Energy baseline adjustment worksheet

What the worksheet calculates

Most compressed-air improvement projects are judged by comparing two electricity totals: one month before the work and one month after. That comparison is rarely fair. If production fell by a fifth after a leak campaign, energy will fall too, and the project gets credit it did not earn. If a new line started, energy may rise even though every cubic metre of air is now produced more efficiently.

The worksheet solves this by building a simple model of the pre-improvement system and then running it forward under the reporting period’s conditions. It produces four numbers that should always be shown together:

• Measured baseline energy – the actual kWh recorded during the baseline period.
• Adjusted baseline energy – the kWh the old system is predicted to have used under reporting-period conditions.
• Measured reporting-period energy – the actual kWh after the change.
• Avoided energy – adjusted baseline minus reporting-period energy, with its uncertainty and any documented exclusions.

Keeping the raw totals visible beside the adjusted figures matters. A plant manager can then see both what the meter said and why the adjusted comparison differs.

Which inputs does an energy baseline need?

A baseline is only as good as the variables that genuinely drive compressor energy. For a typical Indian plant the useful candidates are listed below. Not every site needs all of them; the test is whether a variable explains a real, physical part of the variation.

The baseline and reporting periods must use the same measurement boundary. If the baseline included the refrigerated dryer and cooling-tower pump but the reporting meter covers only the compressors, the comparison is broken until the boundary is reconciled.

Method: how the adjustment is calculated

The worksheet follows the general approach used in recognised measurement and verification practice, such as the International Performance Measurement and Verification Protocol (IPMVP) and ISO 50015 / ISO 50006 for energy performance indicators. The steps are:

1. Define the baseline period. Choose a period that covers the normal range of production and seasons – often 12 months, or at least several weeks of logged data covering every shift pattern.
2. Clean the data. Remove meter faults and periods with known abnormal events, and record every exclusion and why it was made.
3. Fit a model. Usually a linear regression: E = a + b × production + c × hours (+ d × ambient temperature where justified). The constant a represents base load such as leakage, idling and dryer running.
4. Check the fit. Look at the coefficient of determination (R²), the coefficient of variation of the root-mean-square error (CV(RMSE)) and the residuals. Guidance such as ASHRAE Guideline 14 sets statistical acceptance criteria; the worksheet reports these values rather than hiding them.
5. Apply routine adjustments. Insert reporting-period values of each variable into the baseline model to obtain the adjusted baseline.
6. Apply non-routine adjustments. Add or subtract documented, separately estimated effects – for example a new paint line that would have added air demand to the old system too.
7. Calculate avoided energy. Adjusted baseline minus measured reporting energy, presented with its uncertainty.

Where the plant has a stable, narrow operating pattern, a simpler ratio method may be acceptable: baseline kWh per unit of production multiplied by reporting-period production. The worksheet should state which method was used and why.

Worked example

A pharmaceutical packaging plant near Hyderabad logs compressor-station energy and production for twelve baseline months. Regression gives:

Monthly kWh = 38,000 + 0.85 × packs (thousands) + 120 × production days

The 38,000 kWh constant is large, which already hints at significant leakage and unloaded running. The plant then completes a leak repair campaign and lowers generation pressure by 0.5 bar. In the first reporting month it packs 24 million packs (24,000 in the thousands used by the formula) over 26 production days, and the meter records 52,400 kWh.

• Adjusted baseline = 38,000 + 0.85 × 24,000 + 120 × 26 = 38,000 + 20,400 + 3,120 = 61,520 kWh
• Avoided energy = 61,520 − 52,400 = 9,120 kWh for the month

Had the plant compared the reporting month with the average baseline month of, say, 57,000 kWh, it would have claimed only 4,600 kWh, because production happened to be higher. The adjusted comparison credits the project with the correct, larger figure – and in a month of low production the same method would prevent an inflated claim.

The financial value is then a separate step: avoided kWh multiplied by your own contracted tariff, handled in the project economics tools rather than inside the baseline.

