Walk any plant room and you are surrounded by heat exchangers, whether or not anyone calls them that. The chiller has two. The boiler is one. The cooling tower is a variation on the theme. Every air handling unit has at least one coil, and the plate unit where the district cooling supply arrives is the most photogenic example of the family. Once you see a building this way, equipment that looked unrelated turns out to be the same idea packaged differently, and maintenance symptoms that looked unrelated turn out to share a cause.
The message up front: a heat exchanger transfers heat between two fluids while keeping them separate. Everything else is a trade. More surface area and tighter flow paths give better transfer but foul faster and cost more pressure drop. Sealed construction survives higher pressures but cannot be opened for cleaning. Heat exchangers almost never fail suddenly. They degrade quietly, and the whole maintenance discipline is about noticing that degradation before it becomes a capacity problem.
1. What a heat exchanger actually does
A heat exchanger transfers thermal energy between two or more fluids at different temperatures without allowing those fluids to contact each other. Drop the second half of that sentence and you no longer have a heat exchanger, you have a mixing vessel, which is a much simpler and cheaper thing to build.
That non-mixing constraint is the whole engineering problem. If you were allowed to mix the fluids you would pour them together and be done. Because you are not, every design has to put a solid barrier between them and then spend all its ingenuity making that barrier as thin and as large in area as the operating pressures, the fluid chemistry and the budget permit. Thin for conduction, large for total transfer, strong enough not to burst. Those three requirements pull against each other, and the entire catalogue of heat exchanger types is a set of different resolutions of that tension.
The duties themselves vary widely: rejecting heat from a refrigerant circuit to condenser water, transferring heat from combustion gas into a water loop, hydraulically separating a landlord's chilled water from a tenant's so the loops exchange energy but never water or contamination, or recovering heat from exhaust air into incoming fresh air. All four are the same device family serving four different purposes.
2. The physics, in plain language
Heat moves in three ways, and the distinction matters when you are diagnosing a poorly performing unit.
- Conduction is heat moving through a solid, or between materials in direct contact, by molecular interaction. No bulk movement of matter, just energy passing along. A metal wall with hot fluid on one side and cold on the other conducts heat from the hot face to the cold face.
- Convection is heat carried by the bulk movement of a fluid. Water flowing past a warm surface picks up heat and carries it away as cooler water arrives to take its place. This is what gets heat to and from the wall.
- Radiation is heat transferred as electromagnetic energy, needing no medium. It dominates in furnaces and anything glowing, and is largely a side note in the liquid-to-liquid and air-to-liquid exchangers most facilities teams deal with.
In a typical heat exchanger all three are present but only two do the work. Convection brings heat from the bulk of the hot fluid to the hot face of the separating wall, conduction carries it through, and convection on the other side carries it away into the cold fluid. Radiation contributes very little at the temperatures normal building and plant exchangers operate at.
Conduction through the wall is the mechanism to focus on because it is the one maintenance most directly degrades. The wall material and thickness are fixed at design time, but the effective thickness in service is the metal plus whatever has deposited on both faces of it: scale, biofilm, silt, corrosion product, grease, dust. Those deposits conduct badly. Add a layer you cannot see to both sides of a wall that was deliberately made as thin as safety allows, and you have quietly redesigned the heat exchanger into a worse one. That is fouling, and it is the operational heart of this subject.
The mental model worth keeping
Total heat transferred depends on three things you can influence: how much surface area the two fluids share, how big the temperature difference is across the wall, and how readily heat crosses that wall. A designer sizes the first. Operation sets the second. Maintenance determines the third. Most heat exchanger performance complaints are the third one degrading while everyone looks at the first two.
3. Flow arrangement: counterflow, parallel flow and crossflow
Two heat exchangers can have identical surface area, materials and fluids and still perform differently, because how the streams are routed relative to each other changes how much of the available temperature difference is usable along the length of the device. It can be understood entirely without mathematics.
In parallel flow, both fluids enter at the same end and travel the same way. At the inlet the difference between them is at its maximum, so transfer is vigorous, but as they travel together the hot one cools and the cold one warms and the gap closes rapidly. By the far end the streams are converging on a common middle temperature with very little difference left to drive transfer, so the last stretch of surface area does almost nothing. The cold fluid leaving can never be warmer than the hot fluid leaving, because they are chasing each other toward the same value.
