Your cart is empty
Data Centers — Cooling & Water Infrastructure
Data Center
Cooling & Water
Infrastructure
Pumping and mechanical flow solutions for mission-critical data center cooling, heat rejection, liquid-cooling infrastructure, water distribution, and fire protection.
E4 Industrial helps contractors, engineers, OEMs, and data center operators select and source pumping equipment for critical cooling and water systems.
How Data Center Cooling Works
Data center cooling is fundamentally about moving heat. Servers, storage, networking equipment, and other electrical systems generate heat during operation. The cooling infrastructure must capture that heat, transport it away from the IT equipment, and ultimately reject it outside the facility.
There is no single cooling architecture used by every data center. The design depends on factors such as IT load, rack density, climate, available water, facility configuration, redundancy requirements, and the type of computing equipment being deployed. Modern facilities may use air cooling, chilled-water cooling, direct liquid cooling, or a combination of these approaches.
Regardless of the architecture, the basic objective is the same:
In a chilled-water data center, Computer Room Air Handlers (CRAHs) circulate warm data-hall air across cooling coils supplied with chilled water. The chilled water absorbs heat from the air and carries that heat back toward the cooling plant. Chilled-water pumps maintain circulation between the cooling plant and the cooling loads.
Where water-cooled chillers are used, a separate condenser-water system transfers heat from the chiller to cooling towers for rejection to the outdoor environment. This typically introduces another major pumping service: condenser-water pumps. Air-cooled chiller systems use a different heat-rejection arrangement and do not require the conventional cooling-tower/condenser-water loop.
Direct-to-chip liquid cooling captures heat close to high-power processors using liquid-cooled cold plates. The heated coolant circulates through the Technology Cooling System (TCS) to a Coolant Distribution Unit (CDU). In a common liquid-to-liquid arrangement, the CDU transfers heat through a heat exchanger into the facility-side cooling system while keeping the technology and facility cooling circuits hydraulically separate.
Liquid cooling does not necessarily eliminate air cooling throughout the data center. Facilities can use hybrid cooling architectures, with liquid cooling serving high-heat-load processors while air-based systems continue to cool other IT equipment and components.
A data center may contain several separate water or coolant circuits, each performing a different function.
| Circuit | Function |
|---|---|
| Chilled Water (CHW) | Carries heat between cooling loads such as CRAHs and the central cooling plant. |
| Condenser Water (CW) | In water-cooled chiller systems, carries heat between the chiller condenser and the cooling towers. |
| Facility Water System (FWS) | Provides facility-side cooling water serving liquid-cooling infrastructure. |
| Technology Cooling System (TCS) | Circulates controlled coolant between liquid-cooled IT equipment and the CDU or other technology-side cooling equipment. |
These circuits should not be viewed as one continuous water system. They can operate at different temperatures, pressures, flow rates, water-quality requirements, and equipment boundaries depending on the cooling architecture.
Whenever water or coolant is used to transport heat, flow must be created and maintained. Depending on the facility design, data center pumping applications may include:
Pump selection therefore becomes part of the larger cooling-system design, with flow, pressure, efficiency, redundancy, maintainability, and operating conditions influencing the equipment selected for each service.
Whether heat is first captured by air or directly by liquid, the objective remains the same: reliably move heat away from mission-critical IT equipment and ultimately reject it from the facility. Understanding that cooling path is the first step in selecting the pumping and mechanical infrastructure required to support it.
Where E4 Fits
Data center cooling depends on reliable circulation throughout the facility. From the central cooling plant and heat-rejection system to facility distribution and the facility side of liquid cooling, pumps move the water that carries heat through the system. E4 Industrial focuses on the pumping equipment that supports these critical cooling and water systems.
The central cooling plant provides cooling capacity for the facility and can contain multiple pumping systems depending on its design. E4 can support pumping applications including:
- Chilled-water pumps
- Primary chilled-water pumps
- Secondary / distribution chilled-water pumps
- Condenser-water pumps
- Cooling-plant circulation pumps
In a chilled-water system, pumps circulate water between the cooling plant and cooling loads throughout the facility. In water-cooled chiller plants, a separate condenser-water pumping system circulates water between the chiller and the heat-rejection system.
Ultimately, the heat removed from the IT equipment must be rejected outside the facility. Depending on the cooling architecture, heat rejection may use equipment such as cooling towers, dry coolers, fluid coolers, air-cooled chillers, or combinations of these technologies. Where circulating water or coolant is used, pumping equipment provides the flow required to move heat from the cooling system to the heat-rejection equipment.
E4 can support pumping applications associated with:
Cooling capacity must be distributed from the cooling source to the equipment that requires it. Large data centers can use extensive hydronic distribution systems connecting central plants, mechanical equipment, data halls, cooling equipment, and other facility loads. Depending on the system configuration, pumping may be arranged as:
- Primary pumping
- Primary-secondary pumping
- Variable-primary pumping
- Dedicated equipment pumping
- Distributed pumping
E4 helps customers select pumping equipment around the required flow, head, system configuration, operating conditions, and project requirements.
Direct liquid cooling changes where heat is captured, but it does not eliminate the need to move that heat through the facility. In common liquid-to-liquid cooling architectures, the Facility Water System (FWS) supplies cooling water to a Coolant Distribution Unit (CDU). The CDU transfers heat from the technology-side coolant circuit into the facility-side cooling circuit.
The pumps contained inside a CDU or serving the Technology Cooling System are typically part of the specialized liquid-cooling equipment package. E4's core focus is the larger facility infrastructure that delivers and circulates cooling water to and from that equipment.
Cooling represents the largest pumping opportunity, but data centers also require supporting water systems throughout the facility. Depending on the project, E4 can support pumping applications for:
These systems are covered in Section 8 — Fire Protection & Utility Water.
