How DG Load Logs Help Assess RECD Suitability Before You Buy

Two DG sets can share a rating, an engine model and an exhaust route, and still treat a retrofit emission control device very differently. The difference is how they are used. One runs for hours at a healthy load during long outages. The other starts for a few minutes several times a day, carries a handful of lights at night, and sits through weekly tests with no load at all.

The RECD’s particulate filter feels that difference directly. It traps soot all the time and burns it off when the exhaust is hot enough. How hot the exhaust gets, and for how long, depends on load and run length. So the most useful document for judging whether an RECD will thrive on your set is often not a data sheet but a log: the record of when the set ran, for how long and how hard.

This article treats the load log as an assessment tool. It covers where to find the data, the seven numbers worth pulling from it, a worked reading of a 30-day log, three common usage patterns and what each means for an RECD, the fixes that improve a weak profile, and a log format you can start using today.

Why the Log Matters More Than the Nameplate

A particulate filter cleans itself through passive regeneration. Soot trapped in the filter oxidises when exhaust temperature stays high enough for long enough, helped by a catalyst coating that lowers the temperature needed. Our explainer on what a diesel particulate filter is and how it helps and the Wikipedia entry on diesel particulate filters both describe the process.

Exhaust temperature rises with load. A set at 60% load produces much hotter exhaust than the same set at 15%. Run length matters too: the exhaust system needs time to heat up after a cold start, and a run that ends after five minutes may never reach working temperature.

The nameplate tells you the set’s capacity. The log tells you how much of that capacity you actually use, and in what pattern. Two things follow:

• A set with a healthy log will usually keep its filter clean through normal running, with routine attention.
• A set with a weak log can still take an RECD, but it may need operating changes, a different device specification or both.

Light loading also harms the engine itself. Unburned fuel and oil can collect in the exhaust, a condition known as wet stacking, which adds sticky deposits that reach the filter. Our article on understanding generator load factor and duty cycle covers the engine side.

Where Load Data Comes From

Most sites have more load data than they realise. Check these sources.

Source	What it typically records	Strengths	Limitations
DG controller event log	Start and stop times, alarms, sometimes kW and kVA snapshots	Accurate times; already on site	May only hold recent events; needs software or a technician to download
AMF panel	Grid failure and restoration, DG start and stop	Shows the link between outages and runs	Often no load data
Energy meter on DG output	kWh, kW, kVA, power factor, often with interval logging	Good load data	Interval memory may be short; needs download
Building management system	Run status, load, sometimes exhaust temperature	Long history, easy export	Depends on what was integrated
Manual DG logbook	Start, stop, hours meter, sometimes amps and voltage	Long history	Accuracy varies; load often not recorded
Fuel log	Litres added, dates	Cross-check on total running	Indirect
RECD or IoT monitoring (after fitment)	Exhaust temperature, backpressure, run time	Direct view of filter conditions	Only available once a device is installed

Combine sources where you can. An AMF event log gives you accurate start and stop times; a meter gives you load during those runs; the manual logbook fills gaps.

How Much Data Is Enough

A single week rarely represents a full pattern. Aim for at least 30 days of data, and longer if possible. Sites with strong seasonal patterns should look at the worst season for the generator:

• Industrial and commercial sites often see more outages and heavier DG use in summer.
• Sites in NCR may see winter DG use change sharply when GRAP restrictions apply. Our post on GRAP rules for DG sets in Delhi NCR explains the background.
• Sites with weekly test routines should include enough weeks to capture several tests.

If you have no load data at all, start logging now. Even a month of manual readings improves the assessment.

The Seven Numbers to Pull from a Load Log

You do not need advanced analysis. These seven figures tell most of the story.

1. Average load while running

Total kWh delivered by the DG divided by total running hours, expressed as a percentage of the set’s rated kWe. This is the headline figure, but on its own it can hide problems.

2. Share of running hours below 30% load

Generator engine makers commonly advise against prolonged running at light load, and 30% of rating is a widely used rule of thumb for the point below which problems start. The share of hours spent below that line matters more than the average.

3. Share of running hours above 40% load in runs of an hour or longer

These are the hours that do the most for passive regeneration: sustained, reasonably loaded running. The higher this share, the better.

