- October 25, 2026
- Posted by: editor
- Category: RECD & Emission Control Devices
Many large sites in India run a bank of diesel generators: four, five or more identical sets on a common bus. On a normal outage day only one or two start. The third comes on during a heavy afternoon. The last two run only when everything goes wrong at once, such as a long grid failure during peak production or a feeder fault that pushes the whole site onto backup.
That pattern creates a planning problem for Retrofit Emission Control Devices. The lead set works hard and its RECD behaves predictably. The reserve sets sit for weeks, start for a few minutes on a test, and are then suddenly asked for full output for six hours while the site is under pressure. If their RECDs, sensors and controllers have not been looked after with the same discipline, that is exactly when problems appear.
This guide covers RECD generator bank planning for sites like this. It starts by clearing up common beliefs, then organises the plan by the role each set plays. It closes with a worked example for five 1,010 kVA sets, a maintenance schedule by role, and a way to phase costs without leaving any in-use set outside the rules.
Myth vs fact: what owners assume about reserve sets
Myth 1: “The reserve sets barely run, so they don’t matter”
Fact: Low hours do not mean low importance. A reserve set exists for the worst day of the year. If it fails to start, trips on a high back-pressure alarm, or smokes heavily when it takes load, the bank loses capacity when capacity matters most. Its few hours are also often the hardest ones for emissions: a cold start after a long idle period, followed by a sudden step to high load.
Low hours do not reduce the compliance obligation either. Regulators generally ask whether a set is installed and available for use, not how many hours it logged last year.
Myth 2: “Only the lead set needs an RECD”
Fact: Under the approach followed by CPCB and most State Pollution Control Boards, each DG set in use within the covered capacity range is generally expected to meet the applicable emission requirement. For older sets this usually means a CPCB type-approved RECD achieving at least 70% particulate matter reduction. Rules, timelines and exemptions differ between states and regions, with Delhi NCR under GRAP often the strictest. Check the position with your SPCB and your consent conditions before assuming any set can be left out. Our explainer on why RECDs are mandatory in India and the DG set emission regulations page give the background.
Myth 3: “A weekly no-load start keeps an idle set healthy”
Fact: A short off-load run often does more harm than good. The engine never reaches operating temperature, fuel burns incompletely and moisture condenses in the exhaust. Unburnt fuel and oil collect as wet deposits (wet stacking), and the RECD gathers soot without getting hot enough for passive regeneration. A start check proves the starting system; it is not a substitute for loaded exercise. DieselNet’s overview of particulate filter regeneration explains why exhaust temperature matters so much for soot burn-off.
Myth 4: “If it passed commissioning, it will be fine when we need it”
Fact: Commissioning proves the system worked on one day under controlled conditions. Months of idle time can bring corroded flanges, a drifting sensor, a controller that lost its settings after a battery failure, or a filter partly loaded with soot from repeated test runs. Readiness has to be demonstrated periodically, not assumed.
Why idle periods are hard on an RECD-equipped set
Each idle-period risk has a matching control, so it pays to be specific.
| Idle-period risk | What happens | Effect on the RECD | Matching control |
|---|---|---|---|
| Condensation | Moisture condenses in cool pipework and housing | Water mixes with soot and sulphur compounds into acidic deposits | Loaded runs long enough to dry the system |
| Corrosion | Condensate and humid or coastal air attack flanges and clamps | Joint leaks release unfiltered exhaust and upset pressure readings | Inspect joints; replace gaskets as advised |
| Soot from short runs | Off-load starts give cool, sooty exhaust | Filter loads without regeneration; back-pressure rises | Swap short off-load runs for loaded exercise |
| Wet stacking | Unburnt fuel and oil collect in the exhaust | Oily deposits are harder to clear than dry soot | Periodic runs at meaningful load |
| Sensor and controller drift | Long power-off periods; sensors rarely see real values | False alarms or missed faults on the day it matters | Keep controller powered; compare readings at exercise |
| Stale baseline | Last good back-pressure reading is months old | Operators cannot tell normal from abnormal | Record differential pressure (ΔP) at each loaded run |
Planning by set role
Plan by role: a set running 400 hours a year and one running 20 need different attention, even with identical hardware.
Role 1: the always-on lead set
The lead set starts on every outage and carries the base load. Its RECD sees regular, sustained operation, which is generally favourable: higher exhaust temperatures, easier soot burn-off, and plenty of data for reading back-pressure trends.