Data center pumping equipment must match the hydraulic and operating requirements of the system it serves. E4 Industrial works with contractors, engineers, OEMs, and facility operators to help evaluate and source pumping equipment based on project requirements, including:
- Required flow
- Total dynamic head
- Pump configuration
- Fluid and operating temperature
- Materials of construction
- Motor and electrical requirements
- Efficiency requirements
- Redundancy and equipment arrangement
- Installation configuration
- Project specifications and documentation
Our role is straightforward: understand the application, select the appropriate pumping solution, and support the equipment from specification through procurement.
Whether the system uses conventional chilled water, water-cooled heat rejection, or modern liquid-cooling infrastructure, reliable heat transport depends on maintaining the required flow throughout the system. E4 Industrial provides the pumping expertise and equipment needed to keep that infrastructure moving.
Chilled Water Systems
Chilled-water systems are widely used to remove heat from large, mission-critical facilities, including data centers. In these systems, water serves as the medium for transporting heat between cooling equipment in the data center and the central cooling plant.
The chiller produces chilled water and sends it through the facility distribution system. Pumps provide the flow and pressure needed to circulate that water between the cooling plant and the connected cooling loads.
At the cooling load, heat transfers into the chilled water. In an air-cooled data hall using Computer Room Air Handlers (CRAHs), fans move warm data-hall air across chilled-water coils. The water absorbs heat from the air and returns warmer to the central cooling plant.
Inside the chiller, heat is removed from the returning chilled water so that the water can be supplied back to the facility and the cycle can continue. How the chiller ultimately rejects that heat depends on whether the plant uses water-cooled, air-cooled, or another heat-rejection configuration.
Two terms appear constantly on data center mechanical drawings and pump schedules:
The temperature difference between the supply and return water represents the useful temperature rise being achieved as the water passes through the cooling loads. Flow rate and temperature difference together are fundamental to how much heat a chilled-water system can transport.
A pump is not selected simply by pipe size or nominal system capacity. The required flow and hydraulic resistance of the system determine the pumping duty.
Large chilled-water systems can be configured in several ways. Two common arrangements are primary-secondary pumping and variable-primary-flow pumping. The appropriate arrangement depends on the plant design, chillers, distribution system, controls, operating requirements, and redundancy strategy.
A primary-secondary system separates the chiller circulation loop from the facility distribution loop. The primary pumps circulate water through the chillers, while the secondary pumps move chilled water through the distribution system serving the cooling loads. A hydraulically decoupled connection allows the two circuits to operate with different flow requirements.
In a Variable Primary Flow (VPF) system, a common pumping system circulates chilled water through both the chillers and the distribution system, with flow changing as system demand changes. Variable-speed pumping is commonly used to adjust flow while maintaining the operating requirements of the chillers and distribution system.
Neither configuration is universally preferred. Chilled-water plant design requires consideration of system hydraulics, equipment limitations, controls, redundancy, operating range, and energy performance.
Data center cooling demand is not necessarily constant across every operating condition. Modern chilled-water systems can use variable-speed pumps and control strategies to adjust water flow as cooling demand changes.
This makes pump performance important across the expected operating range, not only at a single maximum design point. For mission-critical applications, the pumping system must also work as part of the facility's overall equipment staging and redundancy strategy.
A chilled-water pump is part of a larger thermal-management system. Loss of required water circulation can reduce the ability of connected cooling equipment to remove heat from the data hall.
For this reason, data center chilled-water systems are designed around project-specific requirements for capacity, redundancy, controls, maintainability, and continuity of cooling.
E4 Industrial supports pumping equipment for chilled-water circulation throughout the cooling infrastructure, including applications such as:
Pump selection begins with the system requirements: flow, total dynamic head, fluid conditions, operating range, installation configuration, efficiency requirements, redundancy, and project specifications. E4 works with customers to translate those requirements into an appropriate pumping solution.
The chiller creates the cooling capacity. The chilled-water system distributes it. Reliable chilled-water pumping is what allows that cooling capacity to reach the CRAHs, cooling coils, heat exchangers, and other loads that depend on it throughout the facility.
Condenser Water & Heat Rejection
Removing heat from the data hall is only part of the cooling process. That heat must ultimately be rejected outside the facility. The method used depends on the cooling architecture. Data centers may use cooling towers, dry coolers, fluid coolers, air-cooled chillers, or combinations of heat-rejection technologies. A condenser-water system is specifically associated with water-cooled chillers and certain other water-cooled equipment; it should not be assumed to exist in every data center.
In a water-cooled chilled-water plant, the chiller separates two different water circuits:
Inside the chiller, refrigerant absorbs heat from the chilled-water circuit at the evaporator. The refrigeration cycle moves that heat to the condenser, where it is transferred into the condenser water. The condenser-water system then transports the heat away from the chiller.
Condenser-water pumps circulate water between the water-cooled chiller and the heat-rejection system.
The cooling tower reduces the temperature of the returning condenser water before it is circulated back through the chiller. The process operates continuously while the associated cooling equipment is in service.
CHW pumps move cooling water between the chiller plant and facility cooling loads. Condenser-water pumps move heat from the chiller toward the heat-rejection equipment. These are separate services with separate hydraulic requirements.
Cooling towers are widely used for heat rejection in water-cooled central plants. In an evaporative cooling tower, warm condenser water is brought into contact with moving air. A portion of the water evaporates, removing heat from the remaining recirculating water. The cooled water collects and returns to the condenser-water system for reuse.
Because evaporation is part of the heat-rejection process, cooling-tower systems require make-up water to replace water lost through evaporation, blowdown, drift, and other losses. Water treatment and blowdown management are also required to control the concentration of dissolved solids in the recirculating water.
For data centers, the heat-rejection system therefore affects more than cooling capacity. It can also influence:
The appropriate heat-rejection strategy is project-specific and involves balancing thermal performance, energy use, water use, site conditions, and operating requirements.
