Heat exchanger repair questions

Usually, where the shell is sound, because the shell, channels, nozzles and supports are the expensive fabrication and they are all being kept. It stops being true once the shell itself needs significant work, and the honest answer needs the unit inspected rather than guessed from a drawing.

There is no universal figure, because it depends on the margin the exchanger was designed with. A unit specified with generous surface area tolerates a surprising number of plugged tubes. One running close to its duty does not. The number that matters is the thermal performance you are still getting, not the percentage plugged.

Yes, and where only a handful have failed that is the sensible route. It stops being sensible when the failures are spread across the bundle, because each new tube then sits among tubes that are close to failing themselves.

No, and matching it is sometimes the mistake. If the original material was wrong for the water chemistry or process fluid, a retube is the natural point to change it. The change has to be checked against the design conditions and the tubesheet, since not every material pairs safely with every tubesheet.

A correctly fitted plug in a sound tube end is a permanent repair, not a temporary one. What is temporary is the situation around it: the reason that tube failed is usually still acting on the tubes next to it.

Yes. Plugging is done through the open channel with the bundle in place, which is why it suits a mid campaign failure. It is also why plugging tells you nothing about the condition of the rest of the bundle.

Yes. A tube sealed at one end still fills, and stagnant fluid in a dead tube corrodes it from the inside and can put the tubes around it at risk.

The frame and plates can last for decades on a mild duty, while the gaskets are a wearing part measured in years and driven by temperature and fluid rather than age. Plate condition at the last strip down predicts remaining life; the date on the nameplate does not.

They do, in three ways. Gaskets harden and leak outward, plates crack or perforate and let the two fluids mix, and fouling closes the channels and takes the duty away. The first two are repairs, the third is a clean.

The gaskets set the limits. They cap the pressure and temperature the unit can take and they are a consumable, so a gasketed plate exchanger needs opening periodically in a way a welded unit does not. In exchange it is compact, efficient at a close approach temperature, and can be opened for inspection.

Not on a plate. A cracked or perforated plate is replaced, because a patched plate leaves the cross contamination that brought the unit in. On tubular units a failed tube can be plugged or sleeved, which is a different thing from patching.

Most gasketed plate and tubular units can be repaired, and the question is whether it is economic. What decides that is the condition of the frame or shell, since that is the expensive part and the one that is being kept.

No. A cracked plate is replaced. A cracked plate is replaced. Dye penetrant testing finds the crack, and the plate comes out of the pack.

Tightening a leaking pack is a short term fix that compresses gaskets already at the end of their life, and every tightening takes the pack closer to its minimum dimension. Below that dimension it cannot seal at all, and the plates themselves start to deform.

All of them. A gasket set is renewed complete, because replacing one seal in an aged set brings the unit back within months.

Clean in place is normal for routine duty and will hold fouling back. It does not remove hard scale or established biofilm, and it cannot inspect a plate for cracks, which is the part that stops cross contamination.

It depends on what the strip down finds, on tube and plate lead times, and on whether the pressure envelope needs weld repair. Any duration given before the unit is opened is an estimate. What can be fixed in advance is the inspection stage, and that is the point at which a real programme becomes possible.

For most duties, a unit with a sound pressure envelope, new tubes or plates, new gaskets and a current pressure test is fit for its design conditions. What it does not do is change the design, so a unit that was undersized stays undersized.

Cleaning, plugging and some plate work can be. A full strip down, NDT and weld repair generally needs a workshop, and moving the unit is usually cheaper than setting up equivalent facilities on site.

Design pressure is the figure the exchanger was designed to, and maximum allowable working pressure is the highest pressure the finished unit is permitted to operate at in its current condition, which can be lower once corrosion allowance has been consumed. Both are on the nameplate and both matter, because test pressure is derived from them rather than from the operating pressure.

From the design pressure and the code the unit was built to, corrected for the temperature and the material rather than taken as a flat multiple. Testing to an assumed figure risks damaging equipment that is otherwise serviceable, which is why an illegible nameplate is a real obstacle.

Often, yes, if it can be isolated and drained and the channel can be opened. What in situ testing cannot do is inspect the bundle, so it will tell you the unit is leaking without telling you the state of the tubes that have not failed yet.

