Somebody has a stream to cool and a budget to spend, and they want to know what kind of heat exchanger to put in. It is a fair question and it usually gets a bad answer, because most people answering it are familiar with one type and will steer towards it.
The useful way to think about it is that there are perhaps six families in common industrial use, each one is genuinely the best available answer for some situations and a poor answer for others, and the choice is driven by a short list of conditions rather than by preference. This guide sets out what each family does well, where it stops working, and the questions that decide between them.
Shell and tube
A bundle of tubes inside a cylindrical shell. One fluid passes through the tubes, the other around them, with baffles directing the outer flow across the bundle.
It dominates heavy industry for reasons that have nothing to do with fashion. It handles high pressure and high temperature better than any alternative, it can be built in essentially any material and almost any size, it is mechanically robust, it is well understood by every engineer and inspector you will encounter, and its design is codified in TEMA standards that everybody works to.
Its weakness is compactness. Achieving a given heat transfer area takes more space and more metal than several alternatives, and where plot space is tight or the approach temperature is very close, other types will beat it. The internal arrangement also decides maintainability, which is a subject in itself; our note on reading TEMA type letters covers how fixed tubesheet, U tube and floating head designs differ in service.
Choose it for: high pressure, high temperature, dirty service, large duties, anything requiring a code stamped pressure envelope in an exotic alloy.
Plate and frame
Thin corrugated plates clamped in a frame, with gaskets creating alternating channels so the two fluids flow in adjacent gaps.
The corrugations induce turbulence at low velocity, which makes the heat transfer coefficient very high. The practical consequences are striking: a plate unit achieving the same duty as a shell and tube can be a fraction of the size and hold a fraction of the fluid volume. It can also work to a very close approach temperature, which makes it excellent for heat recovery duties. It can be opened for cleaning in a couple of hours, and capacity can be changed later by adding or removing plates, which no other type allows.
The limits come from the gaskets. Pressure and temperature ranges are considerably narrower than shell and tube, gasket material has to be compatible with both fluids, and gaskets are a consumable with a service life. The narrow channels also block readily, so plate units are unsuited to fluids carrying fibres or particles.
Brazed and fully welded plate variants remove the gasket and extend the pressure and temperature range, at the cost of no longer being openable for mechanical cleaning.
Choose it for: clean liquid to liquid duties, close temperature approach, heat recovery, hygienic service, tight plot space.
Air cooled, the fin fan
Finned tubes in a bundle with fans blowing ambient air across them. There is no second fluid to supply, which is the entire argument for it.
Where cooling water is scarce, expensive, or would require a cooling tower and its associated treatment, air cooling removes that infrastructure altogether. There is no fouling on the air side worth speaking of, no water treatment chemistry, and no discharge to manage. For refineries and plants in dry regions it is often the only sensible choice.
What you accept in return is that your cooling medium changes temperature with the weather. Design has to be based on a summer design air temperature, which means the equipment is oversized for most of the year and at its weakest on the hottest afternoon. The approach temperature achievable is much wider than with water, the plot area is large, and the fans consume power and make noise.
Choose it for: water scarce sites, large condensing and cooling duties, situations where avoiding a water system is worth the plot area.
Double pipe and hairpin
One pipe inside another. It is the simplest configuration there is, and it remains genuinely useful for small duties, for high pressure service where a small shell would be expensive, and for true counter current flow where you need the temperature profiles to cross.
Multitube hairpin units extend the concept upward in capacity. Beyond a modest duty, though, the surface area per unit of space becomes uneconomical and a shell and tube takes over.
Choose it for: small duties, high pressure, temperature cross situations, sampling and side stream service.
Spiral
Two long metal sheets rolled into a spiral, forming two curved passages that wind around a centre.
The geometry gives it one specific virtue: a single continuous passage per fluid, with curvature that promotes turbulence and a self cleaning scouring action. That makes it the answer for genuinely dirty service, for slurries, fibrous liquids and sludges that would block anything with a narrow gap or a tube bundle. It is a common sight in effluent treatment and pulp and paper.
