- November 13, 2026
- Posted by: editor
- Category: RECD & Emission Control Devices
The controller says “Regeneration complete.” The operator ticks the log, moves on and the generator carries on supplying the building. Two days later differential pressure (ΔP) is back above its alarm level and the same message appears again.
What went wrong? Perhaps nothing. Perhaps the message meant something narrower than the operator assumed. An RECD regeneration completion message is the controller’s statement about its own process: a timer ran out, a temperature target was met, a ΔP threshold was crossed or an internal flag was set. It is not a guarantee that the filter is clean, that the soot burned evenly, or that the device will stay clear until the next scheduled cycle.
This article shows how to treat a completion message as a claim to be verified. We look at what messages usually mean, the five facts that confirm or contradict them, annotated sample logs and a checklist that an operator can apply in two minutes. Message wording differs from one controller to another, so the examples are illustrative; always match the terms to your own device manual.
What a controller means by “complete”
Controllers decide that a regeneration has finished in one or more of these ways. Know which one yours uses.
| Completion logic | What it means | What it does not guarantee |
|---|---|---|
| Timer-based | The heater or burner ran for a set time, or the cycle ran for a fixed duration | That enough heat reached the filter, or that all soot burned |
| Temperature-based | Bed or outlet temperature stayed above a target for a set time | That temperature was even across the filter face |
| ΔP-based | ΔP fell below a recovery threshold at a stated flow or load | That ΔP was measured at a comparable load, or that the sensor is accurate |
| Model-based | An internal soot model estimated that soot dropped below a limit | That the model matches reality after interruptions or sensor drift |
| Combined | Several of the above, often with time-outs | That every condition was weighted the way you assume |
The same word, “complete”, can therefore sit on top of very different evidence. A timer-based completion in a cold, damp installation might be technically correct and still leave a loaded filter. A ΔP-based completion at light load might look good because low flow produces low ΔP, even though the filter is far from clean.
For background on how regeneration is triggered and ended in different designs, see active versus passive RECD regeneration.
The message families you are likely to meet
Message names vary, but most fall into a handful of families. The table suggests how much weight to give each family before you check anything else.
| Message family | Typical wording (illustrative) | First interpretation | Weight before verification |
|---|---|---|---|
| Normal completion | “Regeneration complete”, “Regen OK”, “Cycle finished” | Controller believes its end conditions were met | Moderate, verify |
| Completion with warning | “Complete, low temperature”, “Complete, time limit” | Cycle ended but conditions were marginal | Low, treat as partial |
| Time-out | “Regeneration timeout”, “Max duration reached” | Cycle ran out of time before the target was met | Treat as failure |
| Abort by controller | “Regeneration aborted”, “Interlock”, “Sensor fault” | Controller stopped it because a condition failed | Treat as failure; read the reason |
| Abort by external event | “Generator stopped”, “Load drop”, “Supply lost” | Something outside the controller ended it | Treat as failure |
| Inhibited | “Regeneration inhibited”, “Not allowed” | Cycle never started | Not a completion. See RECD regeneration inhibit signals |
| Request pending | “Regeneration requested”, “Awaiting conditions” | Controller wants a cycle but conditions are not met | Not a completion |
If your controller reports only a green light or a single bit on a Modbus register, you have even less information. Our guide to RECD Modbus data, units and register mapping explains how to find out what the bit means.
The five facts that verify a completion message
A completion message is believable when five facts line up. Check them in order. The whole routine takes minutes once you have the log in front of you.
Fact 1: ΔP fell, and fell at a comparable load
Compare ΔP just before the cycle with ΔP after it, at a similar load and flow. A fall of ΔP from 55 to 38 mbar at 50% load is meaningful. A fall from 55 to 30 mbar because the load dropped from 50% to 20% is not.
How to check. Note the load at both readings. If they differ by more than about 10 percentage points, wait for the next stable run at the reference load and read ΔP again. See comparing RECD pressure readings at different generator loads.