Usually because the leak only opens under thermal load or under differential pressure the test did not reproduce, or because it is small enough to sit below what a hydrostatic test can resolve. Helium testing exists for exactly that case.

From the design pressure on the nameplate and the code the unit was built to, not from the operating pressure. Testing to an assumed figure risks damaging equipment that is otherwise serviceable.

Set by the unit's own design conditions and by the gasket material, not by a general figure. The nameplate carries the design pressure and temperature, and the gasket specification narrows it further, which is why the plate is the first thing asked for.

Fouling and corrosion account for most of it. Fouling costs performance first and can lead to under-deposit corrosion; corrosion thins the metal separating the two fluids until it leaks. Mechanical causes such as vibration and thermal cycling account for much of the rest.

The written scheme of examination sets the formal interval, and the competent person who drew it up decides that. Beyond compliance, the useful interval is the one that catches a trend before it becomes a failure, and that depends on how aggressive the duty is.

No. Tube inspection is done through the open channel with the bundle in place. Removing it lets the shell side and the outside of the tubes be examined too, which matters where the corrosion is coming from the shell side.

Which tubes have lost wall thickness and how much, mapped to the tubesheet. Set against the previous inspection, that is what turns a guess about remaining life into a plan for the next shutdown.

On condition rather than on a calendar. Monitoring approach temperature and pressure drop tells you when fouling is costing more than the clean would, which is a better trigger than a fixed interval that is either too often or too late.

The deposit, the tube material and the effluent route decide it. Chemical suits scale and reaches the whole surface; jetting suits hard deposits and avoids chemistry altogether. Sometimes it is both, in that order.

Yes, if the method does not match the material. Wrong chemistry attacks the tube wall, and excessive jetting pressure can damage thin walled or already thinned tubes. Knowing the condition of the bundle first is what prevents that.

No. Cleaning restores performance. A leak is a tube, plate or joint failure and needs a repair, though a clean is often what makes the failure visible in the first place.

The pack is recorded and measured before anything is disturbed, then stripped, the plates cleaned and inspected, gaskets replaced as a set, and the pack rebuilt to its recorded dimension and pressure tested. The plate inspection is what separates a service from a clean.

The unit is isolated, drained and allowed to cool, and the tightened pack dimension and plate order are recorded. A pack reassembled in the wrong order does not work, and the sequence cannot be reconstructed afterwards.

Comparatively, yes, because the pack is built to be opened, the plates come out individually and capacity can be changed by adding or removing plates. What that ease depends on is a recorded plate order and gaskets that are still available for the model.

On condition rather than on a fixed interval. Monitor approach temperature and pressure drop: when performance has drifted enough to cost more than the service would, it is due. Duties that foul hard need it far more often than the same unit on clean water.

A repair fixes something that has failed. A service is done before it fails, and its main output is knowing the condition of every plate. The two overlap when a service finds a cracked plate, which is common enough that it should not be a surprise.

The pack can be opened and regasketed in place where there is room to work. What cannot be done on site is proper tank cleaning and inspection under ultraviolet light, so an in situ service leaves the plate condition unproven.

It depends on the deposit and the plate material. Scale is usually removed with an acidic cleaner, organic and protein deposits with an alkaline one, and a unit on a food duty has to be left fit to return to product service. Identifying the deposit comes first, because the wrong chemistry attacks the plate.

Yes. Cleaning in place circulates chemistry through the assembled unit, and a strip clean removes the plates and cleans them individually in heated tanks. Only the second one lets the plates be inspected, which is why a strip clean answers both questions at once.

Easier than most tubular units, because the pack is designed to be opened. Gasketed plates come out individually. Brazed units cannot be opened at all and are cleaned chemically or not at all.

For routine maintenance on a duty that does not foul hard, often yes. It is not enough when the unit is already well down on performance, when the fouling is hard scale or established biofilm, or when you need to know the condition of the plates. CIP cannot inspect anything.

When the numbers say so. Trending approach temperature and pressure drop gives a genuine trigger; a fixed interval tends to be either too early or too late.