It costs more per unit area than a shell and tube, and its pressure and temperature range is more limited.
Choose it for: slurries, fibrous fluids, sludge, high fouling duties where everything else blocks.
Compact and printed circuit
Printed circuit heat exchangers are made by chemically etching flow channels into flat metal plates and then diffusion bonding the stack into a solid block with no gaskets and no welds in the pressure boundary.
The result is extraordinary surface area density and a very high pressure capability in a small, heavy block. Where space or weight is genuinely constrained, offshore being the obvious case, they can be a fraction of the size of the alternative.
The channels are small, so the fluids must be scrupulously clean and filtration becomes part of the system design. They cannot be mechanically cleaned. They are expensive.
Plate fin exchangers, typically in aluminium, occupy the cryogenic niche: air separation, liquefied natural gas, and other very low temperature duties with multiple streams in one block.
Choose them for: extreme space or weight constraints, very high pressure, cryogenic multi stream duties, always with clean fluids.
The same duty, three different machines
An abstract comparison only goes so far. Take one duty and walk it through three answers.
Cooling a clean process liquid against cooling water, at moderate pressure, in a plant with plenty of water and limited maintenance staff. A gasketed plate unit wins comfortably. It is compact, it achieves a close approach so the water flow is modest, and when it eventually fouls it opens in an afternoon. The gaskets are a maintenance item, but the duty is clean and mild, so they will last.
Now change one variable. The same duty, but the process fluid carries fine solids. The plate unit is now the wrong answer, because the channels between plates are narrow and will block. Shell and tube with a removable bundle and square pitch becomes the sensible choice, accepting more space and a wider approach in exchange for a machine that can be cleaned.
Change a different variable instead. Clean fluid again, but the site has no reliable cooling water. Now neither answer works as stated, because both assume a water supply. An air cooled unit takes the plot space and the fan power and removes the water system entirely, and the design gets built around the hottest afternoon of the year rather than the average.
One duty, three correct answers, decided by cleanliness and by what utilities the site actually has. This is why a type cannot be chosen from a duty alone, and why any supplier who recommends a configuration before asking about fouling and utilities is recommending what they build rather than what you need.
Cost comparisons that mislead
Two habits distort heat exchanger comparisons more than any others.
The first is comparing on cost per square metre of surface area. Different types achieve very different heat transfer coefficients, so they need very different areas for the same duty. A plate unit with a high coefficient may need a third of the area of a shell and tube doing the same job, and comparing their prices per square metre makes the shell and tube look cheap while telling you nothing useful. Compare complete units against the same duty.
The second is ignoring what the exchanger obliges the rest of the plant to do. Pressure drop is paid for in pump power for the life of the installation. A wide approach temperature is paid for in additional heating or cooling duty elsewhere. An exchanger that needs a cooling tower commits you to the tower, its pumps, its water treatment and its blowdown. None of that appears on the equipment quotation and all of it appears in the plant operating cost.
The questions that actually decide it
- What are the design pressure and temperature? High values eliminate gasketed plate units immediately.
- How clean are both fluids? Fouling or solids rule out narrow channel designs and push you towards shell and tube or spiral.
- How close is the required temperature approach? A close approach favours plate and other counter current designs.
- Is cooling water available, and at what cost? This is the whole air cooled argument.
- What plot space and maintenance access exist?
- What material is needed? Exotic alloys are cheaper in a thin plate than in a thick shell, which sometimes reverses the usual cost order.
- What happens if the two fluids mix? Some configurations have an internal joint; others do not.
One number decides more than the rest
If you only have time to establish one thing before this conversation, establish how dirty the fluids are and how they foul.
Pressure and temperature are usually known precisely and they narrow the field quickly. Fouling is the variable that is usually estimated, is frequently wrong, and quietly determines whether the equipment you install will still be achieving its duty in five years. Get that right and most of the remaining decisions follow on their own.