How to interpret the result

An adjusted saving is a well-supported estimate, not a proof that the project alone caused the change. Read it with these questions in mind:

• Is the saving larger than the model’s uncertainty? A saving smaller than the typical monthly scatter is not yet demonstrated.
• Do the residuals show a trend? A steadily drifting baseline may mean leakage was growing before the project, or that a meter was recalibrated.
• Did specific power (kW per m³/min) improve as well as total energy? If flow fell but specific power worsened, the controls may not be following the lower demand.
• Are reporting-period conditions inside the range used to build the model? Extrapolating to a production level never seen in the baseline is weak evidence.
• Have all non-routine events been recorded? An undocumented shutdown or new machine can make or break the result.

Common mistakes we see in baseline comparisons

Limits of the worksheet

The worksheet adjusts energy comparisons; it does not size compressors, set safe operating pressures, certify air quality or produce a carbon inventory. It cannot rescue poor measurement: if station energy is estimated from nameplate kW rather than measured, the adjusted saving inherits that weakness. It also assumes the old system’s behaviour stayed stable through the baseline. If a compressor was failing during the baseline months, the model will include that inefficiency, and a later “saving” may partly be the effect of a repair. Record such events and treat the result accordingly.

For electrical measurement, uncertainty and carbon conversion, use the dedicated worksheets linked below, so that each step stays traceable.

How Reddy Compressors helps

Reddy Compressors has served industry from Tarnaka, Secunderabad, since 2002, and our team of more than 20 service engineers and technicians covers the whole of Telangana and Andhra Pradesh. As an independent multi-brand specialist we work on ELGi, Atlas Copco, Ingersoll Rand, KAESER, Chicago Pneumatic and other makes, so our view of a baseline is not tied to selling one brand.

For energy baselines we can log compressor power, pressure and, where a suitable point exists, flow; review your production and meter data; help you choose sensible baseline variables; and repeat the measurement after improvement work so that both periods share the same boundary. The improvements themselves – leak repair, pressure reduction, control changes, compressor replacement or right-sized rental cover during changeover – are carried out by our service teams. Our audits follow the ISO 11011 method, although we do not issue conformity certificates.

What to send us

• Monthly or daily compressor-station energy data for at least the last year, with meter locations
• Production output and production days for the same periods
• Compressor makes, models, ratings and working pressure settings
• Dryers, filters and auxiliaries inside the metered boundary
• A list of changes already made or planned, with dates
• Your site location and contact person

Frequently asked questions

What is an adjusted energy baseline for compressed air?

It is an estimate of the energy the original compressed-air system would have used under the reporting period’s production, hours and site conditions. It is built from a model of baseline data, usually a regression on production and operating time. Comparing it with metered energy after an improvement gives a like-for-like saving instead of a simple before-and-after bill comparison.

How long should a compressed-air energy baseline period be?

Ideally it covers a full year so that summer heat, monsoon humidity and seasonal production changes are all represented. Where that is not practical, several weeks of logged data covering every shift pattern and production mix can work, provided the reporting period stays within the same range of conditions. The worksheet should state the chosen period and why.

Why did my electricity bill go up after a leak repair?

Usually because something else changed: higher production, more working days, hotter weather or a new air user. A raw bill comparison cannot separate those effects. Adjusting the baseline for production, hours and temperature shows whether the leak repair reduced energy for the same service. Also check that compressor controls actually unload or slow down when demand falls.

Which variables should be used to normalise compressor energy?

Use only variables that physically drive air demand or compressor efficiency: production output, operating hours or shifts, ambient temperature and, where it changes, generation pressure. Test each one statistically and keep it only if it improves the model meaningfully. Too many variables on a short dataset produce a model that fits the past and predicts the future badly.

Can a baseline model prove that a project saved energy?

It provides strong supporting evidence but not absolute proof. The saving is credible when it clearly exceeds the model’s uncertainty, the measurement boundary did not change, all non-routine events are recorded and specific power improved as well. Periodic re-checks against the same model show whether the saving persists or fades as leaks and settings drift back.

Does Reddy Compressors measure energy before and after improvements?

Yes. Our engineers can log compressor power and pressure, and flow where a suitable measuring point exists, before improvement work and again afterwards under the same boundary. We serve sites across Telangana and Andhra Pradesh from our Secunderabad office and work on most compressor brands. Send us your energy and production records to start.

Related pages

• Measurement, verification and energy performance
• Motor input and electrical measurement worksheet
• Measurement uncertainty worksheet
• Power and flow logging
• Compressed-air energy optimisation
• Compressed air knowledge and calculation centre