Yes, if the method does not match the material. Wrong chemistry attacks the plate, and aggressive mechanical cleaning damages the pressing that does the heat transfer. It is one reason a workshop tank clean is gentler than field methods, not harsher.

No. Cleaning restores performance. A leak is a gasket or a cracked plate, and it needs a repair, though a clean is very often what makes the crack visible in the first place.

It depends far more on temperature and fluid than on age. A gasket in a hot or aggressive duty ages quickly; the same gasket on mild heating water lasts a long time. Condition when the pack is opened tells you more than the date it was fitted.

Clip-on gaskets are faster to fit and can be changed on site, which suits units that come apart often. Glued gaskets, cured in an oven under compression, generally hold better in demanding duties. The plate model usually decides it for you.

It sometimes stops a weep for a while, and every time it does, the pack gets closer to its minimum dimension. Below that dimension the pack cannot seal at all and the plates begin to deform, so tightening only postpones a larger job.

The nameplate, what has failed and how you know, and whether the plant is down now or still running. Those three answers are usually enough to decide the approach before anyone attends.

Sometimes. Tube plugging, gasket work and cleaning can often be done in place where there is access. A strip down, inspection and weld repair generally need a workshop.

It can be. A correctly fitted tube plug is permanent. What is temporary is the situation around it, because whatever caused that tube to fail is usually still acting on the ones next to it.

Partly. The numbers and a pressure test narrow it down a long way and will often identify the type of fault. Readings and a pressure test cannot establish the condition of surfaces nobody has looked at.

Usually because the repair addressed the failure and not the cause. If water treatment, flow velocity or a temperature profile drove the failure, the same failure returns on the same schedule.

It depends on the unit and where it is, but unlike the repair it can usually be established up front, which is why an assessment is quoted before the work rather than alongside it.

A repair fixes something that has failed. A service is done before it fails, and its main output is knowing the condition of the unit. The two overlap when a service finds a fault, which is common enough that it should not be a surprise.

On condition. Trending approach temperature and pressure drop gives a real trigger, and duties that foul hard need attention far more often than the same unit on clean water.

Where there is a spare or a bypass, yes. Without one, the duty is off line for as long as the work takes, which is usually what decides the approach.

There is no single figure. A unit on clean water at modest temperatures can run for decades; the same unit on an aggressive duty may not. Condition, measured by inspection, tells you far more than age does.

When the pressure envelope itself is failing, when enough plates or tubes have failed that the repair approaches the cost of a unit, or when parts are no longer obtainable. All three are inspection findings rather than judgements from the outside.

Usually yes, provided the duty, the connections and the space all work. It is worth confirming the replacement suits the duty the plant has now rather than the one it had when the original was specified.

When the pressure envelope is the problem rather than the internals. Tubes and plates are renewable; a corroded shell, frame or channel generally is not. Parts that can no longer be obtained end a unit's life just as effectively, which is why availability is checked at the same time as condition.

Often, provided the duty, the connections and the space all work. What matters is that the new unit does the job the plant needs now rather than matching a label.

The old unit is removed and either disposed of or recovered. Where it has any remaining value that is worth establishing before it is scrapped.

Inlet and outlet temperatures on both sides, pressure drop across each side, and flow. Together they give the approach temperature and the trend, which is what shows condition without opening the unit.

The design figure for that unit, which is on its datasheet. There is no general number worth quoting, because it depends on the duty, the geometry and the flow. What matters is the change from the clean condition.

The unit's own design decides it, which is why the datasheet is the benchmark and a general rule is not. What matters is the change from where it started.

Yes, and in most cases that is the point. The readings are taken in normal operation, since the question is what the unit is delivering in service.

Testing tells you how much duty you are losing and how quickly. What the deposit or damage actually is still needs the unit opened.

A repair addresses the fault that stopped the unit. A refurbishment takes the whole unit back to a known condition, which is why it is usually planned rather than reactive.

For most duties, a unit with sound plates, new gaskets and a current pressure test is fit for its design conditions. What refurbishment does not do is change the design, so a unit that was undersized stays undersized.