- Maintain by hours and ΔP trend, as described in our RECD maintenance guide.
- Watch for long low-load spells (for example overnight with only lighting), which still produce cool exhaust.
- Use its data to calibrate the bank. The lead set shows how this RECD model behaves on this engine; use that to set realistic alarm thresholds for the other sets.
Role 2: the frequent second set
The second set runs on most long outages and busy days. It stays healthy if its runs are not mostly short.
- Watch run length, not just hours. A set that keeps starting for 10 or 15 minutes as load briefly crosses a threshold will load its RECD with soot. Review load-sharing and start/stop settings; our guide to RECD selection for frequent short start-stop cycles explains the effect.
- Make it the first candidate for swapping roles with the lead.
Role 3: the rare third set
The third set starts only on heavy days or long outages, perhaps a few dozen hours a year. Here idle-period risks start to bite.
- Exercise under load if it has not had a sustained run for a month.
- Bring it into rotation if it is identical to the first two.
- Run pre-season checks before the summer peak and before the winter GRAP period in Delhi NCR: joints, sensors and controller logging.
Role 4: the emergency reserve
The reserves run only when the whole bank is needed, perhaps two or three times a year. Their condition is the least visible, so they carry the most risk.
- Exercise under load on a fixed calendar, entered in the maintenance system with a named owner.
- Keep the RECD controller and sensors powered while the engine is idle, if the design allows. A controller silent for three months is an unknown. Our article on RECD controller and heater power supply requirements covers the options.
- Record a ΔP baseline at every exercise run, at a known load.
- Script the all-units event: who starts which set, what operators check in the first 30 minutes, and what happens if a reserve RECD alarms. Our RECD alarm response matrix for operators is a good template.
Compliance across the bank
Start from one principle: every set installed and available for use should be treated as needing to comply, unless your SPCB has confirmed otherwise in writing.
- List every set with its consent reference, rating, engine details and emission status. Identical sets can still have different engine histories.
- Confirm the rule for your state and region. Requirements can vary by capacity band, installation date and region. Our comparison of state-wise pollution control norms for DG sets is a starting point, and the CPCB website publishes national notifications.
- Verify type approval for each RECD against the engine it is fitted to. See how to verify a CPCB-tested RECD for your DG set.
- Keep records per set, not per bank: certificate, installation date, test reports, maintenance and exercise logs. Inspectors look at individual sets.
- Avoid informal assumptions that reserves are exempt. If a set is kept purely as a spare and is not to be used, discuss its status with the SPCB and document the outcome.
Readiness for the rare all-units event
The all-units event is the reason the bank exists, and the day every weakness shows up together.
Before: a quarterly readiness check
- All RECD controllers powered, logging and healthy.
- Last loaded exercise for each reserve within its planned interval, with ΔP and exhaust temperature recorded.
- No open alarms or overdue RECD maintenance on any set.
- Spare sensors, gaskets and clamps on site.
- Operators know the expected ΔP range for each set at typical event loads.
- RECD supplier contact details and response commitments current.
During: the first hour
- Check each set’s load share. If the bank is oversized for actual demand, five sets may each carry too little load. Where the control system and your reliability rules allow, fewer sets at higher load with one in hot standby may suit the RECDs better. Our guide on RECDs with synchronised generators and load sharing covers the interaction.
- Compare reserve-set ΔP with its last exercise baseline at similar load.
- Note visible smoke at start-up; it should clear as the engine warms.
After: the debrief
Log run hours, peak ΔP and alarms for each set, and use this real-load data to refresh baselines.
Exercise and loaded test regimes
Exercise is the most effective control for idle-set risk. The aim is to bring engine and exhaust to normal operating temperature so the RECD sees conditions close to real operation. Follow your engine OEM and RECD supplier’s guidance, but a typical run looks like this:
- Warm-up at light load for a few minutes.
- Loaded period at a meaningful share of rated load (50% or more is often suggested for diesel sets, though your OEM may say otherwise) for 30–60 minutes.
- Readings of load, exhaust temperature upstream of the RECD, ΔP and controller alarms, taken at a stable point.
- Cool-down at light load, as the OEM advises.
In a bank, the simplest load source is usually the site itself: run the reserve in parallel during a planned grid isolation or test window. Where that is not possible, hire a resistive load bank. Keep start checks (a few minutes, to prove batteries and starter) separate from exercise runs in your schedule, and do not count one as the other.
