Air Cooler Header Box

Definition and Purpose

An air cooler header box is a pressure-containing enclosure fabricated from welded plate assemblies, designed to receive, distribute, and collect the tube-side process fluid at each end of the finned-tube bundle in an air-cooled heat exchanger. Unlike shell-and-tube exchangers, where the tube-side fluid is enclosed in a cylindrical shell, air-cooled exchangers expose the outer tube surface directly to forced-convection airflow; the header box therefore provides the sole pressure boundary for the process fluid at the bundle terminations.^[2]

The fundamental operating principle of the header box is hydraulic distribution: the inlet nozzle introduces process fluid into a defined compartment, from which it is distributed uniformly across all tubes of the corresponding pass. After traversing the tube length and exchanging heat with the ambient airstream, the fluid is collected in the outlet compartment of the opposing header box and discharged through the outlet nozzle. Partition plates welded horizontally within the box divide the internal volume into discrete pass compartments, enabling multi-pass tube-side flow arrangements that increase the tube-side fluid velocity and convective heat-transfer coefficient without increasing the physical tube length.^[3]

Key performance metrics for the header box include its pressure-drop contribution to the overall tube-side circuit, the uniformity of fluid distribution across all tubes within a pass, and its structural integrity under combined internal pressure, nozzle loads, and thermal-expansion forces. Non-uniform distribution caused by undersized nozzles, poor internal geometry, or maldistributing partition arrangements elevates the wall temperature of low-flow tubes, accelerating fouling, corrosion, and thermal degradation in those elements. A maldistribution factor below 5% is generally acceptable for most refinery and petrochemical services.^[3]

Historical Development

The evolution of air cooler header boxes parallels the broader development of air-cooled heat exchangers during the twentieth century. Early air-cooled exchangers, introduced in the 1920s and 1930s for natural-gas compression cooling in remote locations where cooling water was unavailable, employed simple flat-plate headers bolted to tube bundle end frames with full-face gaskets. These rudimentary designs were limited to low pressures and were susceptible to gasket leakage under thermal cycling.^[3]

The growth of petroleum refining in the post-World War II era, and in particular the widespread adoption of high-pressure catalytic reforming and hydrocracking processes in the 1950s and 1960s, created demand for header boxes capable of withstanding pressures above 100 bar in hydrogen-rich services. This drove the development of the plug-type header, in which machined threaded shoulder plugs replace the gasketed face plate, providing individual tube access while maintaining a robust pressure seal that does not depend on a large-area gasket. The American Petroleum Institute first codified plug-type header design requirements in API Standard 661, the initial edition of which established the foundational geometry and material specifications that remain the basis for current practice.^[2]

The 1970s brought additional design refinements driven by IMO environmental regulations and the oil price shocks that incentivized heat recovery. The split header was introduced to address the differential thermal expansion problems encountered in large multi-pass air coolers processing high-temperature residue streams, where pass-to-pass temperature differences could exceed 150°C in a single rectangular box, imposing unacceptable bending stresses on the tubesheet and partition welds. By the 1980s, finite element analysis (FEA) had been adopted to validate header-box structural designs under combined pressure, thermal-gradient, and nozzle-load conditions, replacing the purely empirical design rules of earlier decades.^[3][2]

The modern era has integrated computational fluid dynamics (CFD) into header-box design, enabling engineers to optimize internal geometry nozzle placement, partition-plate configuration, and compartment volume for uniform flow distribution. Advanced non-destructive examination techniques, including phased-array ultrasonic testing (PAUT) and digital radiography, have improved the reliability of weld-quality verification in fabrication, while risk-based inspection (RBI) methodologies under API 580 and API 510 now govern in-service inspection intervals and remaining-life assessments.^[2]

Pressure Vessel Mechanics

The header box is classified as a pressure vessel and designed in accordance with ASME Section VIII, Division 1, which governs the minimum required thickness of flat plates, nozzles, and weld-joint efficiency factors. For a flat rectangular plate under uniform pressure, the governing bending stress at the plate centre is given by:

${\displaystyle \sigma ={\frac {\beta pb^{2}}{t^{2}}}}$

where ${\displaystyle \sigma }$ is the maximum bending stress, ${\displaystyle p}$ is the design pressure, ${\displaystyle b}$ is the short dimension of the plate, ${\displaystyle t}$ is the plate thickness, and ${\displaystyle \beta }$ is a dimensionless coefficient dependent on the plate aspect ratio and boundary conditions. Since the header box plates are welded on all four edges, the boundary condition approximates a fully fixed plate, and ${\displaystyle \beta }$ values typically fall in the range 0.28–0.49 depending on the length-to-width ratio of the box face.^[4]