4. Median run length

Half of all runs are shorter than this figure. A median of five minutes means most runs end before the exhaust system warms up.

5. Share of runs shorter than 15 minutes

Short runs add soot and condensate without much chance to burn it off.

6. Starts per month

A high number of starts, particularly short ones, suggests unstable grid supply or a sensitive transfer scheme. Each cold start adds a burst of soot.

7. No-load running

Hours run with no load at all, usually during tests. These are the least useful hours for a filter and among the worst for the engine.

A Worked Reading: 30 Days on a 320 kVA Set

Here is an illustrative month for a 320 kVA set (about 256 kWe) at a commercial building in Noida. The set backs up the whole building during grid failures and has a weekly 10-minute no-load test on Saturdays. The data is composite, built to show the method.

Run	Date	Start	Duration (min)	Average kW	Load (% of 256 kWe)	Note
1	2 Sep	10:05	8	70	27%	Brief outage
2	4 Sep	14:20	95	150	59%	Afternoon outage
3	6 Sep	11:00	10	0	0%	Weekly test
4	9 Sep	18:40	22	120	47%	Evening outage
5	11 Sep	03:10	180	55	21%	Night outage
6	13 Sep	11:00	10	0	0%	Weekly test
7	15 Sep	12:30	240	165	64%	Planned grid shutdown
8	18 Sep	21:15	35	80	31%	Evening outage
9	20 Sep	11:00	10	0	0%	Weekly test
10	23 Sep	16:05	6	140	55%	Brief outage
11	25 Sep	02:00	150	50	20%	Night outage
12	27 Sep	11:00	10	0	0%	Weekly test

Calculating the seven numbers

Total running time: 776 minutes, or about 12.9 hours.

Energy delivered: the sum of kW × minutes for each run is 78,090 kW-minutes, or about 1,300 kWh.

1. Average load while running: 78,090 ÷ 776 ≈ 101 kW, or about 39% of rating. On its own, that looks reasonable.
2. Hours below 30% load: runs 1, 3, 5, 6, 9, 11 and 12 total 378 minutes, about 49% of running time.
3. Hours above 40% load in runs of an hour or longer: runs 2 and 7 total 335 minutes, about 43% of running time.
4. Median run length: sorting the twelve durations gives a median of 16 minutes (midway between the sixth and seventh values, 10 and 22).
5. Runs shorter than 15 minutes: runs 1, 3, 6, 9, 10 and 12, so 6 of 12, or 50%.
6. Starts per month: 12.
7. No-load running: four tests of 10 minutes, 40 minutes.

What the reading says

The 39% average conceals a split personality. Almost half the running time is at light load, mostly from two long night outages when the building is nearly empty. Half the starts are short. The good news is that two long, well-loaded runs (one of four hours) give the filter real opportunities to regenerate.

For this set, an RECD is likely to be workable, with three changes:

• Make the weekly test a loaded run of 30 to 60 minutes by transferring building load to the DG.
• Review night outages. The DG carries only about 50 to 55 kW at night. If the building has a second, smaller set or can schedule some heavier loads (such as water pumping or chiller pre-cooling) to run on DG during long night outages, the load improves.
• Ask the vendor whether the proposed catalyst formulation and filter capacity suit a set that spends this much time at light load, and how often backpressure should be reviewed.

Three Usage Patterns and What They Mean

Most logs fall into one of three broad patterns. Recognising yours speeds up the conversation with a vendor.

Pattern A: The steady worker

Typical log: Long runs, mostly above 40% load, few short starts. Common at factories with continuous processes, sewage and water pumping plants, and sites where the DG runs as prime power.

RECD outlook: Favourable. Sustained hot exhaust supports passive regeneration. Routine monitoring and maintenance are usually enough.

Watch for: Overload. A set that runs near or above its rating smokes more and can load the filter quickly. Check peak load as well as average.

Pattern B: The sprinter

Typical log: Many short runs, often under 15 minutes, triggered by brief outages or voltage dips. Load during runs may be reasonable, but runs end before the exhaust heats up. Common in areas with unstable grid supply and at sites with sensitive changeover settings.