Rotation within the bank
Rotation changes which set takes each role during the year, spreading hours and wear. For RECD planning:
- Rotate on a fixed interval (monthly or quarterly) rather than leaving it to operators.
- Rotate among identical sets where possible, since different engines or RECD models need different baselines.
- Move the maintenance interval with the role. A set promoted from reserve to lead needs shorter intervals.
- Decide whether reserves join the rotation. Including them adds hours but reduces idle time; excluding them keeps hours low but makes exercise essential.
Monitoring idle sets
A powered RECD controller on a stopped set can show that sensors read plausible static values and that the log is intact, so data is captured from the first second of a start. If the site has a BMS or SCADA, show every set’s RECD status on the same screen, idle or running. Flag long silences in a reserve set’s log rather than ignoring them, and keep controller clocks aligned with the generator runtime log using a consistent format such as ISO 8601, so event-day data can be compared across sets.
Worked example (illustrative): a bank of 5 × 1,010 kVA sets
The numbers below are illustrative only, chosen to show the method. Replace them with your own load logs.
The site
A manufacturing plant has five identical 1,010 kVA sets (A to E), each fitted with a CPCB type-approved RECD on a common bus. The outage pattern produces this annual demand:
- One set needed for 420 hours (lead role).
- A second set for 150 hours.
- A third set for 45 hours.
- All five together during 3 events of about 6 hours each, so 18 hours for each of the fourth and fifth sets. These 18 hours fall within the third set’s 45.
Current practice: fixed roles (A always lead, E always last) and a weekly 15-minute off-load start on every set, which is 52 × 0.25 = 13 hours of off-load running per set per year.
Current annual hours per set
| Set | Role | Operational hours | Off-load test hours | Loaded exercise hours | Total |
|---|---|---|---|---|---|
| A | Lead (fixed) | 420 | 13 | 0 | 433 |
| B | Second (fixed) | 150 | 13 | 0 | 163 |
| C | Third (fixed) | 45 | 13 | 0 | 58 |
| D | Reserve | 18 | 13 | 0 | 31 |
| E | Reserve | 18 | 13 | 0 | 31 |
| Total | 651 | 65 | 0 | 716 |
Check: 420 + 150 + 45 + 18 + 18 = 651. Off-load 5 × 13 = 65. Total 651 + 65 = 716.
What it shows:
- Set A carries 433 of 716 hours, about 60% of bank running.
- For D and E, 13 of 31 hours (about 42%) are off-load runs that never heat the RECD.
- D and E see real load only at the 3 events, so their RECDs go roughly four months between meaningful runs.
The recommended plan
1. Rotate lead, second and third roles among A, B and C. Each spends four months in each role. Assuming outages are spread evenly (a simplification, as real outages are often seasonal), each takes a third of the combined demand: (420 + 150 + 45) ÷ 3 = 615 ÷ 3 = 205 hours.
2. Replace weekly off-load runs with loaded exercise. Brief start checks continue as the OEM advises but are not counted as exercise.
- D and E: a 1-hour loaded run in 8 months and a 2-hour run in the other 4 (one per quarter): 8 × 1 + 4 × 2 = 16 hours.
- A, B and C: a 1-hour loaded run in each of the 4 months spent as third set, if no sustained run has occurred that month: up to 4 hours.
3. Add one full-bank test a year of about 1 hour before the highest-risk season, all five sets synchronised on as much site load as practical, topped up with a load bank if needed.
Recommended annual hours per set
| Set | Role pattern | Operational hours | Loaded exercise | Full-bank test | Total |
|---|---|---|---|---|---|
| A | Rotating | 205 | 4 | 1 | 210 |
| B | Rotating | 205 | 4 | 1 | 210 |
| C | Rotating | 205 | 4 | 1 | 210 |
| D | Reserve | 18 | 16 | 1 | 35 |
| E | Reserve | 18 | 16 | 1 | 35 |
| Total | 651 | 44 | 5 | 700 |
Check: operational 3 × 205 + 2 × 18 = 615 + 36 = 651, unchanged. Exercise 3 × 4 + 2 × 16 = 12 + 32 = 44. Full-bank 5 × 1 = 5. Total 651 + 44 + 5 = 700.