Not every data center uses an evaporative cooling tower. Other heat-rejection technologies include:
Liquid cooling does not eliminate heat rejection. It changes how heat is collected and the temperatures at which it can be transported. After heat is transferred from the IT equipment through the Technology Cooling System and into the Facility Water System, that heat must still be removed from the facility.
Depending on the design, the facility-side system may ultimately reject heat through chillers, cooling towers, dry coolers, adiabatic or evaporative fluid coolers, or other heat-rejection or heat-recovery systems.
Higher facility-water temperatures associated with some liquid-cooling architectures can expand the use of dry cooling and other forms of compressor-reduced or compressor-free heat rejection when climate and operating conditions permit. This is one reason the transition toward high-density liquid cooling can change the design of the entire facility cooling plant, not just the equipment located inside the data hall.
E4 Industrial supports the pumping equipment used to circulate water and heat-transfer fluids through condenser-water and facility heat-rejection systems. Applications can include:
- Condenser-water circulation
- Cooling-tower circulation
- Central cooling-plant pumping
- Closed-loop heat-rejection circulation
- Dry-cooler and fluid-cooler circulation
- Water and water-glycol systems
- Cooling-system make-up water
Pump selection is based on the hydraulic and operating requirements of the system, including required flow, total dynamic head, fluid properties, operating temperature, open-loop or closed-loop configuration, equipment elevation and suction conditions, operating range, efficiency requirements, redundancy requirements, installation configuration, and project specifications.
Whether a data center uses traditional chilled water, direct liquid cooling, evaporative heat rejection, or dry cooling, the thermal path is not complete until the captured heat is transferred out of the facility. The condenser-water and heat-rejection systems provide that final link between the cooling infrastructure and the outdoor environment.
Liquid Cooling Infrastructure
The rapid growth of AI and high-performance computing is increasing heat density inside data centers and changing how cooling systems are designed. For many high-density AI and HPC applications, direct-to-chip liquid cooling has become a primary thermal-management approach, moving heat away from processors through liquid rather than relying entirely on air.
For E4, the primary opportunity is the facility-side pumping infrastructure that delivers cooling water to liquid-cooling equipment and carries the transferred heat back toward the cooling plant or heat-rejection system.
In Direct-to-Chip (DTC) cooling, liquid-cooled cold plates are mounted directly to major heat-generating components such as CPUs and GPUs. Heat transfers from the electronic component into the cold plate and then into circulating coolant. This allows a substantial portion of the heat to be captured close to its source instead of first transferring that heat into the surrounding data-hall air.
The technology-side coolant and the facility-side water are commonly separate cooling circuits rather than one continuous water loop.
The Technology Cooling System (TCS) is the cooling circuit closest to the IT equipment. In a typical direct-to-chip arrangement, coolant circulates between the CDU and the liquid-cooled IT equipment:
The TCS operates under coolant-quality, temperature, pressure, flow, cleanliness, and material-compatibility requirements established for the IT cooling equipment. This technology-side circuit is generally part of the specialized liquid-cooling infrastructure associated with the servers, racks, and CDU.
The Coolant Distribution Unit (CDU) is one of the key pieces of equipment in a liquid-cooled data center. In a common liquid-to-liquid CDU arrangement, the CDU creates and controls the secondary technology cooling loop while transferring heat into the facility-side water system through a heat exchanger.
A CDU can incorporate equipment such as:
The CDU therefore performs more than simple circulation. It helps provide the controlled coolant conditions required by the technology-side cooling system while maintaining separation from the facility water circuit.
The heat crosses the CDU through heat exchange. The two water or coolant circuits do not need to physically mix for heat to move from the technology side to the facility side.
The Facility Water System (FWS) connects the data center's mechanical cooling infrastructure to the liquid-cooling equipment. In liquid-to-liquid CDU systems, facility water enters the CDU heat exchanger, absorbs heat transferred from the technology-side coolant, and then returns toward the facility cooling or heat-rejection system.
The exact facility architecture varies. Depending on system temperatures, climate, and project design, the facility side may connect to chillers, dry coolers, cooling towers, fluid coolers, heat exchangers, or combinations of heat-rejection equipment.
One important development in high-temperature liquid cooling is the ability to operate some technology cooling systems with warmer facility water. Higher allowable cooling-water temperatures can increase the opportunity to reject heat directly to outdoor air through equipment such as dry coolers when climate and operating conditions permit. In these designs, mechanical chillers may operate less frequently or, in suitable applications and climates, may not be required for portions of the cooling system.
This can simplify portions of the cooling architecture, but chiller-less operation is not universal. The appropriate design depends on IT equipment requirements, facility-water temperatures, climate, heat-rejection equipment, redundancy, and overall system design.
Liquid cooling is not simply equipment added inside the server rack. Moving cooling liquid closer to the IT equipment introduces additional requirements throughout the mechanical infrastructure, including:
This means the transition to liquid-cooled computing can affect the entire cooling path from the IT equipment to the outdoor heat-rejection system.
E4's primary liquid-cooling focus remains on facility-side cooling and pumping infrastructure, rather than the specialized technology-side hardware used inside direct-to-chip or immersion systems.
E4 Industrial supports the facility-side pumping infrastructure required to move cooling water between the heat-rejection or cooling plant and liquid-cooling equipment. Potential pumping applications include:
Pump selection must account for the complete hydraulic system, including required flow, total dynamic head, supply and return temperatures, fluid properties, operating range, system pressure, equipment pressure limitations, variable-flow requirements, redundancy, materials of construction, efficiency requirements, installation configuration, and project specifications.
Liquid cooling brings the mechanical cooling system closer than ever to the computing equipment. The CDU may form the interface, but the heat still has to travel from the data hall through the facility and ultimately outside. E4 Industrial supports the pumping infrastructure that keeps the facility side of that cooling path moving.
