What Does a Rising Regeneration Duration Tell You About an RECD?

A rising RECD regeneration duration is one of the quietest warnings a retrofit emission control device can give. Nothing trips. No lamp comes on. The cycle still finishes, only it needs forty minutes where it once needed twenty-five. Because the device “still works”, the trend tends to sit unread in a controller log until the day a cycle fails to finish at all.

This article treats regeneration duration the way a doctor treats resting heart rate: a single reading means little, a drifting trend means a great deal, and the useful skill is knowing which underlying cause produces which kind of drift. By the end you will have a simple three-factor model for any lengthening cycle, a table that links each pattern to likely causes, a worked twelve-month example, and a short list of what to record before you ring your supplier.

A note on scope. Regeneration is the process of burning soot out of a filter so that back-pressure falls again. RECD designs differ: some rely on passive regeneration through a catalyst, some add an electric heater, some use fuel assistance, and some ask the operator to run the set at higher load for a while. Your own design decides which numbers your controller logs. If a term below does not match your device, ask the supplier what the equivalent is.

First, what “duration” actually measures

Before reading a trend, be sure what the number means. Controllers define duration in different ways.

  • Heater-on time. The minutes during which the heater or burner was energised. This ignores the time soot continues to burn after heat input stops.
  • Cycle time, start to end flag. From the moment the controller declares regeneration started to the moment it declares it finished. This includes cool-down or hold periods on some designs.
  • Time above a temperature threshold. The minutes during which the inlet or bed temperature stayed above a set point.
  • Time to ΔP target. The minutes needed for differential pressure to fall to a recovery value.

These four can move independently. A cycle with a longer heater-on time and a shorter hold phase may show a flat “cycle time”. Ask your supplier which definition applies, write it on the log sheet, and never mix numbers from different definitions in one trend line.

Also record the conditions that produced each cycle. A cycle run at 70% load in March and one run at 30% load in September are not the same experiment, so duration alone cannot be compared across them. For the background on how regeneration is triggered and ended, see our guide to active versus passive RECD regeneration.

The three-factor model: soot in, heat in, heat out

Every regeneration is a race between three things. If a cycle takes longer, at least one of them has shifted.

Factor 1: how much soot the filter holds at the start

More soot means more material to burn, and a thicker soot layer also insulates and slows heat transfer. A cycle that starts at a higher soot load simply has more work to do. Soot at start depends on how long the set ran since the last cycle, how hard the engine was producing soot in that time, and whether earlier cycles removed everything.

Clue in the log: ΔP at the start of the cycle is higher than it used to be, at a comparable load.

Factor 2: how much useful heat reaches the soot

Heat comes from the engine exhaust, from the heater or burner, and from the soot’s own combustion once it ignites. A weaker heater, a lower exhaust temperature or a cooler ambient all reduce the heat that reaches the filter bed.

Clue in the log: inlet temperature or bed temperature reaches its target later, or never reaches the target at all, while ΔP at the start looks normal.

Factor 3: how fast heat leaks away or is carried off

A cold, windy installation, a missing insulation blanket, a bypass leak or a high exhaust flow all drag heat away from the filter. More airflow through the device means more soot can be oxidised if oxygen is the limit, but too much flow cools the bed.

Clue in the log: the heater runs at its normal power and soot starts at a normal level, yet the temperature rise is lower than before. Insulation damage, wet insulation or a changed exhaust arrangement are typical suspects.

The trick is to read three columns side by side: ΔP at start, temperature rise and heater behaviour. A longer duration with a high start ΔP points to Factor 1. A longer duration with a slow temperature rise points to Factor 2 or 3. A longer duration with both suggests more than one thing is going on.

Reading the pattern: a diagnostic table

The table links the pattern in your log to the most likely explanations. Treat the right-hand column as a list of suspects, not a verdict.

What you seeMost likely factorSuspects to check
Duration rising, start ΔP rising, temperature rise normalSoot inLonger light-load running between cycles; engine producing more soot (injectors, air filter, turbo); previous cycles incomplete
Duration rising, start ΔP normal, temperature rise slowerHeat inHeater element degraded; heater supply voltage low; fuel-assist burner fouled; catalyst less active so soot ignition needs more heat
Duration rising, start ΔP normal, heater power normal, bed temperature lowerHeat outWet or missing insulation; air leak at flange or bellows; exhaust path changed; colder ambient
Duration rising and start ΔP rising together with slow temperature riseMixedPartial blockage by ash or unburnt oily deposit; sensor drift; thermal ageing
Duration rising only in some monthsSeasonalCooler nights, damp monsoon conditions, load pattern changes
Duration rising and post-cycle ΔP not returning to baselineResidual loadingAsh accumulation; oily soot not fully burned; filter nearing cleaning interval
Duration suddenly jumps then stays highEventSensor replaced, firmware or setpoint changed, engine repair, new fuel batch
Duration falling sharply with low post-cycle ΔPPossible bypass or crackNeeds inspection, not celebration

That last row matters. A cycle that suddenly gets shorter is not always good news. If the filter has cracked or a seal has failed, gas can bypass the soot and the ΔP signal collapses, so the controller sees “done” early. Always pair a duration trend with a ΔP trend.