Not fully. Tank cleaning, inspection under ultraviolet light and oven cured gaskets all need a workshop, and moving the unit is usually cheaper than trying to recreate that in a plant room.

No. Cracks run through the pressing and cannot be sealed, so the plate is withdrawn from the pack and a replacement hung in its place.

A gasket failure normally leaks to the outside. Fluids mixing between the circuits normally means a plate. A pressure test on each side separately settles it.

Yes, on any duty where cross contamination matters. One pinhole in one plate is enough, and it will not be the plate you decided to sample.

The inside of the tubes, yes, through the open channel. The outside of the tubes needs the bundle withdrawn, and on a fixed tubesheet unit that is not possible at all.

It depends on the deposit, the tube material and what can be done with the effluent. Chemical reaches the whole surface, jetting handles hard deposits without chemistry. Sometimes it is both, in that order.

Yes, if the method does not match the material and condition. Wrong chemistry attacks the wall, and excessive jetting pressure can damage tubes that have already thinned.

The written scheme of examination sets the formal interval. Beyond compliance, the useful interval is the one that catches a trend before it becomes a failure, which depends on how aggressive the duty is.

Not for tube side work. It does for shell side cleaning and for inspecting the outside of the tubes, and on a fixed tubesheet unit it cannot come out at all.

Which tubes have lost wall and how much, mapped to the tubesheet, plus the condition of the tubesheet, baffles and shell. Against the last report, that turns a guess about remaining life into a plan.

Usually, depending on where the leak is and what the unit is made of. Tube and plate failures are often repairable; corrosion of the shell is a different question and needs assessment.

Most often fouling on one of the surfaces, which insulates it. Flow problems and refrigerant charge are the other common causes, and the numbers separate them.

Sometimes. Circulated chemistry clears soft deposit on an assembled unit, and anything harder, or any need to see the surfaces, means opening it.

The two circuits are connected, through a failed tube, plate or joint. The unit needs to come off line, and the source is found by pressure testing each side separately.

If the problem is fouling and lost cooling, yes, cleaning is usually the answer. If the two fluids are mixing, cleaning does not touch the cause.

The cooler type and the strip down findings decide it. What can be established early is the assessment, and from there a real programme becomes possible.

The construction decides it, along with where the failure is. Semi-welded units can be opened on the gasketed side and repaired much like a gasketed pack. A fully welded pack cannot be opened, so the options are cleaning, external repair or replacement.

Chemically, circulated through the unit, because the plates cannot be removed. The chemistry is chosen for the deposit and for the plate material.

One where plates are welded in pairs and the pairs are sealed to each other with gaskets. It suits duties where one fluid is aggressive to gaskets and the other is not.

No. The plates are brazed into a sealed block with no gaskets and no way to open the pack, so an internal failure cannot be reached. A failed brazed unit is replaced.

Yes, chemically, circulated through it while the unit is still sound. That is worth doing when performance has dropped, and it is not a fix for a unit that has already failed internally.

A gasketed unit has a frame, tie bars and visible plate edges with gaskets between them. A brazed unit is a solid block of plates with connections, and nothing to undo.

A regasket can often be done in place where there is access. Proper cleaning and plate inspection need a workshop, so an on site repair returns the unit to service without establishing the condition of the plates.

Usually not. Fluids crossing between circuits normally means a cracked or perforated plate, and no amount of tightening or regasketing addresses that.

Sometimes it stops a weep, and every time it does the pack gets closer to the dimension below which it cannot seal at all, at which point the plates begin to deform.

Cleaning, tube plugging and gasket work often can, where access and permits allow. Retubing, weld repair and proper inspection generally need a workshop.

That is how most of it is done. The useful step before an outage is establishing condition while the plant runs, so the scope is known before the window opens rather than discovered inside it.

Where a unit sits under one, what the competent person requires is agreed before the work rather than after, because it constrains the method as well as the paperwork.

Usually, and on occupied buildings it often has to be. It is worth saying so at the enquiry, because it changes the plan rather than the price alone.

Send the details and it can be assessed straight away. What can be done quickly depends on whether the fault is the exchanger, the calorifier or the system around it.

