What sequencing changes, and why it needs evidence
On a station with two or more compressors, the controls decide which machines run, which one trims, and how much time is spent loaded, unloaded or stopped. Poor coordination is common: overlapping pressure bands, two machines part-loaded together, a variable-speed unit pushed to full speed while a fixed-speed unit sits unloaded, or a standby that never stops. Each of these wastes power without delivering air.
A sequencer or master controller can fix much of this, and so can carefully set local pressure bands. The difficulty is proving it. Vendor simulations, a controller screenshot or one good day on the trend screen show intent, not measured performance across real shifts. A project record should show the dispatch pattern before and after, under comparable demand, and should say clearly what was measured and what was modelled.
Drawing the system boundary
Comparisons fail when the boundary moves between periods. Before logging, we define it from each compressor’s electrical incomer and package outlet to the common receivers, treatment and the point where demand is measured. The record should state:
- Every compressor in the boundary, including standby, rental and backup machines.
- Interconnecting valves, parallel plants and any air exported to or imported from other areas.
- Whether measured power includes only compressor packages or also dryers, cooling fans, drains and room ventilation.
- Where delivered air is measured, if at all, and the meter’s accuracy and installation.
- Where the controlling pressure signal is taken, since sequencers and local controllers may read different points.
If a dryer was added to the metered circuit halfway through, or a rental unit was running in one period and not the other, the before-and-after comparison is no longer like for like.
Machine inventory: what to record for each compressor
| Field | What to record | Why it matters |
|---|---|---|
| Identity | Make, model, serial, nameplate photograph | Links logs to the right machine |
| Type and control | Fixed-speed load/unload, modulating, variable-speed, oil-free or oil-injected | Each responds differently at part load |
| Rated capacity | FAD at rated pressure, motor rating | Basis for estimating delivered air from loaded time |
| Condition | Last service, filter and separator state, known faults | A worn machine may deliver less than its rating |
| Settings | Load and unload pressures, run-on timers, speed limits, priority | The core of the intervention, before and after |
| Signals logged | Running, loaded, unloaded, stopped, speed, kW, alarms | Without states, kW alone cannot explain dispatch |
Logging the baseline
A sequencing baseline must cover the full variety of demand: day and night shifts, weekdays and weekends, start-ups, batch events and any regular low-demand periods. A single afternoon rarely shows the worst behaviour, which often happens at light load when machines run unloaded for long periods.
- Synchronise clocks on all loggers and controllers, and record any offset.
- Log each compressor’s state and power, station pressure at the control point and, where possible, delivered flow.
- Choose an interval short enough to catch load/unload cycles; long averages hide short cycling.
- Record production data for the same period, so demand can be compared later.
- Note every abnormal event: trips, manual starts, maintenance, power interruptions.
From these logs we can show loaded, unloaded and stopped hours per machine, the number of starts, how many machines ran together, where the station pressure sat within its band and the station’s specific power in kW per m³/min where flow is measured.
Making the control change
The change itself should be documented as carefully as the measurements. That means recording the old and new settings for every machine, the sequencer or master controller logic (base and trim allocation, rotation, pressure bands, start and stop delays), the date and time of each change and who made it. If the logic was adjusted several times during commissioning, each version should be recorded, and the as-left period should begin only after the final settings were in place.
Safety and machine protection settings are not part of an optimisation and should not be altered to achieve it. Minimum run times, restart limits and protection functions stay as the manufacturer intends.
Comparing matched periods
The as-left period should resemble the baseline in production, shift pattern, machine availability and, as far as possible, ambient conditions. In Telangana and Andhra Pradesh, a baseline in a cool month and an as-left period at 40–45 °C summer ambient can differ for reasons unrelated to the controls, because compressor capacity and cooling change with temperature.
Where periods cannot be matched exactly, normalise. The simplest honest approach is to compare station energy per unit of delivered air, or per unit of production where air demand tracks production closely, and to state the basis. Our energy baseline adjustment worksheet sets out one way to do this.
| Observation | Measured or modelled? | How to state it |
|---|---|---|
| Fewer machines running together | Measured from state logs | Report hours with one, two or three machines running, before and after |
| Less unloaded running | Measured from state logs | Report unloaded hours per machine |
| Tighter pressure band | Measured from pressure log | Report the range and its spread at the control point |
| Lower station kWh | Measured if metered within the boundary | Report with matched or normalised basis and uncertainty |
| Annual saving | Modelled | Label as an estimate with its assumptions |
Attribution: what else changed?
Sequencing projects often happen alongside other work. Leak repairs, a lower setpoint, a new receiver, compressor overhauls or a dryer change can all reduce energy at the same time. If they occurred within the comparison window, the record should either separate their effects with evidence or report a combined result. Lowering pressure, for example, typically saves about 7% of compressor power per bar, and that saving should not be credited to the sequencer.