Pumping Solutions
Pumps are the mechanical drivers that keep water and coolant moving through data center cooling infrastructure. They serve different functions throughout the facility, from large central chilled-water and condenser-water systems to facility distribution, heat rejection, and dedicated cooling loops.
The appropriate pump depends on the hydraulic duty, system configuration, fluid, operating range, installation, redundancy strategy, efficiency requirements, and maintainability requirements of the application.
Horizontal split-case centrifugal pumps are well suited to high-flow water circulation services and are widely applied in large HVAC and cooling systems. In data centers, split-case pumps can be applied to primary chilled-water circulation, secondary chilled-water circulation, condenser-water circulation, large facility water loops, and central cooling-plant circulation.
Their double-suction and between-bearings configurations can provide efficient hydraulic performance for large water duties, while the horizontally split casing allows access to internal components without disturbing major system piping. Split-case pumps are used for both chilled-water and condenser-water service in large data center cooling systems.
Where mechanical-room footprint is limited, vertically arranged pumps can reduce the amount of floor space required by the pumping equipment. Vertical split-case pumps retain many of the hydraulic characteristics of a split-case design while arranging the pump vertically. Vertical in-line pumps place the suction and discharge connections along the piping centerline, allowing the pump to be installed directly within the piping arrangement.
These configurations can be used in chilled-water, condenser-water, and other hydronic cooling applications where the required duty falls within the pump's operating range. Vertical split-case pumps have been used for both primary chilled-water and condenser-water duties in hyperscale data centers, and vertical in-line configurations are well established within commercial chilled-water and HVAC pumping systems.
End-suction centrifugal pumps provide a compact and versatile solution for a broad range of water and cooling applications. They may be applied to chilled-water circulation, cooling-water circulation, dedicated equipment loops, heat-exchanger circulation, fluid-cooler and dry-cooler loops, smaller or distributed cooling systems, and OEM and packaged cooling equipment.
End-suction pumps are available in several arrangements, including close-coupled, frame-mounted, and split-coupled designs.
Vertical turbine pumps are used when the hydraulic source or installation favors a vertical wet-pit, basin, or below-grade pumping arrangement. Potential data center applications can include cooling-tower basin service, water intake, raw-water supply, make-up water, facility water supply, and other high-flow water-transfer applications.
Unlike a conventional horizontal centrifugal pump supplied from a pressurized suction header, the bowl assembly of a vertical turbine pump can be submerged in the water source while the driver remains above the operating water level.
Liquid-cooled data centers create additional pumping duties on both sides of the CDU. On the facility side, pumps circulate water between cooling / heat-rejection equipment, facility distribution, and CDUs. These pumps may range from larger central circulation pumps to smaller dedicated pumps depending on how the facility water system is designed.
On the technology side, pumps integrated into CDUs and other liquid-cooling equipment circulate coolant through manifolds, cold plates, and related technology cooling equipment. These pumps are typically smaller and form part of the specialized CDU or OEM cooling package.
Pump manufacturers now offer equipment specifically for both facility-side data-center cooling and OEM liquid-cooling applications, including pumps incorporated into CDU systems. E4's primary focus is the facility-side pumping infrastructure supporting these systems.
Data center cooling systems frequently operate over changing loads rather than at one fixed operating point. Variable Frequency Drives (VFDs) allow pump speed to change as system demand changes. In variable-flow hydronic systems, this can reduce unnecessary pumping energy while allowing the system to maintain the required operating conditions.
For pump selection, this means the design should consider the expected operating range, not simply the maximum flow and head condition.
Large cooling systems commonly divide the required flow among multiple pumps rather than relying on a single unit. Parallel pumping can provide additional system capacity, better matching of pump operation to changing loads, equipment staging, maintenance flexibility, and redundancy when required by the facility design.
When pumps operate in parallel, their combined performance must be evaluated against the system curve; adding another pump does not automatically double system flow. Data center redundancy requirements are project-specific — pump quantities and arrangements are configured around the facility's selected reliability strategy rather than a universal standard.
Selecting a data center pump requires more than identifying a required GPM. E4 evaluates the complete pumping application:
Flow and head establish the fundamental hydraulic duty, but reliable pump selection also requires evaluation of efficiency, NPSH, the expected operating region, fluid characteristics, installation conditions, and appropriate NPSH margin.
Different pump designs solve different hydraulic and installation problems. A large central chilled-water loop may favor a split-case pump. A compact mechanical installation may favor a vertical or end-suction configuration. A cooling-tower basin or below-grade water source may favor a vertical turbine. A dedicated liquid-cooling loop may require a completely different pump size and configuration.
The objective is not to force every application into one pump type. E4 Industrial combines application review, pump selection, and manufacturer resources to help customers identify the appropriate pumping solution for each data center cooling and water application.
Fire Protection & Utility Water
Data center pumping requirements extend beyond the primary cooling systems. Facilities can also require dedicated pumping for fire protection, cooling-system make-up water, domestic and service water, drainage, and wastewater handling. These systems perform different functions and should be treated as separate applications with their own hydraulic, operational, and regulatory requirements.
Fire protection is a dedicated life-safety function and is separate from the normal HVAC and cooling-water systems. A fire pump provides the required water flow and pressure to a water-based fire protection system when the available water supply cannot provide the required hydraulic performance on its own.
Fire pump selection and installation are governed by the project's fire-protection design and applicable codes and standards. In the United States, NFPA 20 — Standard for the Installation of Stationary Pumps for Fire Protection establishes requirements for the selection and installation of stationary fire pumps. FM Global also publishes detailed requirements for fire-protection pump selection, installation, water supply, drivers, controls, and testing. The fire-protection engineer, project specifications, applicable codes, listings/approvals, insurer requirements, and the Authority Having Jurisdiction (AHJ) govern the final configuration.
