A refinery pump starts running hotter than its datasheet says it should. Nobody changed the pump. Upstream, a cooler that was commissioned four years ago is no longer achieving its outlet temperature, because the tube side has filmed over with a deposit nobody budgeted for. The exchanger still works. It just does not do the job it was bought to do.

That is the most common failure in heat transfer equipment, and it is almost never a manufacturing defect. It is a design decision made against the wrong criterion. The exchanger was bought as a fabricated steel item, priced by weight, when it should have been bought as a thermal duty guaranteed over a service life.

Two designs live inside every exchanger

Any heat exchanger is really two designs stacked on top of each other, and they are done by different disciplines.

The thermal design decides whether the unit will transfer the heat you need. Surface area, tube count, tube length, baffle spacing, the number of passes, the velocities on both sides, the log mean temperature difference you have available to work with. This is process engineering, and it decides whether the equipment performs.

The mechanical design decides whether it survives. Shell thickness, tubesheet thickness, nozzle loads, differential expansion between shell and tubes, vibration, the pressure and temperature envelope. This is pressure equipment engineering, and it decides whether the equipment lasts.

A shop that can only do the second one is a fabricator. You can hand such a shop a fully engineered drawing and get a good product back. What you cannot do is hand them a duty and expect a recommendation. If your enquiry says "cool 180 cubic metres an hour of process fluid from 95 to 40 degrees" rather than specifying tube counts, you need a manufacturer with thermal engineers, not only welders.

Fouling is the number everyone guesses at

Here is the uncomfortable part. The single largest variable in a heat exchanger design is a number that is, in most projects, guessed.

Fouling resistance accounts for the layer of scale, biological growth, coke, polymer or general dirt that will accumulate on the surface during operation. It gets added to the design as an allowance, and it drives the surface area upward. Choose it too low and the exchanger meets its duty on day one and fails it by month eighteen. Choose it too high and you have bought and installed a unit substantially larger than the process needs, which then runs at low velocity, which encourages deposition, which is a genuinely self defeating loop.

Standard fouling factors exist and they are a reasonable starting point. They are not a substitute for knowing your own fluid. If you have run the same service on an existing plant, the operating history from that plant is worth more than any table. Tell your manufacturer what the last exchanger did, how often it was cleaned, and what came off it when it was opened. That single piece of operating history will improve the design more than any other input you can provide.

Velocity is the lever that answers fouling. Keep the tube side velocity high enough and much of the deposition problem manages itself. This is why the cheapest exchanger on a bid list is often the one that will cost the most to run: extra area at low cost per square metre, achieved by dropping velocity, is not a saving. It is a maintenance contract you have not read.

Approach temperature is where the energy bill hides

The gap between your outlet temperature and the temperature of the medium doing the work is called the approach. It is the quietest cost in the whole specification.

A tight approach means you recover more heat, which means less duty on the fired heater or the chiller downstream, which means a lower energy bill for as long as the plant runs. It also means more surface area, because squeezing the last few degrees out of a stream is disproportionately expensive in area. The relationship is not linear, and the last five degrees can cost more area than the first thirty.

That trade is a genuine engineering judgement and it belongs at the design stage, not in a value engineering exercise six weeks before delivery. A manufacturer worth working with will show you the curve, tell you where the sensible point is, and be honest that the tightest approach is not automatically the right one. Over a twenty year plant life, though, an approach chosen for capital cost alone is usually the more expensive decision.

The equipment range, and why breadth matters

Heat transfer is one field with many quite different machines in it, and a manufacturer confined to one of them will tend to recommend that one.

Our thermal work covers shell and tube exchangers across the TEMA configurations, including fixed tubesheet, U tube, floating head and kettle reboilers. It extends to process and refinery service exchangers, condensers and reboilers, air cooled units of the fin fan type, plate exchangers, and the specialised end of the field: spiral units, printed circuit exchangers, compact designs and cryogenic plate fin equipment.

The build envelope runs to shell diameters beyond 5,000 mm, tube lengths past 20 m, tubesheets thicker than 600 mm, pressures from full vacuum to 700 bar and above, and temperatures from minus 196 up to 1,250 degrees Celsius, in single pieces past 250 MT. Design follows TEMA and HEI alongside the ASME pressure code and the European standards where the project calls for them.