Also consider changes the other way: a new production line or a failed drain valve can raise demand and mask a genuine control improvement. A record that lists confounders openly is far more credible than one that claims a clean result.
Common findings on multi-brand stations
Many Indian plants run a mix of makes and ages: perhaps an older fixed-speed screw, a newer variable-speed unit and a piston or rental machine for peaks. Each controller was set up in isolation, so the station behaves as several independent machines. Typical findings include overlapping bands that keep two machines part-loaded, a variable-speed unit set to base load while a fixed-speed machine trims, standby units left in automatic and starting needlessly, and long unload timers on lightly loaded machines. These can often be improved with careful local settings before any new controller is considered, and the record should say which approach was used.
How Reddy Compressors helps
We have worked on compressed-air stations since 2002 from our office at Tarnaka, Secunderabad, with service teams across Telangana and Andhra Pradesh and project support elsewhere in India where feasible. Because we are independent and multi-brand, servicing ELGi, Atlas Copco, Ingersoll Rand, KAESER, Chicago Pneumatic, Sullair, Doosan, Kirloskar and others, we can look at how mixed machines work together rather than optimising one brand in isolation. We are not an authorised dealer of any of these makes. We have tie-ups with all major and reputed brands.
For a sequencing review we can survey the machines and their settings, check compressor condition so a worn unit is not mistaken for a control problem, log states, pressure and power, recommend band and priority changes or a vendor-neutral sequencing approach, and help log the as-left period. Where demand needs covering during changes, we can supply a machine from our own rental fleet. Any results we publish are anonymised and approved by the plant.
What to send us
- A list of all compressors with make, model, rating and nameplate photographs.
- Current load/unload pressures, timers and priority settings on each machine.
- Details of any existing sequencer or master controller.
- A sketch showing receivers, dryers and where pressure is measured.
- Shift pattern and typical production schedule.
- Any existing logs of running hours, loaded hours or energy.
Frequently asked questions
How do I know if my compressor sequencing is wasting energy?
Look at loaded and unloaded hours on each machine, how often two or more run part-loaded together, and whether a variable-speed unit sits at full speed while a fixed-speed machine unloads. Long unloaded running, frequent starts and a wide station pressure band are common signs. Logging each machine’s state and power alongside station pressure for a representative week usually shows the pattern clearly.
Does a master controller always save energy?
Not automatically. A master controller can coordinate machines well, but the saving depends on the mix of compressors, the demand profile, storage, pressure bands and how the logic is set. On some stations careful local settings achieve much of the benefit. The only way to know for a given plant is to compare matched before-and-after logs of machine states, delivered air and station kW.
What data is needed to prove a sequencing improvement?
Synchronised logs of each compressor’s state and power, station pressure at the control point and, ideally, delivered flow, over comparable periods before and after the change. Add production data, a record of every setting change and a list of other work done in the same window. With these, dispatch changes can be measured directly and energy changes can be stated with an honest basis.
Should the variable-speed compressor be the base or the trim machine?
Usually the trim machine, because its efficiency advantage comes from following changing demand while fixed-speed units run fully loaded as base. The variable-speed unit’s usable turndown should cover the typical demand swing between base steps. If it is too large or too small for that swing, the station may still cycle machines. The right allocation depends on the plant’s measured demand profile.
Can we mix compressors of different brands under one sequencer?
Often yes. Many vendor-neutral sequencers control machines through pressure signals or simple run, load and fault contacts, and some can read richer data from individual controllers. Compatibility depends on each machine’s controller and available interfaces. Whatever the method, machine protection settings must stay as the manufacturer intends, and the station needs a safe fallback if the sequencer fails.
How long should we log before and after a sequencing change?
Long enough to include every regular demand pattern: day and night shifts, weekdays and weekends, start-ups and any low-demand periods. For many plants that means at least a full production week each side, but it depends on how much the operation varies. The as-left period should start only after commissioning adjustments have finished, so the record reflects the final settings.
Related pages
Calculators and worked examples
These tools give indicative figures for screening, comparison and for preparing an enquiry. They run in your browser and send nothing. No prices or tariffs are assumed: enter your own to see costs. Efficiencies and other figures shown are example values: replace them with your own. Results are not a performance guarantee or a design certificate; final selection is confirmed by a service engineer against the machine data sheet and the site.
Adjusted energy baseline and avoided energy
Apply a baseline model (from your regression) to the reporting period’s production and days, then compare with the metered energy to get a like-for-like avoided energy figure.
The worked example below shows the method with fixed numbers. Turn on JavaScript to change the inputs.
