How to Calculate HVAC Energy Savings

The building blocks

Every HVAC savings estimate rests on four quantities. Get them right and the rest is arithmetic.

• Capacity in tons. One ton of refrigeration is 12,000 Btu per hour of heat removal. Rooftop and split units often list capacity in MBH (thousands of Btu per hour); divide by 12 for tons.
• Efficiency in kW/ton. Kilowatts of electrical input per ton of cooling delivered. Lower is better. Equipment rated in EER (Btu per watt-hour) converts with kW/ton = 12 ÷ EER, so EER 10 equals 1.2 kW/ton.
• Equivalent full-load hours (EFLH). The number of hours the equipment would have to run at full capacity to use its actual annual energy. It folds part-load operation into one number.
• Electricity price. Separate the energy charge ($/kWh) from the demand charge ($/kW per month). They behave differently and should be calculated separately.

If these terms are new, kW/ton, EER, SEER2 and COP explained covers the metrics in more depth.

Step 1: Estimate baseline energy use

Baseline annual kWh = tons × EFLH × kW/ton

Example: a 100-ton system with 2,500 EFLH at 1.2 kW/ton uses 100 × 2,500 × 1.2 = 300,000 kWh per year for cooling.

Sanity-check this number against your bills. If the building used 600,000 kWh in total last year, a 300,000 kWh cooling estimate is plausible for a cooling-dominated climate. If it used 350,000 kWh, the cooling estimate is too high, and one of the inputs (usually EFLH) needs revising.

Where the inputs come from

kW/ton. Best: measured, from a power meter on the equipment and a calculation of delivered cooling (chilled-water flow and temperature difference, or airflow and enthalpy change). Next best: trend data from the building automation system. Fallback: rated values. Federal purchasing tables show what new equipment achieves at rated conditions: roughly 0.50 to 0.61 kW/ton for full-load-optimized water-cooled centrifugal chillers, 0.56 to 0.73 kW/ton for water-cooled positive-displacement chillers, and EER 10.890 to 10.964 (about 1.1 kW/ton) for air-cooled chillers. Packaged air conditioners of 65,000 Btu/h and larger carry minimum EERs between 9.5 and 12.2 depending on size, or roughly 0.98 to 1.26 kW/ton. Older equipment and equipment with degraded heat transfer typically perform worse than their ratings.

EFLH. Texas program calculations for residential central air conditioners use deemed values of 1,142 (Amarillo), 1,926 (Dallas), 2,209 (Houston), 2,958 (Corpus Christi) and 1,524 (El Paso) hours. Commercial values depend heavily on schedule and internal loads. A better approach for your building is to derive EFLH from your own data: annual cooling kWh ÷ (tons × kW/ton).

Step 2: Estimate energy savings

There are two equivalent ways to express the savings, depending on what you know.

Percentage method: kWh saved = baseline kWh × % improvement.

Efficiency-difference method: kWh saved = tons × EFLH × (kW/ton before − kW/ton after).

The second form is more useful when you have measured before-and-after efficiency, because it forces you to state both values. A move from 1.20 to 1.08 kW/ton is a 10% improvement; on the 100-ton example that is 100 × 2,500 × 0.12 = 30,000 kWh per year.

Savings can also come from reducing hours (scheduling, setpoint changes) or reducing load (envelope, lighting, ventilation). Those measures interact: if you cut the cooling load, a later efficiency improvement saves a percentage of a smaller number. When several measures are planned, calculate them in sequence rather than adding percentages.

Step 3: Convert energy savings to dollars

Energy $ saved = kWh saved × marginal energy rate

The marginal energy rate is what you pay for one more, or one fewer, kWh: the energy charge plus per-kWh riders and per-kWh delivery charges in your tariff. Do not divide the total bill by total kWh. That average includes demand charges, customer charges and taxes that do not scale with kWh. For context, EIA reports an average Texas commercial price of 8.47 cents per kWh for July 2026, compared with 14.53 cents nationally; those averages blend all charges across all commercial customers.