The minimum required plate thickness ${\displaystyle t_{min}}$ incorporating a corrosion allowance ${\displaystyle c}$ is:

${\displaystyle t_{min}=t_{calc}+c}$

For carbon steel in clean hydrocarbon service, the corrosion allowance is typically 1.5 to 3.0 mm; for sour-gas (H₂S-containing) services or corrosive process streams, it may be increased to 6 mm or supplemented by corrosion-resistant cladding or weld overlay. The maximum allowable working pressure (MAWP) is established at the completion of fabrication based on the actual measured thicknesses and materials certified by mill test reports.^[4]

Flow Distribution

Uniform distribution of tube-side fluid across all tubes in a pass is essential for achieving the design heat-transfer performance and for preventing local overheating and accelerated corrosion. The distribution uniformity depends on the ratio of the dynamic pressure at the nozzle inlet to the frictional pressure drop across the tube length. When the nozzle velocity head is large relative to the tube pressure drop, flow is preferentially directed to the nearest tubes, starving the remote tubes of flow. The nozzle inlet velocity head ${\displaystyle \Delta p_{nozzle}}$ is:

${\displaystyle \Delta p_{nozzle}={\frac {1}{2}}\rho v_{n}^{2}}$

where ${\displaystyle \rho }$ is the fluid density and ${\displaystyle v_{n}}$ is the nozzle inlet velocity. For acceptable distribution, API 661 recommends that the nozzle velocity head be maintained well below the tube-side pressure drop, typically by sizing nozzles to limit inlet velocities to 1.5–3.0 m/s for liquid services and 10–15 m/s for gas or vapor services. CFD analysis is employed in critical services to validate that the maldistribution factor defined as the ratio of the maximum to minimum tube flow rate minus one remains below 5%.^[2][3]

Thermal Expansion

The tube bundle expands longitudinally when heated by the process fluid, and this expansion must be accommodated without imposing excessive axial loads on the tubesheet-to-tube joints or the header-box structure. The free thermal expansion of the tube bundle ${\displaystyle \delta _{T}}$ is:

${\displaystyle \delta _{T}=\alpha \cdot L\cdot \Delta T}$

where ${\displaystyle \alpha }$ is the coefficient of thermal expansion of the tube material, ${\displaystyle L}$ is the effective tube length, and ${\displaystyle \Delta T}$ is the temperature rise above the ambient installation condition. For a carbon steel tube bundle 12 m long subjected to a 200°C temperature rise, ${\displaystyle \delta _{T}\approx 12\times 10^{-3}\times 12\times 200=28.8}$ mm a significant displacement that must be absorbed by the sliding supports at the outlet-header end without generating unacceptable tube-end bending moments.^[3]

API 661 designates the inlet header as the fixed end and provides slotted bolt holes in the outlet-header support pads to allow longitudinal sliding, typically ±6 mm for standard bundles and ±13 mm for long or high-temperature bundles. In multi-pass exchangers where adjacent tube passes operate at significantly different temperatures, the differential thermal expansion between pass groups introduces bending loads on the partition welds and tubesheet. When the pass-to-pass temperature difference exceeds approximately 110°C, a split header is required to decouple the thermal expansion of each pass compartment.^[2]

Nozzle Load Analysis

External forces and moments imposed by the connected piping on the header-box nozzles are a critical design input. These loads arise from thermal expansion of the piping system, deadweight, wind, and seismic actions. Excessive nozzle loads distort the tubesheet, open tube-to-tubesheet joints, or crack welds at the nozzle-to-plate interface. API 661 Table 3 specifies allowable nozzle load envelopes as a function of nozzle diameter and schedule, and compliance is verified by a piping flexibility analysis whose results are used as boundary conditions in the finite element analysis of the header box.^[2]

The combined stress at a nozzle-to-plate weld is assessed using the interaction formula:

${\displaystyle \left({\frac {F_{x}}{F_{x,allow}}}\right)^{2}+\left({\frac {F_{y}}{F_{y,allow}}}\right)^{2}+\left({\frac {F_{z}}{F_{z,allow}}}\right)^{2}+\left({\frac {M_{x}}{M_{x,allow}}}\right)^{2}+\left({\frac {M_{y}}{M_{y,allow}}}\right)^{2}+\left({\frac {M_{z}}{M_{z,allow}}}\right)^{2}\leq 1.0}$

where ${\displaystyle F}$ and ${\displaystyle M}$ denote the applied forces and moments in each principal axis, and subscript allow denotes the API 661 limit for that load component. Nozzle loads exceeding these limits require piping layout modifications, the addition of pipe supports, or structural reinforcement of the header box.^[2]

Plug-Type Header

The plug-type header is the predominant configuration in refinery and petrochemical service, and the design mandated by API 661 for all air coolers handling hydrocarbon liquids or gases. Its defining feature is a machined face plate into which one threaded, shouldered plug is installed directly opposite each tube end. Each plug is independently removable, allowing access to a single tube for inspection, re-rolling of the tube-to-tubesheet joint, mechanical cleaning, or installation of a tube plug without disturbing the adjacent tubes, the piping connections, or the pass partition arrangement.^[2]

The plug shoulder compresses a sealing element fiber, spiral-wound metallic, or soft-iron ring gasket against a machined seat in the face plate. The torque applied to the plug during installation is specified by the exchanger manufacturer based on the gasket seating stress required to achieve tightness at the design pressure. Plug material must be compatible with the process fluid and must have a coefficient of thermal expansion close to that of the face plate, since differential expansion under thermal cycling is the primary cause of plug-seat leakage over the service life of the exchanger. Stainless steel plugs in carbon steel headers are therefore avoided in high-cycling services, where austenitic-versus-ferritic expansion mismatch progressively loosens the plug compression.^[2]

The principal limitation of the plug-type header is the cumulative maintenance burden associated with a large number of plugs a single header box in a large air cooler may contain several hundred plugs each requiring inspection and re-torquing at every turnaround. However, this granular access capability is indispensable in services subject to tube corrosion, erosion, or process-side fouling deposits, where individual tube identification and targeted remediation are essential to managing partial bundle degradation without removing the entire unit from service.^[2]

Cover-Plate (Removable Bonnet) Header

In the cover-plate design, the entire face of the header box opposite the tubesheet is formed by a single bolted flat plate. Removal of this plate accomplished by unbolting the perimeter flange exposes all tube ends simultaneously, enabling rapid visual inspection of the full tubesheet face, mechanical hydro-jetting of the tube bundle interior, and re-rolling or plugging of any tube without the sequential plug removal required in plug-type headers. This configuration is particularly advantageous in services where heavy tube-side fouling is anticipated and frequent cleaning outages are required.^[3]

Cover-plate headers are limited to design pressures below approximately 10 bar (150 psig) because the large-area gasket required to seal the cover plate becomes impractical at higher pressures: achieving adequate gasket seating stress across a large flat surface requires bolt loads that exceed the structural capacity of the cover plate itself or the flange ring. API 661 therefore restricts cover-plate construction to auxiliary utility services lube-oil coolers, hot-oil circuits, and cooling-water circuits where design pressures are inherently low and process leak-tightness requirements are less stringent than for hydrocarbon vapor services.^[2]

The cover-plate gasket is typically a full-face or ring-type configuration made from compressed fiber, PTFE-encapsulated, or spiral-wound material, selected for compatibility with the process fluid and the design temperature. Where partition plates are used in a multi-pass cover-plate header, separate partition gaskets seal the pass boundaries at the cover-plate interface, and the alignment of partition gasket grooves in the cover plate with the partition welds in the box body is a critical dimensional requirement during fabrication and reassembly.^[2]

Split Header

The split header comprises two or more horizontally stacked rectangular compartments, each serving a distinct pass of the tube bundle, physically separated by a structural gap or by independent box structures rather than sharing a common interior volume. This design is required when the temperature difference between adjacent passes in a multi-pass bundle exceeds approximately 110°C (200°F), a threshold above which the differential longitudinal thermal growth of the two tube groups generates bending stresses in the tubesheet and partition welds that exceed allowable limits under ASME Section VIII and the supplementary structural rules of API 661.^[2]

By providing each pass compartment with independent support and freedom of thermal movement, the split header converts the intolerable bending load scenario of a monolithic box into a series of independently expanding structural elements, each of which is within its allowable stress envelope. The gap between adjacent split compartments accommodates the differential growth and is typically sealed with a flexible metallic bellows or simply left open on the air-side face where no process containment is required.^[3]

