A maintenance team has an exchanger opened during a shutdown. The plan was to pull the bundle, clean the shell side, and put it back inside the window. Except this unit has a fixed tubesheet, which means the bundle does not come out at all, and it never did. Somebody chose the type twelve years ago on a capital cost comparison, and every shutdown since has paid for it.
Those three letters on an exchanger datasheet look like a filing code. They are not. They describe how the machine comes apart, and therefore what it will cost to own.
What the three letters actually encode
The TEMA designation gives an exchanger a three letter type. The first letter describes the front head, where the tube side fluid enters. The second describes the shell itself. The third describes the rear head, and it is the one that decides almost everything about maintenance.
So a unit described as BEM has a bonnet front head, a single pass shell and a fixed tubesheet at the back. An AES has a removable channel and cover at the front, a single pass shell and a floating head with a backing device at the rear. An AEU has a removable channel and a U tube bundle. Once you can read the third letter, you can tell from a datasheet whether a bundle is removable without asking anybody.
TEMA also assigns a class, R, C or B, which sets the mechanical standard: generally severe petroleum and related processing service, moderate commercial service, and chemical process service respectively. That class affects tolerances, corrosion allowances and construction detail, and it belongs in the enquiry rather than being left to the fabricator to assume.
The three rear head choices, plainly
Fixed tubesheet construction welds the tubesheets to the shell at both ends. It is the least expensive arrangement, it has no internal gasketed joint that could leak one fluid into the other, and it gives the most tubes in a given shell. The two limitations are absolute rather than negotiable. The shell side cannot be mechanically cleaned, because you cannot get inside it. And because the tubes and shell are rigidly connected, any significant difference in thermal expansion between them has to be absorbed, usually by an expansion bellows in the shell, which is then a component with its own fatigue life and its own failure mode.
U tube construction bends the tubes back on themselves so only one tubesheet is needed. Differential expansion stops being a problem, because each tube is free to move. The bundle can be withdrawn, so the shell side can be cleaned. The catch is on the tube side: you cannot pass a mechanical cleaning lance around a U bend. So U tube units suit clean tube side service with a fouling shell side, and they are a poor choice when the tube side is dirty.
Floating head construction leaves one tubesheet free to move inside the shell. It handles differential expansion, allows the bundle to be pulled, and permits mechanical cleaning of both sides. It is the most expensive, it uses some shell diameter for the floating head mechanism, and it introduces an internal gasketed joint. The pull through variant is easier to service; the split ring variant is more compact.
The kettle type deserves separate mention. Its enlarged shell creates a disengagement space above the bundle, which is what makes it the standard arrangement for reboilers and vaporisers where liquid boils and vapour must separate cleanly.
The expansion joint you inherit
Specify a fixed tubesheet exchanger with a large temperature difference between shell and tubes and you will usually be handed an expansion bellows in the shell. It is worth understanding what that component commits you to.
The bellows exists because tubes and shell are rigidly joined and will try to grow by different amounts. Without it, that difference appears as very high stress in the tube to tubesheet joints. With it, the shell can flex and absorb the movement.
What you have introduced is a thin walled component inside the pressure boundary that flexes every time the unit heats and cools. It has a fatigue life measured in cycles, so a plant that starts and stops frequently consumes that life faster than one running steadily. It is more vulnerable to corrosion than the shell around it, because it is thin. And it is difficult to inspect properly in place. Bellows failures are not common, but when they happen they are a pressure boundary failure rather than a performance problem.
The test that answers the question
Selecting a type looks complicated and mostly reduces to a short sequence of honest answers.
- Which side fouls? If the shell side fouls, you need a removable bundle, and fixed tubesheet is eliminated immediately.
- Does the tube side foul as well? If both sides foul, you are looking at a floating head, because a U tube cannot be cleaned through the bend.
- How large is the temperature difference between the two streams? A large difference with a fixed tubesheet means an expansion joint, and an expansion joint is a component you will inspect for the rest of the equipment life.
- What is the consequence of the two fluids mixing? An internal gasketed joint in a floating head is a possible cross contamination path. Where mixing is unacceptable, fixed or U tube construction avoids that joint entirely.
- Is there physical room to pull the bundle? A removable bundle is worthless without the clear length in front of the exchanger to withdraw it into.
That last point is missed with surprising regularity. A floating head unit installed in a pipe rack with no pull space has all of the cost of removability and none of the benefit.