The reason to care about that breadth is not the catalogue. It is that a shop which builds air cooled units and plate exchangers as well as shell and tube can tell you when your duty does not want a shell and tube at all. That advice is worth more than a discount. If you want the comparison in full, our guide to the main types of heat exchangers sets out where each one earns its place.

The metal is a process decision

Tube material is chosen against the fluid, not against the budget, and cooling water is the classic trap. Chloride bearing water will find the weakness in a standard austenitic stainless steel given enough time and enough crevices. Duplex 2205 or super duplex 2507 costs more per tube and frequently costs less per decade.

We work across carbon and low temperature carbon steels, the 300 series stainless grades, duplex and super duplex, nickel alloys including Inconel, Monel, Hastelloy, Alloy 20 and Incoloy, titanium, zirconium, copper nickel and clad constructions where a corrosion resistant layer is bonded to a carbon steel backing. Clad is worth a specific mention, because on a large exchanger it delivers the corrosion performance of an expensive alloy with most of the strength coming from cheaper steel underneath.

What a thermal guarantee actually promises

Most exchanger offers include a performance guarantee, and it is worth reading closely, because the guarantees vary far more than the prices do.

A guarantee is given against a stated set of conditions: those flow rates, those inlet temperatures, those fluid properties, that fouling resistance. Change any of them in operation and the guarantee no longer applies, which is entirely reasonable and also means the guarantee is only as useful as the accuracy of the data you supplied. An exchanger guaranteed against optimistic fouling figures is guaranteed against a scenario that will not occur.

Ask two specific questions. First, how much thermal margin is included over the stated duty, and is it stated explicitly or hidden inside the fouling allowance? A designer who adds margin by inflating fouling is making the unit larger while telling you it is dirtier, and the two are not the same thing. Second, what is the guarantee actually against: achieving the outlet temperature, or achieving a heat transfer coefficient? Those are different commitments, and the first is the one that matters to your process.

Revamps follow different rules

A large share of heat exchanger work is not new plant at all. It is replacing or upgrading a unit inside an existing installation, and that is a more constrained problem than a blank sheet.

The nozzle positions are usually fixed, because the piping already exists and moving it is the expensive part. The foundations and supports are fixed. The available plot space is fixed, often by equipment installed on either side decades ago. And the maintenance access is whatever it happens to be.

Within those constraints the objective is normally to get more duty out of the same envelope, which pushes the design towards enhancement rather than enlargement: tighter baffle spacing, different tube layouts, enhanced or low finned tubing, or a change of tube material that permits thinner walls. Occasionally the honest answer is that the duty cannot be achieved in that envelope and the surrounding pipework has to move.

The most useful thing a plant can supply for a revamp is not a specification. It is the operating history of the unit being replaced, including what it actually achieved rather than what it was designed to achieve, and what was found inside it the last time it was opened.

What a good enquiry contains

  • Both stream compositions, flow rates, and inlet and outlet temperatures.
  • Operating and design pressures on shell and tube sides.
  • Allowable pressure drop for each stream, which constrains velocity and therefore the whole design.
  • Known fouling behaviour, ideally from an existing unit in the same service.
  • Cleaning expectations: whether the bundle must be mechanically cleanable, which points directly at the TEMA type.
  • Plot space and orientation limits, and the maintenance pull space available.

Allowable pressure drop is the one most often left blank, and it is the one that most constrains the designer. Without it, a manufacturer either asks or assumes, and assumptions cost money in both directions.

The twenty year bill

Price a heat exchanger on its purchase order and you are seeing a small fraction of what it will actually cost you. The rest arrives as pumping energy against pressure drop, as fired duty you failed to recover, as cleaning shutdowns, as bundle replacements, and occasionally as lost production while a unit that was never quite right is taken offline again.

None of that shows up on a comparative bid sheet. All of it shows up in the plant accounts. The exchanger that costs a little more and holds its duty for twenty years is not the expensive option, and any engineer who has signed off a cleaning shutdown already knows it.