A fire-protection system may also include a jockey pump, sometimes called a pressure-maintenance pump. The jockey pump is not intended to provide the main fire-flow demand. Its purpose is to maintain system pressure and compensate for small pressure losses so that minor pressure fluctuations do not unnecessarily start the main fire pump.
Fire-pump installations involve more than the pump itself. Depending on the project, a fire-pump package can incorporate:
Fire pump packages may use electric-motor or diesel-engine drivers depending on the project design and applicable requirements. Complete fire-pump packages can also be provided as factory-engineered systems.
Evaporative cooling systems lose water during normal operation and require make-up water to replace those losses. In cooling-tower systems, make-up water primarily replaces water lost through evaporation, blowdown, drift, and other system losses.
Maintaining the required water level allows the cooling-tower and condenser-water systems to continue operating as designed. Water quality and treatment also matter because evaporation concentrates dissolved minerals within the recirculating water.
Data-center water systems can also incorporate filtration, treatment, recovery, or alternative water sources depending on the facility's water-management strategy.
Data centers are also commercial facilities and require water for occupied and support areas. Where the available incoming water pressure is insufficient for the building's distribution requirements, a pressure-booster system can maintain the required pressure throughout the facility — for domestic water distribution, restrooms and occupied spaces, facility service water, maintenance water, and other building water requirements.
Pressure-booster systems typically use one or more pumps with controls that maintain water pressure as demand changes. Multi-pump packaged booster systems are an established solution for facility water distribution in data centers.
Water entering a below-grade space or collecting in equipment areas must also be removed safely. Depending on the facility design, pumping applications may include mechanical-room drainage, equipment drainage, sump pits, groundwater infiltration, wastewater collection, sewage lift applications, and emergency drainage.
These services commonly use submersible drainage or wastewater pumps, selected according to the liquid being handled, solids requirements, required flow, discharge head, sump arrangement, and control requirements.
Each application requires its own equipment selection criteria. Fire pumps must follow the approved fire-protection design and applicable codes and standards. Make-up and booster pumps must satisfy the required flow and pressure conditions. Drainage and wastewater pumps must be selected around the characteristics of the liquid, solids handling requirements, sump conditions, and required discharge head.
Cooling-water circulation may represent the most visible pumping requirement in a data center, but reliable operation depends on supporting systems throughout the facility. From fire protection and cooling-system make-up water to pressure boosting and drainage, E4 Industrial can help customers select pumping equipment for the broader water infrastructure supporting mission-critical facilities.
Mechanical Flow Infrastructure
A pump does not operate independently. It is part of a complete hydraulic network that must deliver the required flow from the cooling source to the equipment being served and return the fluid through the system. In data center cooling systems, that mechanical flow infrastructure can include pumps, supply and return piping, valves, heat exchangers, strainers and filtration, expansion and air-management equipment, instrumentation, and controls.
The piping network connects the cooling source, pumps, heat exchangers, cooling equipment, CDUs, and other loads throughout the facility. Its design directly affects the pumping requirement.
Every length of pipe and every piece of equipment in the flow path creates hydraulic resistance. Pipe size, flow rate, fittings, valves, heat exchangers, coils, and other components collectively determine the pressure loss the pump must overcome. This is why a pump is selected against the system, not simply against a required flow rate.
Valves perform several different functions within cooling-water infrastructure. Depending on the system design, they may be used to:
- Isolate equipment for maintenance
- Prevent reverse flow
- Control or modulate flow
- Balance branches of a distribution system
- Bypass equipment when required by the system design
- Support redundant flow paths
- Allow equipment to be staged on or off
Isolation valves, control valves, balancing devices, and redundant distribution paths can all be part of the cooling-water infrastructure, depending on the system design. The valves and piping arrangement therefore affect both how the system operates and the hydraulic duty seen by the pumps.
This is one reason accurate pump selection requires a complete understanding of the system rather than simply selecting a pump from GPM alone.
Clean fluid is important throughout cooling-water infrastructure and becomes especially critical as flow passages become smaller in liquid-cooled systems. Strainers and filtration can help remove debris and contaminants before they reach sensitive system components. Commercial hydronic equipment commonly incorporates strainers ahead of critical equipment, while modern liquid-cooling systems use filtration to protect heat exchangers, cooling passages, valves, and other components from contamination.
Liquid-cooling systems can place particularly stringent requirements on fluid quality because contaminants can restrict small cooling passages and impair heat-transfer performance. The appropriate filtration and fluid-management strategy depends on the equipment manufacturer's requirements and system design.
Closed hydronic systems must also accommodate changes in fluid volume as temperature changes. Expansion equipment provides a controlled means of accommodating those changes while helping establish and maintain appropriate system pressure.
Air and other gases must also be managed. Entrained or trapped air can interfere with circulation, reduce pumping performance, create noise, and affect hydraulic stability. Closed hydronic systems therefore require appropriate expansion and air-management provisions as part of the overall system design.
Mission-critical cooling systems require operators to understand what is happening throughout the hydraulic network. Common parameters that may be monitored include supply and return temperature, system pressure, differential pressure, flow, pump operating status, equipment operating status, fluid conditions, and leak detection in liquid-cooling systems.
These measurements also support pump control. For example, differential-pressure or flow signals can be used with variable-speed pumping to adjust circulation as system demand changes.
The pump sits at the center of a much larger hydraulic relationship. Changes anywhere in this network — pipe routing, pipe sizing, valve position, equipment selection, heat-exchanger pressure drop, system flow, parallel pump operation — can change the hydraulic conditions under which the pump operates. For that reason, E4 approaches pump selection as a system application, not an isolated equipment selection.