In counterflow, the fluids enter at opposite ends and travel in opposite directions, so the hot fluid meets the cold stream at its warmest at one end and at its coldest at the other. The temperature difference therefore stays comparatively even along the whole length instead of collapsing, and every part of the surface area is working rather than just the first part. Because the cold stream exits where the hot stream enters, the cold outlet can approach the hot inlet far more closely than parallel flow ever permits. Same metal, same fluids, more heat moved. This is why counterflow is the default wherever geometry allows it.
In crossflow, the fluids move at right angles, typically one through tubes and one across them. This is the natural arrangement whenever one fluid is air, because you cannot sensibly duct air along a narrow parallel path. It performs between the other two, and it is chosen because the physical arrangement demands it rather than because it is thermally optimal. Many real coils are multi-pass arrangements where the tube circuiting is routed so the overall effect leans counterflow even though each pass is crossflow.
| Arrangement | Behaviour | Where it is used |
|---|---|---|
| Counterflow | Fluids travel in opposite directions. Temperature difference stays comparatively even along the length, so all the surface area contributes. Permits the closest approach between outlet and opposite inlet. | The default for liquid-to-liquid duty: plate exchangers, most shell and tube arrangements, district cooling interface units, hydraulic separation. |
| Parallel flow | Fluids travel in the same direction. Large difference at inlet that collapses quickly. The downstream surface does little work, and outlet temperatures converge rather than cross. | Rarely chosen for thermal reasons. Used where it eases piping, where the fast initial transfer is wanted, or where a gentler temperature gradient protects the material. |
| Crossflow | Fluids move at right angles. Performance sits between the other two. Often built as multi-pass so the overall effect approaches counterflow. | Anywhere air is one of the fluids: AHU coils, finned-tube coils, air-cooled condensers, air-to-air recovery devices, radiators. |
4. Approach temperature and what a close approach costs
Approach temperature is the most useful operating number in the heat exchanger world. It is simply the temperature difference remaining between the two streams at the point where they have come closest together: if the hot side enters at one value and the cold side leaves slightly below it, that remaining gap is the approach.
A close approach means the exchanger has extracted nearly all the heat physically available from the difference presented to it. That is thermally excellent, and it costs money, because each further increment of extraction demands disproportionately more surface area as the driving difference shrinks. A designer chasing a very close approach specifies a larger, heavier, costlier unit with more plates or tubes and frequently more pressure drop, meaning more pump energy for the life of the installation. Approach is therefore a capital-versus-operating trade with a duty-specific answer: worth paying for where the extracted energy is valuable and continuous, not worth it on a minor duty.
Operationally it doubles as a health indicator. If a unit that reliably produced a certain approach when commissioned now produces a noticeably wider one under comparable flows and inlet temperatures, something has come between the two fluids. That is your fouling signal, and it usually shows before anyone notices a comfort or capacity complaint. Recording approach at commissioning and trending it is one of the cheapest high-value habits a plant room team can adopt, and it is exactly the kind of derived measure that belongs in a plant analytics view rather than a notebook. The chiller plant analytics pillar covers how to build that trending properly.
5. The main types and what each trades off
Many geometries exist, but a facilities or plant audience meets a manageable handful. What follows is an orientation to each, framed around what it gives you and what it costs you.
Shell and tube. A bundle of tubes inside a cylindrical shell, one fluid through the tubes and the other around them, with baffles directing the shell-side flow across the bundle. The oldest and most universally applicable arrangement, and repairable in ways compact units are not: tubes can be plugged or replaced, the bundle can often be withdrawn, and tube interiors can be mechanically cleaned. The trade is bulk. These units are large, heavy and need clearance for bundle withdrawal, which matters in a crowded plant room.
Gasketed plate. A stack of thin corrugated plates clamped between end frames, with gaskets forming channels so the fluids pass through alternate gaps. The corrugations induce turbulence, so transfer per unit of area is very high and the unit is dramatically more compact than an equivalent shell and tube. It achieves a close approach well, and capacity can be adjusted after installation by adding or removing plates. The trades are the gaskets, a consumable with a finite life and a chemical and temperature compatibility envelope, and a lower pressure ceiling. The redeeming feature is that it opens: the frame unbolts, the plate pack comes apart, and every surface can be inspected and cleaned by hand.