The details that decide whether it lasts
Once the type is settled, a small number of construction details carry most of the reliability risk.
The tube to tubesheet joint is the most critical connection in the machine. Tubes may be expanded into the tubesheet, welded, or both, and for severe or hazardous service a strength weld combined with expansion is the conservative answer. Tubesheets on our work run past 600 mm thick, and on thick tubesheets the drilling accuracy and ligament quality between holes matter enormously.
Baffles do two jobs at once. They direct shell side flow across the bundle for heat transfer, and they support the tubes against sagging and vibration. Get the spacing wrong and you can induce flow induced vibration, which destroys tubes at the baffle edges. It is a genuine failure mode, it is predictable at design stage, and it is why baffle cut and spacing are not arbitrary.
Impingement protection at the shell inlet nozzle prevents a high velocity stream from eroding the first tubes it meets. It is a small plate that prevents an expensive and entirely avoidable failure.
Tube layout is a cleaning decision too
How the tubes are arranged in the tubesheet looks like a detail for the designer. It is actually the second maintenance decision in the exchanger, after the rear head type.
Tubes can be laid out on a triangular pitch or a square pitch. Triangular packs more tubes into a given shell diameter, which means more surface area for the same shell, better shell side heat transfer, and a more economical unit. It is the default for clean shell side service and there is no reason to avoid it there.
Square pitch packs fewer tubes and gives up some performance, and in exchange it leaves straight, continuous lanes between the tube rows. Those lanes are what a cleaning lance travels down. On a triangular layout there is no straight path through the bundle, so mechanical cleaning of the outside of the tubes is essentially impossible even after the bundle is withdrawn.
So the logic closes neatly. If the shell side fouls and you intend to clean it mechanically, you need a removable bundle and a square or rotated square pitch. Choosing a floating head for removability and then specifying triangular pitch inside it is a combination that appears more often than it should, and it delivers the cost of removability without the benefit.
Pass arrangement is the related choice on the tube side. Splitting the tube side into multiple passes raises velocity, which improves heat transfer and suppresses fouling, at the cost of pressure drop and of partition plates that consume tubesheet area. More passes also introduce temperature cross considerations that can require multiple shells in series.
Retube, repair or replace
Exchangers rarely fail as a whole. They fail tube by tube, and at some point a decision has to be made about an asset that is still mostly serviceable.
Plugging individual leaking tubes is the immediate response and it is entirely legitimate. Each plugged tube removes a little surface area and a little flow path, so performance drifts down as the count rises. Plants typically set a threshold, often a small percentage of the tube count, beyond which plugging stops being sensible.
Retubing replaces the bundle tubes while retaining the shell and heads. It is substantially cheaper than a new exchanger and it is the right answer when the shell is sound and the failure is confined to the tubes. It is also the moment to reconsider the tube material, because if the original tubes failed by corrosion, fitting the same grade again buys the same life.
Replacement becomes the answer when the shell itself is degraded, when the duty has changed enough that the geometry is wrong, or when the original type was the wrong choice and the plant has been paying for it at every shutdown.
What we build
Our shell and tube work covers the full TEMA range: fixed tubesheet, U tube, floating head and kettle reboilers, in process and refinery service, as condensers and reboilers, and in the compact and cryogenic variants where the duty calls for something other than a conventional bundle.
Shell diameters run beyond 5,000 mm, tube lengths past 20 m, with pressures from full vacuum through 700 bar and above and temperatures from minus 196 to 1,250 degrees Celsius, in carbon steel, stainless, duplex, nickel alloys, titanium, copper nickel and clad construction. Design work follows TEMA and HEI alongside the applicable pressure vessel code.
Where the thermal side of the decision matters more than the mechanical side, our note on buying heat exchangers by thermal duty covers the fouling and approach temperature trade in more depth.
Cheap to buy, expensive to open
A fixed tubesheet exchanger is genuinely the right answer in a great many services, and there is nothing wrong with specifying one. Clean cooling water on the shell side, a modest temperature difference, a service that has run for years without fouling: buy the fixed tubesheet and spend the difference elsewhere.
The failure is not choosing the cheap type. It is choosing it without asking the maintenance question, and discovering the answer during a shutdown when the only remaining options are chemical cleaning or replacement. Ask which side fouls. Everything else follows from that.