Before and after
| Measure | Current | Recommended |
|---|---|---|
| Total bank hours | 716 | 700 |
| Off-load test hours | 65 | 0 (brief start checks only) |
| Loaded exercise and test hours | 0 | 49 |
| Highest single-set hours | 433 (A) | 210 (A, B, C) |
| Longest gap without a loaded run (D, E) | About 4 months | About 1 month |
| Sets with a fresh ΔP baseline before each event | Lead only | All five |
Bank hours fall slightly because 65 off-load hours are replaced by 49 loaded ones (44 + 5). More importantly, every non-event hour on the reserves is now useful: hot exhaust, meaningful sensor values and a recorded baseline. Wear across A, B and C evens out, so their RECD service intervals line up. If D and E joined the rotation as third set for part of the year, their exercise needs would fall but their hours would rise, a trade-off to settle with the maintenance team.
RECD maintenance schedule by role
With roles defined, write the schedule per role rather than one interval for the whole bank. Your supplier’s instructions take priority; adjust intervals using your own ΔP trends.
| Task | Lead and rotating sets | Rare third (not rotated) | Emergency reserves |
|---|---|---|---|
| Visual check of RECD, joints, clamps | Weekly | Monthly | Monthly and before each exercise |
| Controller status and log review | Daily via BMS or weekly | Weekly | Weekly, even when idle |
| ΔP and exhaust temperature record | Every long run | Every run | Every exercise and event |
| Loaded exercise | Rarely needed | Monthly if no sustained run | Monthly, longer each quarter |
| Cleaning or filter service | By ΔP trend and hours | By ΔP trend; check after many short runs | By ΔP trend; inspect at least yearly |
| Sensor check against reference | Six-monthly | Yearly or pre-season | Yearly and pre-season |
| Gasket and fastener inspection | Supplier interval | Yearly | Yearly; more often in coastal or humid sites |
| Full-bank test | Yearly | Yearly | Yearly |
Record every task against the individual set. Our sibling post on building an RECD asset register (linked below) offers a template for this.
Cost phasing without compliance gaps
Fitting five large sets is a significant spend. Phasing should follow your compliance obligations, not replace them. If your sets already need to comply, the programme must cover every in-use set within the required timeline. Where the SPCB has allowed a phased timeline, or some sets are genuinely not yet covered, a sensible order is:
- Sets that run most (the lead and rotation pool).
- Sets likely to run during GRAP or other restricted periods.
- Reserve sets, fitted before the next high-risk season.
Do not run an unfitted reserve set on the assumption that low hours make it acceptable. If a timeline cannot be met, raise it with the SPCB and keep a written record. Our guide to ranking DG sets for a phased RECD rollout shows how to set priorities while keeping backup capacity available.
Budget running costs by role too. Lead and rotating sets take most of the cleaning spend; reserves carry a fixed cost for exercise fuel, operator time and possibly load bank hire. For an identical bank, one stock of spare sensors, gaskets and clamps can serve all five sets.
FAQ
Do reserve DG sets need an RECD if they run only a few hours a year?
In most cases, yes. Regulators generally look at whether a set is installed and available for use, not its annual hours. Rules vary by state, capacity and region, so confirm with your SPCB and do not assume an exemption without written confirmation.
How often should a reserve set with an RECD be exercised?
A common approach is a loaded run of 30–60 minutes about once a month, with a longer run each quarter, but follow your OEM and RECD supplier’s guidance. The goal is normal exhaust temperature and a recorded baseline, not just proof that the engine starts.
Are weekly no-load test runs bad for an RECD?
Frequent short off-load runs tend to produce cool, sooty exhaust and condensation, loading the filter without regeneration. Keep start checks brief and add loaded runs. Our article on whether an RECD suits a DG set that runs only occasionally discusses this further.
What should operators check when all units start during a major outage?
Confirm each RECD controller is healthy, compare each set’s ΔP with its last exercise baseline at similar load, watch for smoke that persists after warm-up, and log alarms against the response matrix for the debrief.
Plan the whole bank, not just the busiest set
Aceget supplies CPCB type-approved Retrofit Emission Control Devices for DG sets from 25 kVA to 2500 kVA, including multi-set banks where some units run daily and others wait for the rare all-units event. Review our certifications or contact our team to talk through your bank’s roles, running pattern and compliance timeline.
Related Reading
- Can a Shared Spare RECD Support a Fleet of Generators?
- How to Build an RECD Asset Register for a Multi-Site DG Fleet
- RECD Upgrades During a Factory Expansion: What Should Be Reassessed?
- Lead-Lag Generator Rotation and RECD Operating Conditions
- Sequencing RECD Installations for N+1 Generators
- RECD Commissioning with Load Bank Testing