- November 13, 2026
- Posted by: editor
- Category: RECD & Emission Control Devices
The short answer is yes, repeated failed regeneration can damage an RECD filter, but not always, and rarely in the dramatic way people imagine. A single interrupted cycle usually does nothing more than leave the filter a little dirtier than before. The harm comes from a pattern: the same failure repeating, each time leaving more soot behind, until the next successful burn becomes a risky one.
This article takes the question apart. Instead of asking “is it harmful?”, we ask which of five distinct damage mechanisms could be at work, what each one needs in order to cause harm, and how many failures it takes before the risk becomes real. Then we give a practical retry framework: how many attempts are reasonable, when to stop and what to do next.
Terms first. A failed regeneration here means a cycle that starts but does not complete: it times out, is aborted by the controller, is interrupted by a generator stop or load change, or ends with the differential pressure (ΔP) still high. It does not include a regeneration that never starts. That is a different problem, covered in our guide to RECD regeneration inhibit signals.
Why a failed cycle is not just “no cycle”
It is natural to assume that a failed regeneration leaves the filter exactly where it was. It does not, for three reasons.
- Part of the soot may already have burned. An interrupted cycle may remove soot from the hottest region and leave it in cooler regions, so the soot is redistributed rather than reduced.
- The filter has been heated and cooled. Every attempt is a thermal cycle, even a failed one.
- The controller’s idea of soot load may no longer match reality. If the controller estimates soot from ΔP and run-time, an incomplete burn can leave it believing the filter is cleaner than it is.
Published engineering literature on particulate filters notes that interrupted regeneration can leave soot concentrated in cooler zones and distort load estimates, and that repeated interruptions can leave regions holding more soot than the filter is designed to carry. That is the starting point for the damage map below.
The damage map: five mechanisms
Mechanism 1: Soot overload leading to a violent next burn
What happens. Each failed cycle leaves some soot behind, and the engine keeps adding more. If the filter reaches a soot load well above its design limit, the next successful ignition can burn very fast. A thick soot layer releases a lot of heat in a short time, so temperatures inside the filter rise steeply.
Why it matters. Ceramic filters tolerate high temperatures but not steep temperature differences between the centre and the edge, or between inlet and outlet. A rapid burn can produce thermal gradients that crack the substrate or, at worst, melt part of it. DieselNet’s overview of diesel particulate filter regeneration discusses why soot load at the start of regeneration governs the peak temperatures reached.
What it needs. A high accumulated soot load, then a sudden supply of oxygen and heat. A loaded set suddenly dropping to idle during a burn is a classic trigger, because exhaust flow falls and the heat in the bed is no longer carried away.
Risk after repeated failures: this is the main one. It grows with each failure that leaves soot behind.
Mechanism 2: Uneven soot distribution
What happens. Heat from the heater or exhaust is not delivered equally across the face. The centre of the filter often gets hot first, while the outer edge stays cooler. A cycle that stops early burns the centre and leaves the rim loaded.
Why it matters. Over several failures the rim can hold far more soot than the centre. When a full cycle finally runs, the rim burns later and violently, while the controller’s estimate, based on overall ΔP, underestimates the real load. The result is the same as mechanism 1, but harder to see in the data.
Clue in the log. ΔP after a failed cycle falls less than the burned fraction would suggest, or the ΔP-versus-load relationship changes shape.
Mechanism 3: Thermal cycling fatigue
What happens. Every attempt heats a heater element, the filter bed and the housing, then lets them cool. Repeated rapid cycling stresses ceramic, cement joints, mounting mats and welds.
Why it matters. Fatigue is cumulative. Ten gentle cycles are not the same as ten aborted cycles in which the heater is switched on and off at the limits of its range.
Risk after repeated failures: low per attempt, but real for heater elements, thermocouple joints and gaskets, which often fail before the filter itself does.
Mechanism 4: Hardening and sintering of ash and deposits
What happens. Ash from lubricating oil additives and fuel is not removed by regeneration. A regeneration that runs hot enough to burn soot but not long enough to finish may bake ash and oily residues into a harder layer at the inlet face or in the channels.
