid, measured to IEC 60156. Total acid number should not climb above twice the value allowed for unused fluid. Color is checked on the ASTM D1500 scale, and a rise of more than two units calls for a full fluid analysis. A flash point drop after a component overheats or arcs points to decomposition products in the bath.
Single-phase tanks run continuous filtration in the circulating loop and keep lids closed to limit moisture pickup from room air. Two-phase systems add finer filtration below one micron, desiccant media to remove water, and activated carbon to strip plasticizers and flux residues leached from boards. Servers lifted out for repair carry fluid with them, so every tank has a top-off routine using the same fluid batch family, logged by volume, because unexplained level loss is often the first sign of a leak or a vapor problem.
Immersion cooling system maintenance schedule
Maintenance covers three fronts: the dielectric fluid and its filters, the tank pumps and exchanger, and the facility water and heat rejection that serve them. The intervals below are typical starting points and are adjusted to the tank and fluid manufacturers’ instructions.
| Interval | Tank and fluid | Facility side |
|---|---|---|
| Daily (monitoring) | Fluid level, supply and return temperature, pump status and leak alarms | Facility water supply temperature, differential pressure and flow |
| Monthly | Walk-down of lids, gaskets, drip trays and drain-down area; log top-off volumes | Strainer differential pressure, valve positions, visible leaks at exchanger connections |
| Quarterly | Filter differential pressure and changes as needed; visual color check | Facility water chemistry: pH, conductivity, inhibitor level, bacteria counts |
| Semiannual | Fluid sample sent for the supplier’s test panel | Dry cooler coil cleaning or tower inspection, pump vibration and seal checks |
| Annual | Exchanger approach-temperature trend review, sensor calibration | Valve exercising, instrument calibration, heat rejection capacity check before summer |
| After any event | Fluid sample after an electrical fault, overheating or major spill | Inspection after a water-side leak, pressure excursion or freeze risk |
Paul’s preventive maintenance crews take on the facility-side tasks, the tank pump and filter service, fluid sampling and top-off, and the records that tie them together.
What drives cost in an immersion cooling data center?
The largest cost lines are the tanks, the fluid inventory, structural work and the heat rejection plant, with installation labor driven mainly by how far facility water has to travel and how difficult the delivery route is. Fluid is a significant line because each tank holds a large volume, and fluorinated two-phase fluids cost considerably more per liter than hydrocarbons.
- Structure. Slab reinforcement or replacing raised floor in a retrofit can outweigh the piping cost.
- Distance to water. Long runs from the plant, new headers, or a new isolating exchanger add the most piping labor.
- Heat rejection. Warm-water designs with dry coolers avoid chillers; two-phase or shared chilled water designs need more compressor capacity.
- Rigging and access. Cranes, floor protection, door removals and night work in a live facility.
- Containment and safety. Leak trays, spill kits, ventilation upgrades and any changes required by the fire code official.
- Fluid logistics. Delivery, storage, filtered transfer, spare inventory for top-off and eventual disposal.
- Server preparation. Conversion work on existing hardware, or immersion-ready servers from the manufacturer.
Offsetting those costs, the IT equipment loses its fans, the room needs far less air handling, and warm-water heat rejection uses less energy than a chiller plant.
How to choose a contractor for immersion cooling for data centers
Choose a contractor that can carry the work from rigging through fluid fill and startup as one scope, because most problems on immersion projects appear at the handoffs between those steps. Questions worth asking:
- Can they rig, set and level heavy skids inside a finished or operating hall without damaging floors or interrupting live equipment?
- Do they fabricate and install facility water piping to a written code, with flushing, hydrostatic test and water treatment records?
- Will they handle dielectric fluid transfer with filtered, dedicated equipment, keep batch records and take a baseline sample?
- Do they understand containment, grounding and the spill rules in the OCP immersion requirements?
- Can they size and install the heat rejection, whether dry coolers, towers or chillers, and tie it to the tank exchangers?
- Will the same team come back to sample fluid, change filters and maintain the plant after turnover?
Our data center mechanical contractors page explains how Paul organizes crews and records on mission critical sites.
What Paul Industries does on immersion cooling projects
Paul Industries delivers the mechanical side of immersion deployments nationwide. Our crews plan the delivery route and rig, set, level, anchor and ground the tanks; fabricate and install the facility water branches, isolating exchangers and headers in stainless steel or carbon steel; install dry coolers, cooling towers, chillers and pumps; flush and hydrostatically test the water side; transfer, filter and sample the dielectric fluid during fill; support startup and load testing; and maintain the tanks, fluid and plant afterward, with emergency response when something goes wrong. Request a quote for an immersion hall, a retrofit or a maintenance program.
Standards referenced: OCP Immersion Requirements · OCP Base Specification for Immersion Fluids · OCP Design Guidelines for Immersion-Cooled IT Equipment · ASHRAE TC 9.9 Thermal Guidelines · ASME B31.3 · ASME B31.9 · NFPA 30 · ASTM D93 · ASTM D1500 · IEC 60156
Frequently asked questions
What are the practical immersion cooling disadvantages, and how are they managed?
The main ones are weight, servicing time and fluid logistics. Floors need a structural check, technicians lift servers out of the fluid and let them drain before working on them, and the fluid has to be stored, filtered, tested and topped off. Good layout, a drain-down station, spare filters and a sampling routine turn each of those into a routine task.
How does immersion cooling reliability compare with air-cooled servers?