Split headers are characteristically found in hot-service air coolers processing fired-heater effluent streams, crude tower overhead vapors, or hydrocracker reactor effluents, where inlet temperatures may approach 350–400°C and the temperature differential between the first and second tube passes is inherently large. Their fabrication is more complex and costly than monolithic headers, requiring careful alignment of the multiple compartment bodies with the corresponding tube rows and precise dimensional control during assembly to prevent tube-end bending at the tubesheet joints.^[2]

Manifold Header

In certain extreme high-pressure services typically gas coolers operating above 100 bar the tube ends are connected to cylindrical manifold pipes rather than to a rectangular box assembly. The cylindrical geometry provides superior pressure containment efficiency: hoop stress in a thin-walled cylinder under internal pressure is ${\displaystyle \sigma _{h}=pD/(2t)}$, whereas bending stress in a flat plate of equivalent wall thickness and span is several times higher for the same pressure, making flat-plate construction prohibitively thick and heavy at very high pressures.^[4]

Manifold headers are fabricated as machined or forged cylindrical vessels with drilled tube-entry bores on their lower face, aligned precisely with the tube bundle pitch. Sealing between the tube ends and the manifold bores is achieved by roller expansion, strength welding, or precision-machined ferrule fittings depending on the design pressure and temperature. The manifold configuration does not provide the same individual-tube access convenience as a plug-type header, and maintenance access typically requires removal of a bolted end cap that exposes a subset of tube ends rather than a single plug per tube.^[3]

Petroleum Refining

Petroleum refineries deploy air-cooled heat exchangers with header boxes across virtually every major process unit. In crude distillation, overhead condenser duties on the atmospheric and vacuum towers require header boxes designed for naphtha vapor and hydrogen sulfide-containing streams at moderate pressures of 3–8 bar. In catalytic hydrocracking and hydrotreating units, reactor effluent air coolers operate at hydrogen partial pressures of 100–200 bar, requiring heavy-wall plug-type header boxes in hydrogen-resistant chrome-moly steels (ASTM A387 Gr. 11 or 22) with PWHT and compliance with Nelson curve criteria for resistance to high-temperature hydrogen attack (HTHA).^[2]

Fluid catalytic cracking (FCC) main fractionator overhead condensers handle large volumes of catalytic naphtha vapors at design pressures of 3–5 bar, with header boxes subject to significant coking potential requiring plug-type access for periodic tube cleaning. Amine treating unit regenerator overhead condensers, which handle H₂S-rich acid gas streams, require header boxes in NACE MR0103-compliant materials with strict hardness controls on all weld heat-affected zones to prevent sulfide stress cracking.^[5]

Petrochemical Processing

In ethylene production, the quench water cooler and pyrolysis gas cooler air coolers must handle highly fouling pyrolysis gasoline streams at elevated temperatures, requiring plug-type header boxes with large-bore nozzles and design margins for tube plugging during on-stream operation. Methanol and ammonia synthesis loops include air-cooled synthesis-gas and product coolers where header boxes are designed for compressed gas streams at 100–300 bar, often necessitating manifold-type headers.^[3]

Ammonia refrigeration compressor inter- and after-coolers, and propylene refrigeration condenser air coolers in olefin plants, employ header boxes designed for NH₃ and C₃H₆ services respectively, with material selections that exclude copper and copper alloys due to their reactivity with ammonia, and that comply with stress-corrosion cracking avoidance requirements for propylene service.^[2]

Natural Gas Processing

Onshore gas compression stations use air-cooled inter- and after-coolers to cool high-pressure natural gas between compression stages and before pipeline injection. Header boxes in these services handle natural gas at pressures of 70–150 bar and temperatures up to 150°C, in carbon steel or low-alloy construction with NACE MR0175 compliance for fields containing H₂S. Amine regeneration overhead condensers and acid-gas removal unit coolers in sour-gas processing plants require the same sour-service materials disciplines as in refinery amine units.^[5]

In liquefied natural gas (LNG) facilities, propane precooler and mixed-refrigerant condenser air coolers operate at cryogenic to moderate temperatures with refrigerant vapors at 15–40 bar. Header boxes in cryogenic refrigerant service are fabricated from 3.5% nickel steel or austenitic stainless steel to maintain adequate fracture toughness at low temperatures, and are designed with attention to thermal-shock risks during start-up and emergency depressurization.^[3]