Fact 2: The temperature curve looks like a regeneration
A real regeneration leaves a visible temperature signature: a rise to a working level, a plateau or gentle hump while soot burns and a fall afterwards. The shape depends on the design, but flat lines and strange spikes both deserve attention.
How to check. Plot or review inlet and bed temperature over the cycle. Compare the peak and duration with the commissioning cycle. A peak far below earlier cycles suggests the soot did not ignite. A very sharp spike suggests an uncontrolled burn. Our article on RECD regeneration temperature explains what ranges mean.
Fact 3: The heater or burner behaved as designed
For assisted designs, the energy input should match the cycle. Heater current, duty cycle and on-time should look like earlier successful cycles.
How to check. Compare heater-on time and current with a good reference cycle. If the log shows the heater on for the full duration at half the usual current, the “complete” message may rest on a timer rather than on delivered heat.
Fact 4: Nothing interrupted the cycle
A cycle that appears complete may have been shortened by a load drop, a transfer or a stop and restart.
How to check. Read the generator’s own event log and the building’s load record for the same window. Look for any dip in load, a transfer to grid, a stop or a controller restart. If the cycle lasted 38 minutes and the set was off for six of them, the message deserves less trust. See what happens when the generator stops during RECD regeneration.
Fact 5: The result persists
A genuine clean-up lasts for a while. If ΔP at the reference load climbs back to pre-cycle levels within a few hours of running, the soot was not removed, or a different problem exists, such as a leak, a sensor fault or oil contamination.
How to check. Read ΔP at the next two or three reference-load runs, and then again after the next full day of operation. If it holds, the cycle was real. If it bounces back, investigate. Our article on RECD regeneration requests after cleaning covers a related pattern.
Annotated sample log extracts
The three extracts below are fictional and simplified. They show how the same words can hide different stories.
Extract A: a believable completion
14:02 Regen requested dP=58 mbar load=52%
14:05 Regen started Tin=402C heater ON 6.0 kW
14:25 Heater OFF Tbed=588C
14:41 Regen complete dP=36 mbar load=51% duration=36 min
Reading it. ΔP fell by 22 mbar at nearly the same load. The bed reached 588 °C and the heater did its work for 20 minutes. Load was steady. This is a believable completion. Verify Fact 5 on the next run and file the log.
Extract B: complete, but only by the clock
02:10 Regen requested dP=61 mbar load=19%
02:12 Regen started Tin=318C heater ON 4.1 kW
02:52 Heater OFF Tbed=441C
02:57 Regen complete dP=50 mbar load=19% duration=45 min
Reading it. The message says complete, but the bed only reached 441 °C, the heater power was lower than the 6 kW seen in extract A, and ΔP fell by 11 mbar rather than 22. The set was at light load, so exhaust heat contributed little. This is a partial cycle that the controller ended on a timer. The right action is to log it as incomplete in the operator’s own record, plan a loaded run, check the heater supply voltage and watch ΔP at the next reference run. See choosing an RECD for a generator running at low load.
Extract C: complete, but interrupted
19:30 Regen requested dP=64 mbar load=47%
19:33 Regen started Tin=395C heater ON 6.0 kW
19:48 Load transfer to grid (DG unloaded)
19:51 DG restart, load 45%
19:53 Regen resumed
20:15 Regen complete dP=44 mbar load=46% duration=42 min
Reading it. The cycle started, lost its heat and flow when the set was unloaded for several minutes, and resumed. The controller treated the whole as one cycle, so the message says complete. ΔP did fall, but only by 20 mbar and the bed temperature trace (not shown) probably dipped. Treat it as complete with doubt. Check ΔP over the next two runs and review how the controller handles resumes. For battery and supply issues that cause such interruptions, see RECD regeneration and battery-supported controls.
Reading a completion alongside the alarm picture
A completion message sits in a context of other alarms and warnings. Pair it with:
- Sensor health. A completion based on a faulty ΔP sensor is meaningless. See separating a sensor fault from an RECD performance problem.
- Back-pressure. If DG-side back-pressure also fell, the clean-up is more credible. See how exhaust backpressure affects a DG set.