Brazed plate and welded plate. The same corrugated-plate idea with the plates permanently joined. Removing the gaskets removes the gaskets' limitations, so these handle higher pressures in a package that is more compact and cheaper for a given duty. The trade is absolute: they cannot be opened. No plate pack to separate, no surface to brush. Cleaning is restricted to chemical circulation, and when that no longer restores performance the unit is replaced rather than refurbished. Semi-welded variants are a deliberate compromise, pairing welded channels for an aggressive fluid with gasketed channels for a benign one.
Air-cooled and finned-tube coils. Tubes carrying liquid or refrigerant, fitted with closely spaced fins to multiply the air-side surface area, with air blown or drawn across them. Air is a poor transfer medium compared with water, so the fins compensate with sheer area. The family covers air-cooled condensers, dry coolers and the coils inside every air handling unit. The advantage is that no water is needed, which in a water-scarce region is a serious consideration. The trades are footprint, fan energy, sensitivity to ambient air temperature, and a fin surface that blocks with airborne debris and bends easily during careless cleaning.
Plate-fin and thermal wheel, for air-to-air recovery. A plate or plate-fin block keeps exhaust and supply air in separate adjacent channels with no moving parts and no cross-contamination, typically in crossflow. A thermal wheel is a slowly rotating matrix that absorbs heat in the exhaust stream and releases it as it rotates into the supply stream. Wheels recover more for a given size and some transfer moisture as well as heat, valuable in a humid climate, but they carry a small inherent air carryover, so they are unsuitable where the exhaust is contaminated. Both live inside air handling units, which is where most FM readers meet them. The AHU guide covers where coils and recovery devices sit.
Shell and coil. A coiled tube inside a shell. Simple, compact and inexpensive at smaller duties, common in domestic hot water calorifiers and small process heating. The trade is that the coil interior is awkward to clean mechanically because of its curvature, so it belongs where the fluids are clean or the unit is cheap enough to treat as replaceable.
| Type | Construction | Strengths | Limitations | Mechanically cleanable? |
|---|---|---|---|---|
| Shell and tube | Tube bundle inside a cylindrical shell with shell-side baffles | Handles the highest pressures and temperatures of the common types; very wide fluid tolerance; tubes can be plugged, replaced or rodded; bundle often withdrawable | Large and heavy; needs withdrawal clearance; lower transfer per unit area so more metal for the same duty | Yes, tube side readily; shell side with more difficulty |
| Gasketed plate | Corrugated plate pack clamped in a frame, gaskets forming channels | Very compact; high transfer; achieves a close approach; capacity adjustable by adding plates; fully openable | Gaskets are consumable with chemical and temperature limits; lower pressure capability than shell and tube; more channels to block | Yes, fully, by opening the plate pack |
| Brazed plate | Plate pack permanently brazed into a sealed block | Most compact; no gaskets to replace; higher pressure capability than gasketed; low cost per unit of duty | Cannot be opened; chemical cleaning only; effectively a replace-not-repair item; intolerant of debris | No |
| Welded / semi-welded plate | Plates welded, or welded channel pairs alternating with gasketed channels | Suits aggressive fluids and more demanding conditions while keeping plate compactness; semi-welded keeps one side openable | Fully welded cannot be opened; repair options limited; higher cost than gasketed | Fully welded no; semi-welded on the gasketed side only |
| Air-cooled / finned coil | Finned tube bank with fans moving air across it | No water required; simple; well suited to condensing and to AHU heating and cooling duty | Large footprint; fan energy; performance tracks ambient air temperature; fins block and bend easily | Air side yes, with care; tube side depends on the circuiting |
| Plate-fin recovery block | Static plate or plate-fin block separating two air streams | No moving parts; no cross-contamination between supply and exhaust; low maintenance burden | Bulkier than a wheel for the same recovery; generally heat only, not moisture; air-side pressure drop | Air side, partially and with restricted access |
| Thermal wheel | Rotating matrix passing alternately through exhaust and supply air | High recovery for its size; some types transfer moisture as well as heat, which matters in humid climates | Moving part with drive, belt and seals to maintain; small inherent air carryover, so unsuitable for contaminated exhaust | Surface cleaning only |
| Shell and coil | Coiled tube inside a shell | Simple, compact and inexpensive at smaller duties; common in calorifiers and small process heating | Limited to smaller duties; coil interior awkward to clean; usually treated as a replaceable item | Largely no, in practice |
6. Where heat exchangers appear in a building or plant
Understanding this device family collapses a lot of apparently separate equipment into one concept:
- Inside the chiller. A chiller contains at least two. The evaporator transfers heat from the chilled water loop into the refrigerant, and the condenser transfers it out of the refrigerant into condenser water or ambient air. Both are usually shell and tube in larger water-cooled machines, and both are why chiller performance is so sensitive to water quality. The chiller guide is the companion to this article: read together, the two show that a chiller is largely a refrigerant circuit wrapped around two heat exchangers.