OCP's hardware guidance credits immersion with fewer solder joint failures, less corrosion of contacts, no dust or particulate exposure and steadier component temperatures, all of which favor long equipment life. Reliability then depends on keeping the fluid within specification, so sampling and filtration matter as much as the hardware.
What sets immersion cooling fluid price on a project?
The chemistry and the volume. Fluorinated two-phase fluids cost far more per liter than hydrocarbons or esters, and the tank design decides how many liters each kilowatt needs. Owners also budget for a spare inventory to cover top-off and for disposal at the end of the fluid's life.
Immersion cooling vs air cooling: what changes in the mechanical plant?
Most of the heat moves from room air to water, so air handlers shrink to what people and air-cooled equipment need, and the facility water plant grows. Because tanks accept warm water, the plant can lean on dry coolers or towers and run chillers far less than an air-cooled hall would.
Can existing servers be converted for an immersion tank?
Many can, after preparation. Fans come out and fan alarms are disabled in firmware, vented hard drives are replaced with sealed helium drives or solid-state storage, optical links are swapped for sealed or direct-attach types, and labels, cables and thermal paste are checked against the fluid. The server maker's warranty position should be confirmed first.
How warm can facility water be for a single-phase immersion tank?
Warmer than most other IT loads accept. Published testing has shown reliable single-phase operation on 40 degrees Celsius facility supply, and many tank makers rate their exchangers for the W32 to W45 ASHRAE classes. The guaranteed figure for a given tank, fluid and server mix comes from the manufacturer.
Does every immersion tank need a coolant distribution unit?
Each tank needs some liquid-to-water exchanger, and the question is where it sits. Some tanks carry an integrated exchanger, pumps and filters on the skid, others share an external CDU across several tanks, and two-phase tanks condense vapor on a water coil inside the lid. Each arrangement ends at a facility water connection.
How heavy is a filled immersion tank on the floor?
Heavy enough that a structural engineer should check every location. Manufacturers publish empty, filled and fully loaded static loads in kilonewtons per square meter. Hydrocarbon fluids weigh a little less than water, while fluorinated fluids weigh roughly twice as much as hydrocarbons, so the same tank volume can load the floor very differently.
How is dielectric fluid transferred into a new tank?
From sealed drums or totes through a dedicated transfer pump and filter, using clean hoses that have never carried another fluid. The tank is inspected clean and dry first, batch numbers and volumes are logged, and a baseline sample is taken once the fluid has circulated.
How often should immersion fluid be sampled?
At the interval the fluid supplier sets, which commonly means a baseline at fill, a check in the first months of operation, then routine sampling once or twice a year. An extra sample is taken after any electrical fault, overheating event or unexplained level change.
Which test results show that immersion fluid is degrading?
A drop in dielectric strength toward 6 kV per millimeter, a total acid number above twice the new-fluid limit, rising water content, a color change of more than two units on the ASTM D1500 scale, or a falling flash point. Any of these calls for a full analysis and a decision on filtering, treating or replacing the fluid.
What should happen if dielectric fluid spills?
Containment under the tank should hold the whole volume, and absorbent should be on hand for the largest tank on site. Stop the source, keep fluid out of floor drains, clean up with absorbent, dispose of it through a proper waste route, and record the volume lost so the tank can be topped off and the cause fixed.
Do immersion cooling rooms need special ventilation?
OCP calls for open-bath tanks to sit in a well-ventilated room as directed by the fluid's safety data sheet. Two-phase systems need closer attention, because vapor escapes when lids open, so sealed lids, condensers and vapor recovery limit losses and room exposure.
Why do two-phase tanks lose fluid, and how is the loss limited?
The fluid boils at around 50 degrees Celsius, so any vapor that leaves the tank is lost inventory. Sealed lids, a freeboard zone above the liquid, a condenser coil supplied with cool enough water, and procedures that let vapor settle before a lid opens keep losses low.
Are PFAS rules a concern for immersion fluids?
They are for many two-phase fluids, which are fluorochemicals and fall within the broad PFAS definition. The OCP fluid specification flags restrictions under development in the EU and in states such as Maine, so owners consider future supply and disposal before committing to a large fluorinated inventory. Hydrocarbon and ester fluids fall outside that definition.
What pressure test applies to the facility water piping serving immersion tanks?
The site piping is hydrostatically tested under its governing code, usually ASME B31.3 or B31.9, before the tank exchangers are connected or while they are isolated. OCP asks tank interfaces to tolerate up to 1,000 kPa gauge, but the test pressure is always held within the exchanger's own rating.
Can immersion tanks share a facility water loop with cold plate racks?
Yes, when the temperatures line up. Tanks and CDUs for cold plates are served from the same headers on separate valved branches, with balancing and controls set so the most demanding load gets the supply temperature it needs.
Can the heat from immersion tanks be reused?
Yes. Warm return water from single-phase tanks can feed building heating or a district heating loop through a heat exchanger, with dry coolers or towers kept available for times when the heating demand falls away.
What belongs in the turnover package for an immersion installation?
Placement and grounding checks, piping weld or joint records, flush logs, hydrostatic test records, water treatment baselines, fluid batch numbers and fill volumes, the baseline fluid analysis, safety data sheets, alarm setpoints, valve tags and as-built drawings.
How long do immersion fluids last in service?
The OCP fluid specification asks fluids to stay within specification for a decade under normal operating conditions. Reaching that depends on keeping the fluid dry, filtered and away from overheating, with sampling to catch problems early.
Related: Data Centers · Data Center Cooling Systems · Data Center Liquid Cooling · Data Center Commissioning · Data Center Mechanical Contractors · Dry Cooler · Request a quote