E4 Industrial's current focus is the pumping equipment at the heart of the mechanical flow system. Our application approach considers the infrastructure surrounding the pump so that equipment selection reflects the actual hydraulic requirements of the system. That includes understanding where the pump sits in the cooling architecture, what equipment it serves, required system flow, total dynamic head, expected operating range, suction conditions, fluid characteristics, parallel and redundant operation, installation constraints, equipment interfaces, and project specifications.
A pump creates flow, but the entire hydraulic system determines where the pump operates. Pipe, equipment, valves, fittings, strainers, and heat exchangers create resistance to flow. The combined resistance of the system determines the head the pump must produce at the required flow rate. Required Flow + System Resistance = Pumping Duty. This is why accurate pump selection starts with the system rather than the pump — and why reliable data center cooling depends on the entire mechanical flow path working together.
Data Center Terminology
Data centers bring together IT, electrical, mechanical, cooling, and water infrastructure. Understanding a few common terms makes it much easier to understand how these systems work together. The following glossary covers terminology used throughout this page and commonly encountered in data center cooling and mechanical infrastructure.
Data Hall / White Space
The data hall, also commonly called the white space, is the area of the data center that contains the IT equipment, including servers, storage, networking equipment, and equipment racks.
Rack
A rack or cabinet is the physical enclosure used to organize and support servers, networking equipment, power-distribution equipment, and other IT hardware. Racks provide a standardized structure for installing equipment while supporting power distribution, cabling, airflow, and increasingly liquid-cooling connections.
Rack Density
Rack density describes the amount of electrical power consumed by IT equipment installed in a rack, commonly expressed in kilowatts per rack (kW/rack). Because most of the electrical energy consumed by the IT equipment ultimately becomes heat that must be managed, increasing rack power density places greater demands on the cooling infrastructure. Rack power density is therefore an important input for both power and cooling-system design.
HPC — High-Performance Computing
High-Performance Computing (HPC) refers to computing systems designed to perform computationally intensive workloads using large amounts of processing capacity. HPC environments have historically been important users of advanced cooling technologies, including warm-water and direct liquid cooling. Many of the cooling approaches now being adopted for AI infrastructure evolved from technologies already used in HPC facilities.
Colocation / Colo
A colocation data center is operated by a third-party provider that supplies data center space, power, cooling, connectivity, and related infrastructure to multiple customers. Customers install or operate their IT equipment within the provider's facility rather than constructing and operating the entire data center themselves.
Hyperscale
Hyperscale generally describes very large computing infrastructure designed to support cloud, internet-scale, AI, and other large distributed computing workloads. Hyperscale facilities are associated with very large IT deployments and highly scalable power and cooling infrastructure. There is no single size threshold that should be assumed from the term alone.
CRAH — Computer Room Air Handler
A Computer Room Air Handler (CRAH) circulates data-hall air through a cooling coil supplied by chilled water. Warm air from the data hall passes across the coil, heat transfers into the chilled water, and cooler air is returned to the IT environment. CRAHs therefore connect the air side of the data hall with the chilled-water side of the mechanical cooling system.
CRAC — Computer Room Air Conditioner
A Computer Room Air Conditioner (CRAC) is precision cooling equipment that uses a refrigeration system to remove heat from the data-center air.
For the purposes of this page: CRAH = chilled-water cooling unit; CRAC = refrigerant-based precision cooling unit. Both are established data-center air-cooling technologies, although their cooling architectures differ.
CHW — Chilled Water
Chilled water is the water circulated between a cooling source, such as a chiller, and cooling loads throughout the facility. The water absorbs heat from cooling coils, heat exchangers, or other equipment and returns to the cooling plant to be cooled again.
CHWS — Chilled-Water Supply
Chilled-Water Supply (CHWS) is the chilled water leaving the cooling source and traveling toward the cooling loads.
CHWR — Chilled-Water Return
Chilled-Water Return (CHWR) is the warmer water returning from the cooling loads after absorbing heat. Together: CHWS → Cooling Load → CHWR → Cooling Plant. This supply-and-return relationship is fundamental to hydronic data-center cooling systems.
Chiller
A chiller is refrigeration equipment that removes heat from the chilled-water system. The chilled-water loop carries heat from the facility to the chiller. The chiller then transfers that heat to another medium so it can ultimately be rejected outside the facility.
Condenser Water
Condenser water is a separate water circuit used with water-cooled chillers to transport heat from the chiller condenser to the heat-rejection equipment. Conceptually: Cooling Load → Chilled Water → Chiller → Condenser Water → Cooling Tower → Outside. Condenser water and chilled water therefore perform different functions and normally operate as separate circuits.
Cooling Tower
A cooling tower is heat-rejection equipment that cools circulating water primarily through evaporative heat transfer to outdoor air. In a conventional water-cooled chiller system, the cooling tower removes heat carried by the condenser-water circuit before that water returns to the chiller.
Dry Cooler
A dry cooler rejects heat from a circulating fluid to outdoor air through an air-cooled heat exchanger. Unlike a conventional evaporative cooling tower, dry operation does not rely on evaporation for heat rejection. Higher-temperature liquid-cooling systems can increase the opportunity to use dry cooling because heat can sometimes be rejected directly to outdoor air without first producing conventional chilled water.
DTC — Direct-to-Chip Cooling
Direct-to-Chip (DTC) cooling places liquid-cooled heat-transfer devices directly on major heat-producing electronic components such as CPUs and GPUs. Instead of requiring all processor heat to transfer first into the data-hall air, a portion of the heat is captured directly into circulating coolant.
Cold Plate
A cold plate is a liquid-cooled heat-transfer device mounted directly to a heat-generating electronic component. Coolant flows through internal passages in the cold plate, absorbs heat from the processor, and carries that heat into the Technology Cooling System.
CDU — Coolant Distribution Unit
A Coolant Distribution Unit (CDU) manages the coolant circuit serving liquid-cooled IT equipment. In a common liquid-to-liquid configuration, the CDU uses a heat exchanger to transfer heat from the technology-side coolant into the facility-side water system while maintaining separation between the two circuits.