Why it matters. Hardened deposits are harder to clean, and some industry commentators note that repeated regenerations on an ash-heavy filter can convert loose ash into a more stubborn mass. Treat this as a plausible risk rather than a universal rule; it depends on the oil, the engine and the duty.
Risk after repeated failures: moderate on engines that burn a lot of oil. See our guide to soot versus ash in RECD regeneration.
Mechanism 5: Catalyst stress
What happens. If the RECD uses a catalyst, repeated high-temperature excursions or repeated exposure to unburnt fuel and oily deposits can accelerate catalyst ageing.
Why it matters. A less active catalyst lowers the NO2 supply that helps soot ignite at modest temperatures, so future regeneration needs more heat, which makes failures more likely. This is a feedback loop. Our article on catalyst ageing and regeneration behaviour explains the process, and catalyst poisoning in an RECD covers the chemical side.
Risk after repeated failures: slow, cumulative and easy to miss.
Summary table: what each failure type puts at risk
| Type of failed cycle | Mechanism most engaged | What is stressed | Typical warning in the log |
|---|---|---|---|
| Heater runs, bed never reaches ignition temperature | 3, 5 | Heater, catalyst | Long heater-on time, low peak temperature |
| Cycle aborted after soot ignites (load drop or stop) | 1, 2 | Substrate, rim soot | Sharp temperature spike then collapse; ΔP only partly recovered |
| Cycle times out with heater faulty or unpowered | 1 (indirect) | Filter via rising soot | ΔP climbing between cycles, heater current zero |
| Controller aborts on sensor fault | 1, 2 | Filter via unseen soot | Sensor faults, implausible readings, ΔP and temperature disagree |
| Generator stops mid-cycle repeatedly | 1, 2, 3 | Substrate, heater | Many short cycles, run hours low |
| Cycle completes, but ΔP does not return to baseline | 4 | Ash layer, deposits | End ΔP creeping upward cycle on cycle |
Notice that the most dangerous row is the second: a cycle that starts to burn and then loses its supply of flow or heat. That is when a loaded filter is most exposed.
How many failures is too many?
There is no standard number that applies to every device. Supplier guidance wins. But a sensible working framework can be built around two ideas: soot margin and cause removal.
Idea 1: soot margin
Think of the filter as having a soot capacity, with the supplier’s alarm level well below the point of danger. Every failed attempt eats into the margin because soot keeps arriving. If ΔP at the start of a retry is still below the level the supplier regards as a normal regeneration trigger, you have margin. If ΔP is above the alarm level, you do not.
Idea 2: cause removal
A retry is only worth doing if the reason for the last failure has been removed. Retrying without a change is hoping, not engineering. If the heater breaker tripped, the first retry should follow a proper check, not simply a reset.
A retry ladder you can adapt
The ladder below is a template. Ask your supplier to confirm or change each step for your device, and write the agreed version into the site procedure.
| Attempt | Condition to proceed | Action if it fails |
|---|---|---|
| 1st failure | ΔP below the supplier’s alarm level; cause obvious and fixable (for example, load dropped mid-cycle) | Log it; fix the cause; allow the next scheduled or manual attempt |
| 2nd failure in a row | ΔP still below alarm; no heater, sensor or supply fault | Do a checklist review (heater current, supply voltage, sensors, insulation, load during cycle); retry only under stable load |
| 3rd failure in a row | None. Stop automatic retries | Contact the supplier with the data pack; operate at higher steady load only if the supplier agrees; consider planned shutdown and inspection |
| Any failure with ΔP at or above alarm | None | Reduce load if safe, follow the alarm response matrix, call the supplier |
| Any failure with abnormal smoke, smell or temperature spike | None | Stop the set safely, inspect before restart |
For turning that last column into operator actions, see the alarm response matrix for operators.
What to check after every failed attempt
Make this a ten-minute routine, not an investigation.
- Read the controller’s reason code or message. Was it a time-out, abort, sensor fault, supply fault or generator stop? Our guide to interpreting an RECD regeneration completion message explains how to read them.