Power Generation

Gas turbine compressor intercoolers and generator air coolers use compact plate-type header manifolds designed for large air and cooling-water flow rates at low pressures. Air-cooled steam condensers (ACC) deployed in arid regions where freshwater scarcity makes evaporative cooling towers impractical employ very large, multi-row header manifolds that receive low-pressure exhaust steam from the turbine at sub-atmospheric pressures of 0.05–0.15 bar. These manifolds handle the high specific volume of the steam (several cubic meters per kilogram) and must be fabricated to avoid condensate pocketing that would reduce the effective condensing surface area.^[3]

Combined-cycle power plants incorporate air-cooled heat exchangers for gas-turbine compressor intercooling and for the cooling of thermal-fluid circuits in heat-recovery steam generators (HRSG). Transformer oil coolers in electrical substations associated with offshore platforms are another application where compact header-box designs are required for space-constrained installations.^[2]

Routine Inspection Practices

In-service inspection of air cooler header boxes is governed by API 510, the Pressure Vessel Inspection Code, which defines inspector qualifications, inspection intervals, and fitness-for-service assessment methodologies. At planned turnaround outages, inspection typically encompasses removal and examination of all plugs in plug-type headers checking for plug-body corrosion, thread damage, and gasket seat condition and visual inspection of all tube ends for corrosion, erosion pitting, and deposit accumulation. Tube-to-tubesheet joints are inspected for evidence of leakage, indicated by mineral deposits, corrosion products, or process-fluid residues around the joint periphery.^[2]

Ultrasonic thickness (UT) measurements are taken on header-box plate sections, nozzles, and welds to track corrosion rates and verify that remaining wall thickness remains above the minimum required thickness at the next inspection interval. Inspection locations are selected based on the geometry of the header box (focusing on areas of maximum expected corrosion: low-velocity zones, vapor-liquid interfaces, and areas near inlet nozzles where turbulent impingement is highest) and on operating history. Inspection findings are documented in an equipment file maintained in accordance with API 510 requirements, and corrosion rates are calculated to set the next required inspection date based on the remaining corrosion allowance divided by the measured rate.^[4]

Tube Plugging

When a tube-to-tubesheet joint or a tube body develops a through-wall leak that cannot be repaired in situ by re-rolling or re-welding, the affected tube is isolated by installing tapered metal plugs one at each header end that seal the tube bore and divert flow to the remaining active tubes. API 661 specifies that a tube bundle may continue in service with up to 10% of its tubes plugged, beyond which the reduction in heat-transfer area is considered excessive and the bundle should be replaced or retubed during the next available maintenance window.^[2]

Plug material must be compatible with the process fluid, must have adequate strength at the design temperature to resist the tube-side pressure, and must be capable of providing a reliable seal against the tube bore. Tapered plugs are driven by a hydraulic or pneumatic plug driver to a specified interference fit calculated to produce a contact pressure at the tube bore interface that exceeds the design pressure by an adequate safety margin. In high-pressure hydrogen services, the taper geometry is designed to prevent hydrogen-induced outward migration of the plug under cyclic pressure fluctuations.^[2]

Chemical and Mechanical Cleaning

Fouling of the tube-side surface by polymerized hydrocarbons, corrosion products, inorganic scale, or biological deposits in cooling-water services increases the tube-side thermal resistance, reduces the overall heat-transfer coefficient, and elevates the tube-side pressure drop. The overall heat-transfer coefficient with fouling ${\displaystyle U_{f}}$ is related to the clean coefficient ${\displaystyle U_{c}}$ by:

${\displaystyle {\frac {1}{U_{f}}}={\frac {1}{U_{c}}}+R_{fi}+R_{fo}}$

where ${\displaystyle R_{fi}}$ and ${\displaystyle R_{fo}}$ are the tube-side (inside) and air-side (outside) fouling resistances respectively, in m²·K/W. Typical tube-side fouling resistances for petroleum streams range from 0.0002 m²·K/W for clean condensates to 0.0010 m²·K/W for heavy residue streams, representing reductions in effective heat-transfer coefficient of 20–50% relative to clean design values.^[3]

Mechanical cleaning by high-pressure water lancing (100–200 bar) is the most common method for removing soft deposits from tube bores in plug-type headers after individual plug removal. Chemical cleaning with inhibited hydrochloric acid (2–5% concentration) is used for inorganic scale removal, with circulation times of 4–8 hours followed by neutralization and water flushing. Chemical cleaning must account for material compatibility: HCl is unsuitable for stainless steel or titanium components, where citric or phosphoric acid formulations are preferred.^[2]