Depending on its design, a CDU may incorporate a heat exchanger, pumps, filtration, temperature control, pressure and flow monitoring, sensors, leak detection, and controls.
FWS — Facility Water System
The Facility Water System (FWS) is the facility-side cooling-water circuit that connects the mechanical cooling or heat-rejection infrastructure to the liquid-cooling equipment. In a liquid-to-liquid CDU arrangement: Facility Cooling → FWS → CDU. The FWS remains hydraulically separate from the technology coolant loop while accepting heat through the CDU heat exchanger.
TCS — Technology Cooling System
The Technology Cooling System (TCS) is the controlled coolant circuit serving the liquid-cooled IT equipment. Conceptually: CDU → TCS Distribution → Manifolds → Cold Plates → TCS Return → CDU. Its fluid quality, temperature, pressure, materials, cleanliness, and operating requirements are governed by the technology cooling equipment it serves.
RDHx — Rear-Door Heat Exchanger
A Rear-Door Heat Exchanger (RDHx) places a liquid-cooled heat exchanger at the rear of an IT rack. Hot server exhaust air passes through the rear-door heat exchanger, where heat transfers from the air into a liquid circuit. Unlike direct-to-chip cooling, the coolant does not necessarily flow through the processors themselves. Rear-door heat exchangers are a form of close-coupled cooling rather than the same architecture as direct-to-chip cooling.
Immersion Cooling
In immersion cooling, electronic equipment is immersed directly in an electrically nonconductive, or dielectric, cooling fluid. Heat transfers directly from the electronic components into the surrounding fluid. Immersion cooling is therefore a different architecture from direct-to-chip cold-plate cooling and should not be treated as simply another type of CDU-fed cold-plate system.
Flow — GPM
Flow describes the quantity of liquid moving through the system over time. For water systems in the United States, pump flow is commonly expressed in GPM — Gallons Per Minute.
Head — Feet of Head
Head represents the energy per unit weight that the pump must provide to move fluid through the system. For water pumping systems in the United States, pump head is commonly expressed in feet of head (ft). Pump selection begins by establishing the required flow and head and then evaluating the pump's performance against the complete hydraulic system.
TDH — Total Dynamic Head
Total Dynamic Head (TDH) is the total head the pump must overcome at the required operating flow. Depending on the system, this can include static elevation effects and the hydraulic resistance created by piping, fittings, valves, heat exchangers, coils, filters, and other equipment. TDH therefore describes the system requirement seen by the pump, not simply the discharge pressure of the pump.
VFD — Variable Frequency Drive
A Variable Frequency Drive (VFD) controls the speed of an AC motor by varying the electrical frequency supplied to it. On centrifugal pumps, changing motor speed allows pump output to be adjusted as system demand changes. Variable-speed pumping is commonly used in chilled-water and other variable-flow hydronic systems to better match pumping capacity with cooling demand.
N
N represents the amount of equipment capacity required to support the intended system load. If three pumps are required to meet the full design condition, for example: N = 3 pumps. This represents the required operating capacity without an additional redundant unit.
N+1
N+1 means the system has the capacity required for the load plus one additional redundant unit. Using the previous example: 3 required pumps + 1 redundant pump = N+1. N+1 arrangements are commonly used in data-center cooling infrastructure when the project requires continued capacity following the loss or maintenance of one unit.
2N
2N generally means providing a complete duplicate of the required system capacity. Conceptually: N required system + second N system = 2N. The exact implementation depends on how the equipment and distribution paths are configured.
Importantly, terms such as N+1 and 2N describe redundancy architecture; they are not themselves equivalent to an Uptime Institute Tier rating. Uptime Institute notes that Tier compliance depends on the complete topology and operating characteristics rather than simply counting redundant components.
PUE — Power Usage Effectiveness
Power Usage Effectiveness (PUE) is a widely used metric for measuring data-center infrastructure energy efficiency. It compares total data-center facility energy use with the energy consumed by the IT equipment:
PUE = Total Data Center Energy ÷ IT Equipment Energy
A PUE closer to 1.0 indicates that a larger proportion of total facility energy is reaching the IT equipment rather than supporting infrastructure such as cooling and power conversion. PUE is an infrastructure-efficiency metric; it does not measure the useful computational work produced by the IT equipment.
WUE — Water Usage Effectiveness
Water Usage Effectiveness (WUE) measures on-site data-center water consumption relative to the energy consumed by the IT equipment. It is commonly expressed as liters of water per kilowatt-hour of IT energy (L/kWh). WUE helps operators understand the relationship between facility water consumption and IT energy use. The metric is particularly relevant when comparing cooling architectures that use different amounts of evaporative cooling or other water-consuming processes.
A modern liquid-cooling path: Chip → Cold Plate → TCS → CDU → FWS → Facility Cooling → Heat Rejection. A conventional chilled-water path: Data Hall → CRAH → CHW → Chiller → Heat Rejection. Different architectures use different equipment, but both accomplish the same fundamental task: move heat away from the IT equipment and ultimately reject it from the facility.
Manufacturer Access
E4 Industrial works with established pump manufacturers to provide pumping solutions across data center cooling, heat rejection, fire protection, facility water, and specialized fluid-handling applications. Different applications require different pump technologies. Our approach is to match the manufacturer, pump configuration, and hydraulic design to the requirements of the system rather than force every application into a single product line.
For data center infrastructure, E4 provides access to broad portfolios of horizontal and vertical centrifugal pumps built for commercial, industrial, municipal, and fire-protection service — including horizontal split-case pumps, end-suction pumps, vertical turbine pumps, commercial HVAC pumping, fire pumps, jockey pumps, and packaged fire-pump systems.