To stored hot water systems, yes, and it shapes both the work and the records. The water safety plan is the reference point rather than a general rule.

Not always. A failed steam trap, a flooded condensate return or a control valve problem all look the same from the temperature reading. The numbers and a check of the steam side usually separate them.

Often, depending on whether the failure is the coil or the vessel. A coil can be repaired or replaced. Corrosion of the vessel is a different question, because that is the pressure envelope.

To steam plant, yes. The written scheme of examination sets what the competent person requires, and that is agreed before the work rather than after.

The two sides are connected, through a cracked plate, a failed tube or a joint. The unit needs to come off line, and pressure testing each side separately locates it.

That is part of the specification for the work rather than an afterthought. The cleaning method, the chemistry, the rinse and the gasket material are all chosen with the return to product service in mind.

Only if they pass inspection. Plates are checked for cracks and pinholes, and any that fail are replaced rather than refitted.

No. This is industrial and commercial plant: plate packs, shell and tube units, condensers, calorifiers and process coolers. A boiler heat exchanger in a house is a different trade, and a heating engineer is the right person for it.

There is no useful general figure, because a gasket set and a retube are different orders of magnitude on the same unit. What can be established early is the cost of the assessment, and from there a firm quotation follows.

The pressure retaining parts are precision fabrications in materials chosen for the duty, often stainless, duplex or titanium, and they are built and certified to pressure equipment standards. That fabrication cost is also why repairing a sound shell or frame usually beats replacing the whole unit.

Usually repair, because the shell or frame is the expensive fabrication and a repair keeps it. The arithmetic turns once that part is itself corroded, or once the plates or tubes failing inspection approach the cost of a new unit.

Duration follows the scope, and the scope follows the assessment. Parts availability is often the constraint rather than the labour, particularly on older models. Any duration offered before the unit is opened is an estimate.

A photograph of the nameplate narrows it down a long way and is worth sending. A photograph cannot replace the assessment, because cost is driven by what the strip down finds, not by what the outside looks like.

Weld repair to a shell, channel or nozzle is routine work, carried out to a qualified procedure, and on a pressure system it has to leave the unit certifiable against its written scheme of examination. Plates in a plate pack are not weld repaired, they are replaced.

A tube failing completely rather than developing a small leak, which puts the two circuits in direct communication at full pressure. Pressure relief and the design conditions on the low pressure side exist to cover that failure mode.

The tube side and the shell side. Fluid inside the tubes is on the tube side, fluid around them inside the shell is on the shell side, and they are tested separately for exactly that reason.

Where the failures are, and what condition the shell is in. A few failures in a sound bundle means plugging. General wall loss across the bundle with a sound shell means retubing. A corroded shell usually means replacement, because you would be spending the cost of a retube on a unit that fails elsewhere shortly afterwards.

Plugging, cleaning and inspection are all done in place through the open channel. Retubing, weld repair and anything needing the bundle out generally means a workshop, and moving the unit is usually cheaper than recreating a workshop around it.

How the bundle is fixed at each end, and whether it can come out. A fixed tubesheet unit is welded at both ends and the bundle stays put. U-tube and floating head designs allow for thermal expansion and let the bundle be withdrawn, which makes them considerably easier to clean and repair.

On air-cooled units, usually the fin side. Fouling between the fins, flattened fins from over-pressure washing, or fin bond loss will all cut the duty without any symptom that shows up as a fault. Fan and drive problems produce the same picture, so both are worth checking.

Often yes, depending on the bundle construction and access. Where enough have failed, or where the fin bond has gone generally, replacing the bundle usually beats replacing tubes one at a time.

Through the bundle in the direction that carries debris out rather than further in, at a pressure the fin material tolerates. Flattened fins block airflow, and that damage is not easily undone.

Coil failure with a sound vessel is usually a repair. A corroded shell is usually a replacement, because the vessel is the pressure envelope and patching it is rarely the right answer. Access often decides it in practice: some plantrooms were built around the calorifier.

Most often scale on the heating coil, insulating it from the stored water. Scale looks like a boiler or a sizing problem and is neither. The giveaway is that performance has fallen gradually in a hard water area.