Step 4: Estimate demand savings separately

Many commercial tariffs bill for the highest demand (kW) recorded in the billing period. NREL's survey of more than 10,000 U.S. tariffs found that about 5 million of roughly 18 million commercial customers can take service on rates with demand charges above $15 per kW, so this part of the bill can be material.

kW reduction at peak = tons operating at the peak hour × (kW/ton before − kW/ton after)

Demand $ saved = kW reduction × demand rate ($/kW-month) × number of months the cooling system sets the peak

Two cautions. First, the savings only count if the HVAC load is running at the moment the building's peak is set; if a production line or a kitchen sets the peak, HVAC efficiency may not move it. Second, some tariffs include ratchets, where a past peak sets a minimum billed demand for months afterward, which delays the benefit. Read your tariff before you count demand savings.

Utility programs also value peak demand. Texas utility incentives, for example, are paid per kW of verified peak demand reduction and per kWh of energy savings, which is why programs ask for both numbers.

Worked example: an air-cooled chiller

Illustration only. The equipment, rates and improvement below are hypothetical.

1. Baseline energy: 250 × 3,000 × 1.25 = 937,500 kWh/year.
2. Energy saved: 250 × 3,000 × (1.25 − 1.15) = 75,000 kWh/year.
3. Energy dollars: 75,000 × $0.09 = $6,750/year.
4. Peak kW reduction: 225 × 0.10 = 22.5 kW.
5. Demand dollars: 22.5 × $8.00 × 5 = $900/year.
6. Total: about $7,650/year, of which demand is about 12%.

Step 5: Show the uncertainty

A single number invites false precision. Present a low, expected and high case by varying the inputs you are least sure of, usually the improvement percentage and EFLH. In the example, an improvement of 4% instead of 8% halves the savings; 3,500 EFLH instead of 3,000 raises them by about 17%. If the decision changes across that range, you need better data before you decide, which is exactly what a measured pilot provides.

Step 6: Adjust measured results for weather and operations

Once a project is installed, comparing this summer's bill with last summer's is not a savings calculation. Weather, occupancy and operating changes move consumption. The accepted approach, set out in the International Performance Measurement and Verification Protocol (IPMVP) and in ASHRAE Guideline 14, is to build a baseline model from pre-retrofit data (for example, kWh against cooling degree days or outdoor temperature), then compare post-retrofit consumption with what the model predicts for the same conditions. ASHRAE Guideline 14-2023 includes procedures for both whole-building and retrofit-isolation approaches, and an annex on calculating the uncertainty of the savings. See baselines and weather normalization for a plain-language version.

Deemed savings vs. measured savings

Utility programs often use deemed savings: pre-approved values per unit of equipment published in a technical reference manual. The Texas Technical Reference Manual, updated each year, contains deemed savings for nonresidential measures in Volume 3 and standardized M&V protocols in Volume 4; Version 13.0 applies to program year 2026. Deemed values are convenient for common replacements. For measures without a deemed value, programs use custom calculations or measurement, covered in custom utility rebates and M&V.

Common mistakes that inflate estimates

• Using nameplate hours instead of EFLH. Multiplying full-load kW by run hours overstates consumption for equipment that cycles or modulates.
• Using average cost per kWh. Total bill ÷ total kWh overstates the value of energy savings on demand-billed accounts.
• Adding percentages from several measures. Two 10% measures do not save 20%; the second acts on a smaller baseline.
• Counting demand savings that do not occur at the peak. Check what sets your billing peak and whether a ratchet applies.
• Comparing raw bills across years. A cooler summer can look like a successful project.
• Using a vendor percentage as an input without testing. Any claimed percentage is an upper or typical figure from other equipment, not a measurement of yours.

Where an oil-fouling treatment fits in the math

A treatment aimed at internal heat-transfer surfaces would show up in the kW/ton term: if heat exchangers transfer heat better, the compressor does less work per ton delivered. CryogenX4 describes its product as lifting insulating oil film from coil surfaces and returning it to the compressor sump, and improving the lubricity of the existing oil. The company reports energy savings of up to 30% and a typical payback on the treatment of 12 to 36 months; results vary by equipment condition. To put your own number in the formulas, measure kW/ton before and after on representative units. See what oil fouling is and how HVAC energy savings are measured.

Next step

Run your own inputs through the HVAC savings calculator, then take the annual savings figure to payback, ROI and NPV to see whether the project clears your organization's investment threshold.