- Temperature alarms. A high-temperature warning during the cycle suggests an aggressive burn.
- Communication status. If the data link dropped during the cycle, the stored record may be incomplete; see detecting missing data in RECD monitoring records.
An operator’s two-minute checklist
Print this and place it near the controller or in the shift handover folder.
| Step | Check | Pass looks like | If not |
|---|---|---|---|
| 1 | What exactly does the message say? | Plain “complete”, no warning qualifier | Treat “warning”, “limit” or “timeout” as incomplete |
| 2 | ΔP before vs after at similar load | Clear fall, load within about 10 points | Note, repeat the reading at the next stable run |
| 3 | Peak temperature vs reference cycle | Within the expected range | Log as partial; call supplier if repeated |
| 4 | Heater or burner behaviour | Normal current and on-time | Check supply; do not reset breakers repeatedly |
| 5 | Any interruption in the window | None | Treat as doubtful; recheck after next run |
| 6 | Entry in log | Written with values, not just “OK” | Complete the entry |
| 7 | Next verification | Scheduled ΔP check at reference load | Set a reminder |
Write the values in the log, not just the word “complete”. The next shift cannot learn from a tick mark.
What a supervisor does with completion messages
Supervisors can improve the quality of the whole process with three habits.
Define “complete” for your site
Write a one-paragraph definition that matches your controller: for example, “A regeneration counts as complete when ΔP at 50% load falls by at least X mbar from the pre-cycle value, bed temperature exceeded Y °C for at least Z minutes and no interruption occurred.” Agree the numbers with the supplier. This turns a vague message into an acceptance rule.
Review completions weekly
Spend ten minutes a week reading the last few cycles. Look for patterns: increasing duration, falling recovery, completions with warnings. Our article on what a rising regeneration duration tells you shows how to read the trend.
Feed the results into the history report
A good regeneration history report keeps the controller’s verdict next to your verification result. See what an RECD regeneration history report should include.
Common misinterpretations
“Complete means the filter is clean.” It means the controller’s end condition was met. Clean-ness is shown by ΔP at a reference load and by the next days of operation.
“No alarm means success.” A cycle can end with no alarm and no real clean-up, especially at light load.
“Complete twice in one day is double cleaning.” Frequent completions can mean the filter reloads quickly, a symptom of engine-side soot production or a poor recovery.
“A controller reset clears the problem.” Resetting a counter or model does not change the soot in the filter. Reset only under supplier instruction.
“Time-out is the same as completion.” A time-out means the cycle ran out of time. Some controllers present it in a neutral colour or with mild wording, so read the text.
“The message is wrong, so ignore it.” A message that contradicts your data is itself information. It might indicate a sensor fault, a configuration problem or a mislabelled output. Investigate; do not ignore.
Worked example: two cycles, two decisions
The following is an illustrative scenario. A data centre’s maintenance engineer reviews two completions from the same week.
Cycle 1, Tuesday: pre-cycle ΔP 57 mbar at 55% load. Post-cycle ΔP 35 mbar at 54% load. Bed peak 592 °C, heater normal, no interruptions. At the next run on Thursday, ΔP at 55% load is 36 mbar. Decision: accept. Log the cycle as verified.
Cycle 2, Saturday: pre-cycle ΔP 60 mbar at 21% load. Post-cycle ΔP 49 mbar at 20% load. Bed peak 455 °C, heater current 70% of normal. The message reads “Regeneration complete”. Decision: mark as partial. Raise a job to check heater supply voltage, schedule a loaded test run on Monday and set a review. On Monday, ΔP at 50% load reads 44 mbar, still above the 36 to 38 mbar range seen after good cycles. The engineer asks the supplier to supervise a full cycle. A loose heater terminal is found and corrected.
Without verification, both messages would have gone into the log as identical successes, and the second would have set up a failed-cycle loop. See can repeated failed regeneration damage an RECD filter for why that matters.