The written scheme of examination sets the interval and the competent person who drew it up decides that. Water hygiene regimes usually drive their own inspection frequency as well, and the two are best planned together rather than separately.

Charge air temperature after the cooler, compared against the cooling water inlet temperature. As the difference widens, the cooler is fouling. Exhaust temperatures climbing across every cylinder points the same way, and both show up well before anything fails.

Most likely a tube leak putting cooling water into the charge air, and it needs investigating before the engine runs on. Condensation in humid conditions produces the same symptom, so a pressure test settles which of the two it is.

In service washing systems help keep the air side down between overhauls, and they are worth using. In situ methods do not replace removing the cooler and immersion cleaning it, because they do not reach the whole fin block.

Helium leak detection is the normal method. Helium is applied to the suspect area while the condenser is under vacuum, and it is drawn in and detected in the extracted gas. With thousands of tubes, elimination by plugging is not a realistic alternative.

That pattern usually means air ingress rather than a tube leak: air being drawn in through joints, glands or the shell. Air ingress degrades vacuum and costs output, and it is found by a different search than a tube leak.

The margin the condenser was designed with decides it, together with what plugging does to the vacuum, which is what actually costs output. The measure that matters is the vacuum you are still achieving, not a percentage of tubes.

CIP keeps the product side clean. Cleaning in place removes neither hard mineral deposit nor the contents of a partly blocked channel, and it inspects no plate for cracks. A service is about proving plate integrity, which is a different job from cleaning and is the one that protects the product.

Usually through product or utility contamination, which is the expensive way to find out. Dye penetrant inspection with the plates out finds cracks and pinholes before they start passing, which is why a strip service is worth more here than in most sectors.

Deposit that CIP is not fully removing, so each run starts from a slightly worse condition than the last. A shortening cycle is a signal to strip and clean properly, not to shorten the cycle again, and product temperature is worth reviewings and flow at the same time.

Usually fouling or partial blockage in the plate pack, particularly if pressure drop has risen at the same time. Worth ruling out coolant temperature and flow first, since a glycol problem produces the same symptom and is quicker to check.

Circulated caustic and acid cycles handle routine fouling and should be the normal regime. Circulated cleans do not clear a channel packed with trub or hop material, and they cannot establish whether a plate has cracked.

Between brews in a brewery, and in the silent season in a distillery. Booking it against the production calendar rather than against a maintenance interval is what makes the difference between planned work and an emergency.

Between turnarounds, precisely because the turnaround is years away. An inspection is what turns a guess into a scope with the right materials ordered, and finding the problem during the shutdown is the expensive way to discover it.

Because deposits shield the metal underneath from corrosion inhibitor and hold water and contaminants against the surface. Corrosion under deposit is a common failure mechanism, so a fouled exchanger is degrading as well as underperforming.

Tube material lead time, the number of units and whether the bundle can be withdrawn off the critical path all decide it. It is a planning question more than a technical one, which is why the inspection that sets the scope matters so much.

Treat it as a cooler tube leak until proven otherwise, and investigate before running on, because the water ends up in the cylinders. Condensation in humid conditions can produce the same sign, and a pressure test of both circuits separately is what tells you which it is.

More often than an equivalent shore unit, and the interval should follow the trade rather than a generic schedule. Seawater temperature, silt, growth and lay-up periods all change the rate, so condition monitoring beats a fixed calendar here.

Opening a pack and replacing gaskets is done on board where there is room and the spares are held. Proper tank cleaning and inspection under ultraviolet light needs a workshop, so on board work leaves plate condition unproven.

Treat it as a condenser tube leak and locate it. Because the shell side is under vacuum, cooling water is being drawn into the condensate, and the chemistry limits on the circuit mean even a small in-leak is worth stopping. Helium leak detection is the usual way to find which tube.

That pattern points at air ingress rather than a tube leak: air drawn in through joints, glands or the shell. Air ingress costs output in the same way as a tube leak and is found by a different search, so establishing which one you have comes before opening anything.

There is no useful percentage. Plugging reduces the surface available and therefore the vacuum you can hold, and it is the vacuum that determines lost output. The number to watch is the vacuum you are still achieving.