RECD outlook: Challenging if nothing changes. Each short run adds soot and condensate without regeneration.

Possible fixes: Review the AMF and changeover settings with your electrical contractor. A short delay before starting the DG can avoid starts for momentary dips, provided critical loads are covered by UPS. Add periodic longer loaded runs to burn off accumulated soot. Our article on whether an RECD suits a DG set that runs only occasionally covers related ground.

Pattern C: The sleeper

Typical log: Long runs at light load, often below 25%. Common where the DG was oversized for future expansion, where loads have shrunk after LED retrofits or equipment upgrades, and where night-time outages leave the DG carrying only a few lights.

RECD outlook: Needs attention. Cool exhaust for long periods means slow regeneration and possible wet stacking.

Possible fixes: Add load during DG running where it is useful, run a smaller set for light-load periods if one exists, schedule periodic loaded runs, and discuss a low-temperature-assisted catalyst or larger filter capacity with the vendor. In some cases, the right long-term answer is a correctly sized set.

Many real logs mix two patterns, as in the worked example above, which combined sprinter-style short runs with sleeper-style night outages.

A Traffic-Light Screen

The thresholds below are illustrative starting points to frame a discussion with your vendor. They are not limits from any standard, and the right figures depend on the engine and the device.

Measure	Green	Amber	Red
Average load while running	Above 40%	25% to 40%	Below 25%
Hours below 30% load	Under 25%	25% to 50%	Over 50%
Hours above 40% load in runs of 1 hour or more	Over 50%	20% to 50%	Under 20%
Median run length	Over 60 minutes	15 to 60 minutes	Under 15 minutes
Runs under 15 minutes	Under 20%	20% to 50%	Over 50%
No-load test running	None, tests are loaded	Some no-load tests	All tests no-load

Our worked example scores amber on most measures and red on short runs. That is typical of a commercial building in an area with patchy supply, and it is fixable.

Bringing Exhaust Temperature into the Picture

Load is a proxy. What the filter actually responds to is exhaust temperature at its inlet. If you can measure or log exhaust temperature, you have a more direct view.

Many catalysed particulate filters need sustained exhaust temperatures somewhere in the range of roughly 250 to 350°C or more to regenerate passively at a useful rate. The exact figure depends on the catalyst and the soot. Ask your vendor what applies to the device offered, then compare it with your logs.

Ways to get temperature data before fitment:

• A temporary thermocouple and data logger at the exhaust outlet for a few weeks
• A handheld reading taken during each run at stable load, recorded in the logbook
• The engine maker’s data sheet, which gives exhaust temperature at several load points; combine it with your load log to estimate the time spent at each temperature

Remember that exhaust cools as it travels along the pipe. A device several metres from the engine sees lower temperatures than the engine outlet. Our companion guide on using OEM exhaust flow data when specifying an RECD covers how to read the data sheet and allow for pipe losses.

Fixes That Improve a Weak Log

If your log shows a weak profile, these measures can help. Discuss them with your electrical contractor, DG service provider and RECD vendor.

Make tests count

Replace short no-load tests with loaded runs. Transfer real building or plant load to the DG for 30 to 60 minutes. If that is not practical, a portable or permanent load bank can provide the load. Weekly no-load tests prove the set starts; loaded tests also keep the engine and filter healthy.

Tune the changeover

Momentary dips and very short outages trigger many unnecessary starts. Changeover timing and DG start delays should be set by a qualified electrician, with attention to critical loads and UPS autonomy.

Rebalance loads during outages

During long outages, deliberately bringing useful loads onto the DG, such as water pumping, chilled water storage charging, battery charging or production that can be shifted, raises load and exhaust temperature.

Use the right set for the load

Sites with two sets of different sizes can run the smaller set during light-load periods. Sites with one large set and persistent light loads may benefit from a smaller second set in the long run. Rating standards assume a set will do meaningful work: the prime rating described in Cummins’ note on ISO 8528 generator set ratings, for example, is built around an average load factor rather than long periods of near-idle running. Our article on standby vs prime power generators explains how duty affects sizing, and load calculation basics before buying a generator helps if replacement is on the table.