- The boiler. A heat exchanger with a flame in it. Combustion gas on one side of a metal surface, water on the other, no mixing. Different boiler types are different geometries for arranging that surface, and the fire-tube versus water-tube distinction is a shell and tube argument about which fluid goes inside the tubes. See the boiler guide.
- The cooling tower. The interesting exception. An open tower deliberately puts water in direct contact with air and relies on evaporation, so it is not a closed heat exchanger at all; closed-circuit coolers restore the separation by putting a coil in the tower. It is worth knowing precisely because it breaks the non-mixing rule, which is why towers carry water treatment, hygiene and consumption issues closed exchangers do not. See the cooling tower guide.
- Air handling units. Cooling coils, heating coils and, where fitted, an air-to-air recovery device. Every one is a heat exchanger, and the coil is where most of an AHU's maintenance attention should go.
- Domestic hot water and the district cooling interface. Calorifiers and plate-type water heaters transfer heat from a primary loop into potable water while keeping the two completely separate, which here is a hygiene requirement as much as an engineering one. The district cooling interface gets its own section below.
For how these pieces fit together as a system rather than as a list, the HVAC systems and components pillar is the wider frame this article sits inside.
7. The plate heat exchanger in district cooling and hydraulic separation
In the Gulf in particular the plate heat exchanger has a starring role, because district cooling is so widespread. Chilled water is generated centrally and distributed to buildings, and at each building an interface station lets the operator's water exchange heat with the building's own chilled water loop across a plate heat exchanger. The two waters never touch.
The reasons for separating the loops rather than piping district water around the building are practical and financial. It draws a clean contractual boundary, with the operator responsible for one side, the building for the other, and consumption metered at the interface. It isolates the systems hydraulically, so the network's pressure regime is not imposed on terminal equipment rated for considerably less. It contains contamination, so a dirty building loop does not foul a network shared with neighbouring buildings. And it keeps water treatment regimes independent.
The same logic without any district operator is called hydraulic separation and is used inside buildings too: splitting a high-rise into pressure zones, isolating a legacy loop of uncertain cleanliness from a new one, or separating a landlord system from a tenant system. In each case a plate exchanger passes energy while blocking water, pressure and dirt.
The performance of that interface matters, and this is where approach temperature becomes a commercial issue rather than a technical curiosity. If the exchanger fouls and the approach widens, the building side can no longer reach the temperature it was designed around. Terminal units struggle, occupants complain, and the usual response is to increase flow, which raises pumping energy and often worsens the return temperature the operator cares about. A fouled interface exchanger can therefore generate a comfort complaint, an energy penalty and a commercial dispute at once, from a cause two thermometer readings would have identified months earlier.
The habit I would recommend
Record the four temperatures and both differential pressures across every significant heat exchanger at commissioning, while it is clean and performing to design. Without that reference you end up arguing about whether the unit was ever capable of what you now want from it, which is an argument nobody wins.