- Note ΔP and load before and after. The shape matters: a big drop followed by a rise suggests soot redistribution or a bypass.
- Check heater supply. Breaker status, supply voltage during the attempt, current draw.
- Check the generator’s behaviour. Did load fall, did the set trip, was there a transfer?
- Look at the exhaust. Visible smoke, odd smell, any unusual noise.
- Record the cycle in the event log. Use the format from our post on what to record in an RECD event log.
A worked scenario: three failures in a week
The following is an illustrative example, not a real case. A hospital runs a 320 kVA DG set with a heater-assisted RECD. During a week of grid instability the set starts eight times, mostly at light load for short runs.
- Monday, 11:20: regeneration starts at 52 mbar ΔP. The grid returns after 14 minutes and the set is stopped. Failure 1: interrupted, ΔP stays at 49 mbar.
- Tuesday, 02:10: regeneration starts at 56 mbar. The set is at 22% load. Peak bed temperature reaches only 430 °C after 40 minutes, then the controller times out. Failure 2.
- Wednesday, 23:45: the set is running at 18% load and regeneration is requested at 63 mbar. The operator notices the heater current reads zero. A breaker had tripped on Tuesday and was never reset or investigated. Failure 3, never actually heated.
What a careful team does. After Failure 1, they log it and note the cause (set stopped on grid return). After Failure 2, the checklist reveals low load as a probable cause, so they arrange a loaded run. After Failure 3, the ladder says stop. The maintenance lead calls the supplier, shares the log, and learns that the heater supply breaker tripped because of a loose terminal. The supplier supervises a regeneration with the set at stable higher load after the repair.
What would have happened otherwise. Soot would have kept building at light load while the heater sat dead. The eventual full burn might well have happened in conditions that nobody had planned: sudden load, sudden flow, too much soot. That is precisely the Mechanism 1 scenario.
In this example the failures were handled early, so the filter was most probably unharmed. The tripped breaker was the root cause, and none of the failures themselves caused damage. The value of the ladder is that it prevents a fourth, fifth and sixth failure from creating damage.
Signs that damage may already have occurred
If you suspect harm has already been done, these are the typical symptoms. None is proof, and some have other causes, so inspection by the supplier is the only confirmation.
- ΔP no longer recovers to the earlier baseline after complete cycles, even after cleaning.
- ΔP unusually low for the load, or falling suddenly without regeneration, which may indicate a crack, a bypass or a failed seal.
- Visible smoke increases after the filter. See our article on using opacity and particulate readings to separate causes.
- Hot spots or discolouration on the housing.
- Rattling or loose debris noises on removal, suggesting a broken substrate or mat.
- Sudden temperature spikes in earlier cycle logs that no one acted on.
A cracked or melted substrate usually needs replacement of the filter core. Our post on RECD filter core versus complete assembly replacement explains what to ask for.
Preventing the loop
Most failed-regeneration loops start with a small number of avoidable causes.
- Protect the cycle. Tell operators not to stop the set or shed load while a regeneration is in progress, unless safety requires it. See regeneration while supplying building loads.
- Power the heater properly. A heater on a non-essential board fails exactly when needed. See controller and heater power supply requirements.
- Avoid prolonged light-load running. Our guide to choosing an RECD for a generator running at low load explains why this builds soot.
- Keep sensors honest. A drifting ΔP or temperature sensor makes the controller blind. See separating a sensor fault from an RECD performance problem.
- Watch the trend, not just the alarm. A rising duration or falling recovery is an early warning. See our article on what a rising regeneration duration tells you.
- Keep a written retry procedure and make sure night-shift staff know it.
A soot-margin calculation on the back of an envelope
The idea of “soot margin” becomes more useful when you put numbers on it. The figures below are invented for illustration; your supplier’s limits will differ.