Beyond conventional commercial HVAC pumping, E4 provides access to engineered pump lines for demanding industrial and infrastructure applications — including overhung centrifugal pumps, between-bearings pumps, vertical and wet-pit pumps, specialty fluid-handling pumps, and engineered pumping systems for high-flow, high-pressure, and specialized fluid-circulation duties.
Specific manufacturers and product lines are identified during the selection and quotation process, matched to the requirements of each application.
No manufacturer is the best answer for every pumping application. A central chilled-water pump, vertical turbine cooling-water pump, fire pump, facility booster system, and CDU circulation pump can have very different hydraulic, mechanical, regulatory, and installation requirements.
E4 Industrial evaluates the application first and then works within the available manufacturer portfolio to identify the appropriate solution. Our objective is not simply to provide a brand. It is to provide the right pump for the system.
Application & Technical Support
Selecting the right pump requires more than matching a model to a flow rate. The pump must operate correctly within the hydraulic system, satisfy the mechanical and electrical requirements of the project, and provide acceptable performance across the expected operating range. E4 Industrial provides application and technical support from initial equipment selection through procurement and submittal documentation. Our role is to understand the application, evaluate the pumping requirements, coordinate with the appropriate manufacturer, and help customers select equipment that fits the system.
Every pump selection begins with understanding what the equipment is expected to do. Key application information can include:
These inputs collectively determine which pump configurations and hydraulic selections are appropriate for the application.
The required flow and total dynamic head establish the fundamental pump duty, but they are only the starting point. A proper pump selection also considers where the required operating point falls on the pump performance curve. Important performance considerations can include:
The objective is not simply to find a pump whose curve crosses the required duty point. The objective is to select a pump that fits the expected system operating conditions.
The pump and the hydraulic system interact with each other. The pump curve describes the performance available from the pump. The system curve describes the head required by the piping and equipment at different flow rates. Where those two curves intersect establishes the operating point of the pump.
Changes in system resistance, pump speed, valve position, equipment operation, or the number of pumps operating in parallel can change that operating point. This is why E4 approaches pump selection as a system application rather than an isolated equipment choice.
Many data center cooling systems use variable-speed pumping and multiple pumps operating in parallel. These configurations require evaluation beyond a single full-load condition. Depending on the project, E4 can review individual pump duty, multiple-pump operation, duty / standby arrangements, parallel pump performance, variable-speed operating range, minimum and maximum expected flow, motor loading, and efficiency across the expected operating range.
Adding pumps in parallel does not automatically produce a proportional increase in system flow. The actual operating points are determined by the interaction between the combined pump performance and the system curve.
Proper suction conditions are critical to reliable centrifugal pump operation. Pump selection requires consideration of both:
For applications such as vertical turbine pumps, cooling-tower basins, tanks, and other open-water sources, water level, pump submergence, suction geometry, and installation conditions must also be considered during equipment selection.
Once the application requirements are understood, E4 works within the available manufacturer portfolio to identify suitable equipment.
Manufacturer selection tools and application resources provide detailed hydraulic information that can include pump curves, power requirements, efficiency, NPSH, construction options, dimensional information, and technical documentation. E4 uses manufacturer resources together with the project requirements to support the equipment-selection process.
Equipment selection is only part of the procurement process. Data center projects can require detailed technical documentation for engineering review, construction coordination, and equipment approval. Depending on the manufacturer and project, pump documentation can include:
E4 helps coordinate the appropriate manufacturer documentation required for the equipment being supplied.
When a project specification or pump schedule is provided, E4 can use those requirements as the basis for equipment selection. Typical information reviewed can include pump tag, service, required flow, required head, pump type, motor requirements, materials, efficiency criteria, VFD requirements, required redundancy, manufacturer requirements, testing requirements, and documentation requirements.
Where information is incomplete or conflicting, the objective is to identify the issue before equipment is ordered rather than after it reaches the jobsite. The project engineer and approved project documents remain the governing authority for the system design. E4's role is to support the application, selection, manufacturer coordination, and procurement of the pumping equipment.
E4 Industrial brings together the project requirements and manufacturer resources needed to move a pumping application from specification to procurement. Our process is straightforward:
The result is a pumping solution selected around the actual requirements of the system.
The pump is only one component of the cooling infrastructure. Understanding the system, operating conditions, and project requirements is what allows the right pump to be selected for the application. That is the role E4 Industrial is built to provide.
Request a Quote
Data center cooling systems depend on the right pumping equipment, selected around the actual hydraulic and operating requirements of the application. E4 Industrial supports contractors, engineers, OEMs, and facility operators with pump selection, manufacturer coordination, technical documentation, and equipment procurement. Whether the requirement is for a new project, expansion, replacement, or retrofit, send us the application details and we will help evaluate the appropriate pumping solution.
To begin a pump selection or quotation, provide as much of the following information as available:
- Pump schedule or equipment specification
- Required flow
- Required total dynamic head
- Fluid
- Operating temperature
- Suction conditions
- Pump type, if specified
- Motor requirements
- Materials of construction
- VFD requirements
- Redundancy requirements
- Quantity
- Required delivery date
- Project location
- Drawings or supporting documents
If some information is not yet available, E4 can help identify the additional application data needed to complete the selection.
E4 can support pumping requirements for:
Tell us what the system needs to do. E4 will help evaluate the pumping requirements and coordinate the appropriate equipment solution. A contractor may have a complete pump schedule. An owner may only know that an existing cooling-water pump needs replacement. An OEM may have a completely different set of requirements. The form works for all three — no technical field is required to start the conversation.
Prefer to talk it through first? Talk to an Application Specialist: 281.664.8000 or email sales@e4industrial.com
Cooling Infrastructure Starts With
Understanding the Application
E4 Industrial helps turn those requirements into the right pumping solution.
Request a Quote Talk to an Application Specialist- Choosing a selection results in a full page refresh.