8. Fouling: the honest operational core
If you remember one operational thing about heat exchangers, make it this. They rarely fail suddenly. They degrade, quietly and progressively, and the degradation is almost always fouling: the accumulation of unwanted material on the transfer surface. The metal is still intact, the pumps still run, nothing alarms, and yet the device is doing measurably less than it did. Fouling is also the failure mode most within your control, so it is worth understanding rather than treating as inevitable. What fouls, and with what, depends on the fluid and the type:
- Scale. Mineral deposits, predominantly calcium and magnesium compounds, precipitating out of water onto warm surfaces. Worse with harder water and hotter surfaces, so heating and condenser duty suffer most. Hard, adherent, and the classic target for chemical cleaning.
- Biofouling. Biofilm, algae and microbiological growth, thriving in open systems, warm water and low-velocity regions. It insulates extremely effectively for its thickness and can drive under-deposit corrosion as well as reducing transfer. Condenser water fed from open cooling towers is the prime candidate.
- Particulate and silt. Sand, dust, construction debris, corrosion product, pipe scale. It settles in low-velocity zones and blocks narrow channels outright. Plate exchangers are particularly exposed because their channels are narrow by design, which is exactly why a strainer upstream of one is not optional.
- Corrosion product. The exchanger generating its own fouling layer, usually a symptom of a water treatment problem rather than a cleaning problem. Cleaning it without addressing the chemistry just resets a clock.
- Air-side fouling. Dust, lint, fibre, pollen, insects and in some regions sand, packing into the fin block of a coil or air-cooled condenser. It restricts airflow and insulates the fins at the same time, hitting both mechanisms at once.
The symptoms are consistent regardless of type, and they tend to appear in a recognisable order. First, approach temperature widens, usually while the plant is still meeting demand, because a deposit layer directly impedes conduction so the streams cannot get as close as they used to. Nobody notices unless someone is trending it. Then duty falls: inability to hold a setpoint at design conditions, longer run times, more machines running than the load should require, comfort complaints on the hottest days first. And pressure drop rises, because deposits and debris narrow the flow path, showing as differential pressure across the unit and pumps working harder on the water side, or reduced airflow on the air side.
Read together those symptoms indicate the mechanism. A widening approach with stable pressure drop points at a thin insulating layer, scale or biofilm. A sharply rising pressure drop points at blockage or particulate, and on a plate pack can mean whole channels have closed. Either way, by the time anyone complains the unit has been degrading for a long time, with the cost paid in pump and fan energy and lost plant headroom throughout.
The uncomfortable part
Fouling is usually a water treatment and filtration failure presented as a heat exchanger problem. Cleaning restores performance temporarily, but if the chemistry, the strainers, the makeup water quality or the system flushing regime caused the deposit, it will come back on the same timetable. A cleaning interval that keeps shortening is not telling you the exchanger is old. It is telling you nobody has fixed the upstream cause.
9. Cleaning approaches and the maintainability consequence
There are three broad routes back to a clean heat transfer surface, and which are available to you was decided the day the equipment type was selected.
Chemical cleaning in place, often called CIP. A solution matched to the deposit is circulated through the unit without dismantling it, then neutralised and flushed. Acidic formulations address mineral scale, alkaline ones organic and biological deposits, biocides microbiological growth. It is the least invasive route, needs little downtime, and is often the only route available. Its limits are real: chemistry dissolves what it is formulated for and leaves the rest, so silt and debris often survive entirely, and the solution follows the path of least resistance, so a fully blocked channel gets very little circulation and therefore very little cleaning. The worst-affected areas are the ones the method reaches least well. Aggressive chemistry also has to respect the materials it contacts, gaskets very much included.
Opening the unit for mechanical cleaning. On a gasketed plate exchanger the frame is released, the plate pack separated, and every plate inspected and cleaned by hand. This is the most thorough option on a plate unit: you see the true condition of every surface, you find the cracks and erosion and the plates starting to fail, and you remove deposits no chemistry would have shifted. The costs are longer downtime, skilled labour, careful reassembly with correct plate orientation and clamping, and gaskets that usually need replacing. Worth it, particularly the first time you open a unit nobody has looked inside.
Tube-side mechanical cleaning on shell and tube. Brushing, rodding or high-pressure water jetting through the tubes, with the bundle withdrawn where design and clearance allow. Direct, effective and verifiable, and it pairs naturally with tube integrity testing. Shell-side cleaning is harder and more likely to need chemical circulation. The constraints are access and plant room clearance for bundle withdrawal, which is precisely the clearance that gets built over during the first refurbishment.