Worked example
Inputs
- Model constant a (kWh per period): 38000
- Coefficient b (kWh per unit of production): 0.85
- Coefficient c (kWh per production day or hour): 120
- Coefficient d, 0 if not used (kWh per degree of mean ambient temperature): 0
- Reporting period: production (same units as b): 24000
- Reporting period: production days or hours (same basis as c): 26
- Reporting period: mean ambient temperature, used only if d is not 0 (°C): 30
- Non-routine adjustment, + or − (kWh): 0
- Measured reporting-period energy (kWh): 52400
- Average measured baseline-period energy, for the raw comparison (kWh): 57000
Result
- Production term b × production: 20,400 kWh
- Days/hours term c × days: 3,120 kWh
- Temperature term d × temperature: 0 kWh
- Adjusted baseline energy: 61,520 kWh (incl. non-routine 0 kWh)
- Measured reporting-period energy: 52,400 kWh
- Avoided energy (adjusted): 9,120 kWh = 14.8% of the adjusted baseline
- Raw comparison (baseline average − reporting): 4,600 kWh
- Difference made by the adjustment: 4,520 kWh
Method. Adjusted baseline = a + b × reporting production + c × reporting days (or hours) + d × reporting mean ambient temperature + non-routine adjustment. Avoided energy = adjusted baseline − measured reporting-period energy. The raw comparison (average baseline period − reporting period) is shown alongside, so both what the meter said and the adjusted result stay visible. Coefficients come from a regression on baseline data (see the fit tool below or your own analysis); set d = 0 if temperature is not used. Money is a separate step with your own tariff.
What to take from it. A saving smaller than the model’s normal period-to-period scatter is not yet demonstrated; and a reporting period outside the baseline range of production is weak evidence.
Technical content reviewed by D. Srinivas Reddy, Managing Partner and service engineer, Reddy Compressors. Last updated 17 October 2026. How our engineering content is reviewed.
Load/unload cycle energy
Turn measured load and unload times (or the logged percentage of time in each state) and measured package kW into average power, annual energy, the energy spent running unloaded, and specific power based on the air actually delivered. An optional second set of times compares a change.
The worked example below shows the method with fixed numbers. Turn on JavaScript to change the inputs.
Worked example
Inputs
- Times entered as: Measured cycle times (seconds)
- Loaded package power (kW): 82
- Unloaded package power (kW): 28
- Loaded time (s per cycle, or percent): 40
- Unloaded time (s per cycle, or percent): 60
- Stopped time, motor off, 0 if none (s per cycle, or percent): 0
- Power while stopped, controls only (kW): 0
- Running hours per year with this cycle pattern: 6000
- Free air delivered when loaded, FAD (m³/min): 12.5
- After change, optional: loaded time, all three 0 = not used (s or percent): 0
- After change, optional: unloaded time (s or percent): 0
- After change, optional: stopped time (s or percent): 0
Result
- Cycle time: 100 s (36 cycles per hour)
- Loaded fraction: 40 %
- Unloaded fraction: 60 %
- Unloaded power as a share of loaded power: 34.1 %
- Average package power: 49.6 kW
- Annual energy: 2,97,600 kWh per year
- Unload energy (buys no air): 1,00,800 kWh per year = 33.9% of the annual energy
- Loaded specific power: 6.56 kW per m³/min
- Average air delivered: 5 m³/min
- Effective specific power over the cycle: 9.92 kW per m³/min
- Energy cost per year: Enter your electricity tariff (cost per kWh) above
Method. Two input modes. Cycle timing: enter measured seconds per cycle. Fraction: enter the share of time loaded, unloaded and stopped (for example 40, 60, 0, from controller loaded/running hours); no cycle time is reported in this mode. Average kW = (P_load × t_load + P_unload × t_unload + P_stopped × t_stopped) ÷ (t_load + t_unload + t_stopped); the states must not overlap. Annual kWh = average kW × the hours this pattern runs. Unload energy = P_unload × unloaded fraction × hours. Average air delivered = loaded fraction × loaded FAD; effective specific power = average kW ÷ average air delivered. Unloaded power is never assumed to be zero: enter measured package kW (unloaded read after blowdown has settled). Scenario comparison: enter the predicted times after a change (bigger receiver, wider band, auto-stop) in the “after change” fields; the same powers, FAD and hours are used and the kWh difference is shown. An auto-stop timer is modelled by moving unload time into the stopped state (check the motor’s permitted starts per hour). For different day and night patterns, run the tool once per pattern and add the results. The rapid-cycling warning below 30 s per cycle is a screening value used by this tool. Money is shown only when you enter your own tariff.
What to take from it. A large unload-energy share, or an effective specific power well above the loaded figure, means a fixed-speed machine is running lightly loaded: look at storage, pressure band, sequencing, auto-stop or a smaller/VSD trim machine.