Slow drift versus sudden step: two different stories

The slow drift

A gradual lengthening over months is usually a wear or loading story. Ash builds up and does not burn off. Catalyst activity fades, so soot ignites later. A heater element ages and delivers slightly less. The engine drifts out of tune and makes more soot. None of these needs an emergency response, but all of them have a cost if ignored, because cycles that run longer use more heater energy or fuel, last longer in the window where the engine is carrying extra load, and give more chances to be interrupted.

The sudden step

A step change that appears on one date and stays is almost always an event. Something was changed, replaced or broken. Check the maintenance log for the week before the step. Sensor replacements, controller firmware updates, a repaired exhaust joint, a new insulation wrap, a change in the generator’s usual load, or a fresh batch of diesel all produce steps. If you have kept a regeneration and event log, the cause is usually visible within minutes.

The saw-tooth

Cycles that alternate between short and long are often a mix of two different load patterns. For example, weekday cycles run on a loaded set and weekend cycles on an idling set. Group the cycles by load band before judging.

A worked twelve-month example

The following is an illustrative scenario, not a real customer. A 250 kVA DG set at a manufacturing unit has a heater-assisted RECD. The plant maintenance engineer extracts the controller’s regeneration records and tabulates quarterly averages for cycles run at a similar load band (roughly 40 to 60% of rating).

QuarterAvg start ΔP (mbar)Avg peak bed temp (°C)Avg heater-on time (min)Avg cycle time (min)Avg end ΔP (mbar)
Q152590223831
Q254585244032
Q363560314936
Q466545365741

First reading. Cycle time rose from 38 to 57 minutes, which is a 50% increase. That is the headline.

Second reading: Factor 1 (soot in). Start ΔP rose from 52 to 66 mbar, an increase of 14 mbar, or about 27%. Soot loading at the start is higher, but not by enough to explain a 50% jump in time on its own.

Third reading: Factor 2 (heat in). Peak bed temperature fell from 590 to 545 °C, and heater-on time rose from 22 to 36 minutes, a 64% increase. The heater is working longer and reaching less. That is the strongest signal: heat delivered is shrinking.

Fourth reading: end ΔP. End ΔP crept from 31 to 41 mbar. The device is not returning to its earlier clean baseline, which suggests material that regeneration does not remove, or soot that is no longer burning out completely.

What the engineer does. The first check is the heater supply. A clamp meter on the heater feeder during a cycle shows current 15% below the commissioning value, and the supply voltage dips when the set carries load. A loose lug on the heater breaker is found and re-terminated. The second check is insulation: one section of the insulation blanket around the device is damp from a leaking roof penetration. It is dried and replaced. The third step is a supplier-supervised inspection of ash load, because the end ΔP drift suggests a growing ash layer. Ash cleaning is then scheduled, since regeneration cannot remove it. See our explanation of soot versus ash in RECD regeneration.

Result (illustrative). After the repairs and cleaning, the next quarter’s cycle time returns to about 41 minutes with end ΔP near 33 mbar. The drift was a three-cause problem: weak heater supply, wet insulation and accumulated ash. Any one of them alone would have produced a smaller change.

The lesson: a single duration number could not have told the engineer this. The four extra columns did.

Why longer cycles matter even when they finish

It is tempting to say “it completes, so leave it”. There are good reasons not to.

  • More exposure to interruption. The set may be stopped, loaded down or switched off mid-cycle. A thirty-minute window is easier to protect than a sixty-minute one. See what happens when the generator stops during RECD regeneration.
  • More energy used. Heater kWh or fuel per cycle goes up roughly in step with heater-on time.
  • Higher thermal stress. Cycles that need to push harder to ignite soot may run closer to the upper temperature limit, and a thicker soot layer burns more violently once it ignites. DieselNet’s overview of diesel particulate filter regeneration explains why soot load at the start of regeneration matters for filter temperatures.
  • Back-pressure on the engine. While the filter is loaded, the engine works against higher exhaust back-pressure. Our article on how exhaust backpressure affects a DG set covers the effect.
  • A shrinking margin. If a cycle already runs close to the controller’s time-out, any additional slowdown converts a “long cycle” into a “failed cycle”.

What to record before calling the supplier

A supplier can diagnose a rising duration far faster if you send a compact data pack. Gather the following for the last twelve months, or as much as exists.

  1. The definition of duration your controller uses, and the firmware version.
  2. Cycle records: date, start time, end time, heater-on time, start ΔP, end ΔP, peak and average temperatures at the inlet and bed (if measured).
  3. Load at the time: average kW or percentage during each cycle.
  4. Run history between cycles: hours run, average load and number of starts. The article on DG load logs and RECD suitability shows how to summarise this.
  5. Heater data: supply voltage and current during a cycle, compared with the commissioning sheet.
  6. Events: maintenance, engine repairs, sensor replacements, setpoint changes, fuel changes.
  7. Photographs of the insulation and exhaust arrangement around the device.