Suppose the supplier states that a regeneration normally triggers at 55 mbar at the reference load, that the alarm level is 80 mbar at the same load and that the clean baseline is 30 mbar. The usable band between baseline and alarm is therefore 50 mbar.
| Situation | ΔP at reference load | Position in the band | Margin left |
|---|---|---|---|
| Normal trigger point | 55 mbar | 50% used | 25 mbar |
| After first failed cycle (partial recovery) | 58 mbar | 56% used | 22 mbar |
| After second failed cycle | 64 mbar | 68% used | 16 mbar |
| After third failed cycle | 72 mbar | 84% used | 8 mbar |
Notice how quickly the margin shrinks. Each failed attempt leaves a few extra millibars behind, and the engine adds more while the set keeps running. By the third failure in this example the filter has used more than four-fifths of its usable band, and the remaining eight millibars could disappear within a single long run at light load.
Two points follow. First, the ladder’s “stop after three” rule is not arbitrary: it is roughly the point at which the margin becomes too thin to gamble with. Second, margin is not the same on every site. A set that runs at very light load for hours adds soot faster than one that runs at 60%, so the same number of failures can use up the margin sooner. Calculate your own band with your supplier’s figures and write it into the procedure.
Remember that the reference load matters. A reading of 72 mbar at 25% load is not the same as 72 mbar at 70% load, because flow changes ΔP. Our article on comparing RECD pressure readings at different generator loads explains how to keep readings comparable.
A note on standby sets
Standby DG sets add a twist. They may run for only a few hours a year, so there may be few regeneration attempts at all, and a failure in a rare outage could be the first real cycle in months. For such sets, the main protection is preparation: run loaded tests so that the filter is not left heavily loaded, check the heater supply before the monsoon and winter and make sure operators know the retry rule before an emergency, not during one. See is an RECD suitable for a DG set that runs only occasionally.
What not to do
- Do not hammer the manual regeneration button hoping the next try works.
- Do not bypass inhibit signals or safety interlocks to force a cycle.
- Do not reset a tripped heater breaker repeatedly without finding why it tripped.
- Do not run at very high load purely to burn soot if the generator or the building cannot safely carry it.
- Do not ignore ΔP above the alarm level because the set “seems fine”.
- Do not disassemble or clean the filter yourself unless the supplier has told you how; handling cracked ceramic and soot has its own hazards.
FAQ
Can one failed regeneration damage the filter?
It is unlikely. One failed cycle normally leaves the filter loaded a little more than before. Damage becomes a concern when failures repeat, soot accumulates beyond the supplier’s limits, and a later burn runs under poor conditions.
How many failed regenerations are too many?
Ask your supplier for a limit for your device. As a working rule, treat two consecutive failures as a prompt for a checklist review, and three as a prompt to stop retrying and escalate. Any failure with ΔP at or above the alarm level should be escalated at once.
Does an interrupted regeneration harm the engine?
The engine does not feel the interruption directly, but the filter stays loaded, and a loaded filter raises exhaust back-pressure. Sustained high back-pressure can affect the engine. See our article on how exhaust backpressure affects a DG set.
Is a failed regeneration worse than no regeneration?
Not necessarily. A failed cycle can redistribute soot and stress components, but a filter that is never regenerated fills until back-pressure forces action. The goal is not to avoid attempts, but to avoid blind repetition.
Can a cracked filter be repaired?
Usually the filter core is replaced rather than repaired. Ask the supplier for an inspection to confirm the cause before ordering parts, and use the failure history to decide what else needs attention.
Does a catalyst help protect against failed regeneration?
A healthy catalyst lets soot ignite at lower temperatures, which makes cycles more likely to succeed under modest loads. That benefit fades as the catalyst ages, so monitor trends over years as well as weeks.
Bringing it together
Repeated failed regeneration can damage an RECD filter, mainly by letting soot build beyond design limits and then burning it under poor conditions, and secondarily by thermal fatigue, ash hardening and catalyst stress. The cure is not fear but discipline: understand why each cycle failed, remove that cause, use a written retry ladder, stop after the agreed number of failures and send the supplier a clean log.
If you would like a second opinion on a regeneration history, or want to understand which devices suit your duty cycle, see our range of retrofit emission control devices, review our DG set emission regulations overview or contact the Aceget team.
[…] Regeneration discipline. Avoid runaway burns by not letting soot load climb to extremes. See can repeated failed regeneration damage an RECD filter. […]