Where the message appears matters
The same event can reach you in three ways, and each carries a different amount of detail.
| Where you read it | What you usually get | Risk |
|---|---|---|
| On the RECD controller screen | Full text, reason codes and often the stored cycle values | Operators may not look at it unless there is a prompt in the routine |
| On the DG controller or panel | A summary lamp or a short text such as “RECD regen OK” | A one-line summary may hide a time-out or warning |
| On the BMS or remote dashboard | A status bit or short label mapped from a register | Mapping errors; stale values if communication drops |
A message that has passed through two translations (RECD controller to Modbus register to BMS label) can lose its qualifier. “Complete, low temperature” becomes “Complete”. If your operators rely on the BMS, check the mapping once with the supplier and the BMS contractor, and test it by triggering or simulating each message type during commissioning. Our guides to integrating RECD monitoring with a building management system and to RECD alarm versus generator trip control response explain how status signals can be arranged so that the important distinctions survive.
Turning the checklist into one line of the site procedure
A long article is not an operating procedure. Condense what you need into a short rule that fits the site’s standard operating document. For example:
After any regeneration message, record the message text, ΔP and load before and after, peak temperature and any interruption. Mark the cycle “verified” only when ΔP at a similar load has fallen by the agreed margin and the next reference run confirms it. Otherwise mark it “partial” or “failed” and follow the retry rule.
Notice what the rule does. It tells the operator to write down values, it defines “verified” in measurable terms and it points to the next action. Agree the numbers with your supplier, and add the rule to the shift handover form so that it is seen at every changeover. See the RECD shift handover log for a layout that works.
Questions to ask your supplier about completion logic
Next time the supplier visits, ask:
- How does the controller decide a regeneration is complete: timer, temperature, ΔP, model or a combination?
- What are the exact thresholds and can they be viewed or changed? Who may change them?
- What message does the controller show for a time-out, abort, interruption and resume?
- How does the controller handle an interruption: does it restart, resume or abandon the cycle?
- Does the controller log the reason when it aborts?
- Which values are stored for each cycle, and how long are they kept?
- Which sensors does the completion decision rely on, and how are they calibrated?
- What should an operator do after each type of message?
Write the answers into the site procedure and the alarm response matrix.
FAQ
Does “regeneration complete” mean the RECD is clean?
Not on its own. It means the controller’s end conditions were met. Confirm with a drop in ΔP at a comparable load, a normal temperature profile, normal heater behaviour, no interruptions and a result that persists over the next runs.
Why does ΔP rise again soon after a completed regeneration?
Possible reasons include an incomplete burn, engine-side soot production, wet or oily soot, ash accumulation, sensor drift or a leak. A rapid rebound should be investigated, not accepted.
What is the difference between a time-out and a completion?
A completion means the end conditions were reached. A time-out means the cycle reached its maximum allowed duration without meeting them. Controllers sometimes present both with similar wording, so read the full message and the stored values.
Should operators reset the controller after a completion?
No, unless the supplier’s procedure says so. Resetting does not remove soot. It may hide information that the supplier needs.
Can I rely on the green light on the controller?
Treat it as a status indicator, not proof. Use the logged values, ΔP and temperature to confirm.
How many past completions should I keep?
At least twelve months, ideally the whole life of the device, so you can compare recent cycles with commissioning. Our article on preserving RECD data during controller replacement explains how to keep records safe.
Bringing it together
An RECD regeneration completion message is a starting point, not a verdict. Learn how your controller defines completion, check the five facts (ΔP at a comparable load, temperature curve, heater behaviour, interruptions and persistence), record values rather than ticks, and write a site definition of “complete” that you agree with your supplier.
If you want help reading controller logs or choosing a device with clear completion reporting, see our retrofit emission control devices, browse the FAQ or contact the Aceget team.
[…] start, end, ΔP, temperatures and energy input, and should report failures clearly. Our guides on interpreting an RECD regeneration completion message and what an RECD regeneration history report should include set out the evidence you should be able […]
[…] 61 to 38 mbar, a drop of 23 mbar, at a steady DG load. That supports a successful clean-up. See interpreting an RECD regeneration completion message for the checks behind that […]