Fix the engine

A healthy engine produces less soot. Check injectors, air filter, turbocharger and oil consumption before fitment. Our guide on whether to complete an engine overhaul before fitting an RECD helps decide.

Insulate the exhaust

Insulating the pipe between engine and device keeps exhaust hotter at the device inlet, helping regeneration on lightly loaded sets.

A DG Load Log Format You Can Start Today

If you have no electronic logging, a simple sheet in the DG room is enough. Ask the operator to fill one line for every run.

Field	What to record
Date
Reason for run	Grid failure, test, planned shutdown, other
Start time and stop time	From the controller or a clock
Hours meter at start and stop	From the DG hours meter
Load reading	kW from the controller or meter; if only amps are available, record amps on each phase and voltage
Time of load reading	At least one reading after the set has stabilised; more for long runs
Exhaust temperature	If a gauge or handheld instrument is available
Smoke observed	None, light, heavy; at start or under load
Alarms	Any controller or RECD alarms
Operator initials

If only current is available, approximate three-phase kW as 1.732 × voltage (line-to-line) × average current × power factor ÷ 1,000. Assume a power factor of about 0.8 if you have no better figure, but treat the result as approximate.

Our monthly maintenance checklist for diesel generators pairs well with this log.

Using Logs After the RECD Is Fitted

The log does not stop being useful once the device is installed. It becomes more useful, because you can now set it against the device’s own readings.

• Backpressure versus load. Plot RECD backpressure against DG load for each run. A steady relationship means the filter is in balance. A rising trend at the same load means soot is accumulating faster than it burns.
• Recovery after loaded runs. After a long loaded run, backpressure at a given load should fall as regeneration clears soot. If it does not, discuss it with the vendor.
• Early warning. A sudden rise in backpressure with no change in load pattern often points to an engine problem such as a failing injector or oil leak.
• Evidence for inspections. Logs showing run hours, loads and device readings support your records during a pollution control board visit. See what happens during a pollution control board inspection.

Our RECD maintenance guide explains routine checks, and the environmental compliance checklist for generator owners covers wider record keeping.

What to Send Your Vendor

When requesting an RECD proposal, include a short load summary with the seven numbers, the raw log or export, and a note on any planned changes to the site’s loads. A vendor who sees real data can choose a device specification and recommend operating practices with far more confidence than one working from the nameplate. The article on DG nameplate photos for an RECD enquiry covers the rest of the enquiry pack, and the guide to choosing CPCB approved RECD vendors helps you compare the replies.

Frequently Asked Questions

Why does an RECD vendor ask for DG load logs?

Because the filter’s ability to clean itself depends on exhaust temperature, which depends on load and run length. Logs show whether the set runs hard enough and long enough for passive regeneration, and whether operating changes or a different device specification are needed.

What if my DG has no load logging at all?

Start a manual log now, recording start and stop times and a load reading for each run. Download the controller’s event log if possible. Even a month of data helps.

Is a 40% average load good enough for an RECD?

It may be, but averages can hide long periods at light load and many short runs. Look at the share of hours below 30% load and the share of runs under 15 minutes as well.

Do weekly no-load tests harm the RECD?

They do little for it and can harm the engine through wet stacking. Replace them with loaded runs of 30 to 60 minutes where possible.

My DG starts many times a day for a few minutes. Can it take an RECD?

It can, but frequent short runs are hard on the filter. Review changeover settings, add periodic long loaded runs and discuss the pattern with the vendor before fitment.

How long should I log before assessing RECD suitability?

At least 30 days, preferably covering the season when the DG runs most. Longer records give a more reliable picture.

Let the Log Make the Case

A load log turns a guess about how your generator is used into evidence. Pull the seven numbers, recognise the pattern, fix what can be fixed, and give the vendor the data. The result is an RECD matched to how your set actually runs, rather than to how its nameplate suggests it might.

Aceget supplies CPCB type-approved RECDs for DG sets from 25 kVA to 2500 kVA, with IoT monitoring that records exhaust temperature and backpressure so your log keeps working after fitment. For the size in this article’s worked example, see the 320 kVA DG set RECD. To share your logs for a review, contact Aceget or explore the retrofit emission control device range.