If you have a baseline from commissioning, include it. A good baseline turns an argument into a subtraction.

A simple routine for tracking duration

You do not need a data lake. A one-page monthly routine is enough.

Step 1: Fix the comparison conditions

Choose a load band, for example 40 to 60% of rating, and only compare cycles within it. Keep a second line for cycles outside the band.

Step 2: Log five numbers per cycle

Start ΔP, end ΔP, peak temperature, heater-on time and total cycle time. Add load.

Step 3: Plot, do not just tabulate

A rolling average of the last five cycles, plotted against date, shows direction at a glance. Add a horizontal line at the commissioning value, and another at the supplier’s maximum expected duration if they have given one.

Step 4: Set two trigger points

A yellow trigger might be a duration 20% above baseline for three consecutive cycles. A red trigger might be 40% above baseline, any cycle hitting the time-out, or end ΔP failing to recover. These percentages are only starting suggestions; agree your own with the supplier. For turning triggers into operator actions, see the alarm response matrix for operators.

Step 5: Review monthly with a named owner

Trends only help if someone looks. Put the review in the monthly maintenance routine and sign it.

Common misreadings

“The cycle is longer, so the device is cleaning more thoroughly.” Usually not. A longer cycle that still leaves a high end ΔP is a struggling cycle.

“The heater must be failing.” Possible, but not the only candidate. Wet insulation, low supply voltage and higher soot load can each produce the same symptom. Measure before replacing.

“Duration is controller-defined, so I cannot trust it.” The number may be defined by the controller, but trends in it still carry information if the definition is constant.

“A shorter cycle after cleaning proves it is fixed.” Cleaning removes ash and some deposits, but if the underlying cause is a weak heater, the improvement may be temporary. Watch the next three cycles.

“Seasonal change cannot matter for a hot country.” Cooler winter nights, damp monsoon insulation and changing load patterns all shift duration. Our article on passive RECD regeneration during seasonal changes covers what to track through the year.

When to escalate immediately

Do not wait for the monthly review if any of the following happens:

  • Cycles repeatedly time out or end flagged as incomplete. See RECD regeneration not completing.
  • End ΔP is rising from cycle to cycle even though duration looks stable.
  • Heater current is zero or erratic during a cycle.
  • Back-pressure approaches the supplier’s alarm limit.
  • Visible smoke, a burning smell or unusual noise appears during a cycle.

In those cases, follow your site procedure, reduce load if safe, and contact the supplier. Repeated attempts without a diagnosis carry their own risks, discussed in our post on whether repeated failed regeneration can damage an RECD filter.

FAQ

How long should an RECD regeneration take?

There is no universal figure. It depends on the device design, the heater or fuel-assist capacity, soot load and engine load. Your supplier’s commissioning data gives the correct baseline for your unit. Judge your device against its own history, not against a number from a different design or from vehicle DPFs.

Is a rising regeneration duration always a sign of a failing filter?

No. Duration often rises first because of causes outside the filter: a weak heater supply, damp insulation, more soot from the engine or a change in how the set is loaded. The filter itself may be perfectly sound. The trend tells you to investigate, not to replace.

Can I shorten the cycle by running the DG at higher load?

For designs that use exhaust heat, a higher load usually raises exhaust temperature and may help. For heater-assisted designs, extra load changes both exhaust flow and available heat in different ways. Follow the supplier’s guidance and make sure the DG can safely carry the extra load. Do not exceed the generator’s limits to chase a better regeneration.

Why does duration differ between summer and winter?

Ambient temperature changes heat loss from the device, engine load patterns shift with seasons, and damp conditions can affect insulation and heater supply. Compare cycles within similar seasons and load bands.

Does a longer cycle use more fuel?

For heater-assisted designs the extra cost is electrical energy, which a DG set supplies from diesel. For fuel-assisted designs it is direct fuel. In both cases longer cycles consume more. Our article on RECD regeneration energy per cycle shows how to measure it.

Should I clean the filter when duration rises?

Only if the data points to ash or residual loading, for example end ΔP no longer recovering. Cleaning does not fix a weak heater or wet insulation. Diagnose first. Our post on regeneration requests after cleaning explains what to expect afterwards.

Bringing it together

A rising RECD regeneration duration is a signal, not a diagnosis. Read it against start ΔP, temperature rise, heater behaviour and end ΔP, and the three-factor model (soot in, heat in, heat out) will usually point you to the right area within an afternoon. Keep a stable definition, compare within load bands, set yellow and red triggers, and send the supplier a clean data pack when the trend crosses them.

If you operate a DG set with an RECD and want help interpreting your regeneration records, you can read about our retrofit emission control devices, check our certifications, or contact the Aceget team with your cycle log and we will help you read it.

Related Reading



2 Comments

Leave a Reply