The maintainability consequence follows directly, and it is the part of a selection decision that gets least attention at the time and most later. Choosing a brazed or fully welded plate exchanger buys compactness, pressure capability and a lower purchase price, and permanently forecloses mechanical cleaning. Once chemical cleaning no longer restores that unit, the only option is replacement. On a small, non-critical, clean-fluid duty that is a sound trade. On a large, critical duty with water of uncertain quality it converts a maintainable asset into a consumable one, and the whole-life cost picture shifts in a way the capital comparison never showed. None of which argues against sealed units; it argues for making the choice knowingly, with the filtration and water treatment regime specified alongside the exchanger rather than after it. The boiler and chiller preventive maintenance guide and the HVAC preventive maintenance guide both cover where exchanger and coil tasks sit in a maintenance programme.
Competence and statutory scope
Nothing here is a sizing, selection or design method. Thermal and mechanical design, material selection and pressure rating belong to a qualified engineer working to the applicable design code and the manufacturer's data for the specific unit. Separately, pressure equipment is subject to statutory inspection regimes in many jurisdictions, and those schemes differ substantially in scope, thresholds and required intervals, so confirm what applies to your equipment with your local competent authority and your insurer rather than assuming a single global rule.
The idea to walk away with
A heat exchanger transfers heat between two fluids without mixing them, and that one constraint generates everything else. It forces a separating wall that must be thin for conduction, large in area for total transfer, and strong enough for the pressures involved. Every type is a different resolution of those competing demands, and every type's limitations are the price of its strengths. Shell and tube buys robustness and repairability with bulk. Gasketed plate buys compactness and a close approach at the cost of gasket life and a lower pressure ceiling. Brazed and welded plate buy compactness and pressure capability by surrendering the ability to be opened.
Operationally, internalise that these devices degrade rather than fail. Fouling adds an unseen layer to a wall deliberately made thin, and the loss shows as a widening approach, a falling duty and a rising pressure drop, in roughly that order, long before anyone complains. The organisations that manage heat exchangers well are not the ones with the best equipment. They are the ones who recorded clean performance at commissioning, trend against it, fix the water treatment and filtration that caused the fouling rather than only cleaning the symptom, and chose types whose cleaning requirements they can actually meet.
Final thoughts
The payoff of understanding heat exchangers is that it turns a plant room from a collection of unrelated boxes into a system you can reason about. The chiller's evaporator, the boiler's heat transfer surface, the AHU coil, the recovery wheel and the district cooling interface are the same idea, so they share a failure mode and answer to the same three questions. Are the two fluids getting as close as they used to? Is the pressure drop where it was? Has anything changed upstream in the water or the air that would explain a deposit?
Those questions cost nothing but a set of gauges and the discipline to write the readings somewhere they will be looked at again. That discipline, more than any equipment selection, is what separates a plant that holds its design capacity for years from one that quietly loses it and then blames the original design. If you inherit a site with no commissioning baseline, establishing one after a thorough clean is the most useful single day of work available to you.
Disclosure
Alongside advisory work I also build a CMMS and CAFM platform, so I have a commercial interest in this category. Nothing above is a recommendation for it, and no vendor named here has paid for inclusion or had any editorial input. Weigh the analysis accordingly.
Losing plant capacity you cannot account for?
Independent advisory on plant room performance baselines, approach temperature trending, heat exchanger and coil PPM regimes, and getting condition data into a maintenance system where it is actually acted on. 22+ years across utilities, oil and gas, manufacturing, government and facility operations.
Book a conversationRelated reading: What is HVAC: systems and components, What is a chiller, What is a boiler, What is a cooling tower, What is an AHU, Boiler and chiller preventive maintenance, Preventive maintenance for HVAC systems, Chiller plant analytics.
Muhammad Abbas
CMMS / CAFM Manager & Independent Advisor · 22+ years across enterprise CMMS, EAM, CAFM and ERP implementations in utilities, oil and gas, manufacturing, government and facility operations.
Work with me