UPS efficiency and operating cost: calculate losses at your actual load

Start with the operating point, not the biggest percentage

Efficiency is output power divided by input power. A system serving a given load therefore takes more input power when its efficiency is lower. The difference is the UPS loss in the simplified steady-state comparison used here.

ENERGY STAR uses specified load points and weighting to calculate its average efficiency measure. That is a certification metric, not a measurement of your particular facility. In Power Field Guide, the source-reported load-specific percentages refer to the minimum tested configuration and the lowest-input-dependency test mode. Match the configuration and mode before comparing two rows.

Sources: ENERGY STAR UPS key product criteria ↗ · ENERGY STAR Certified Uninterruptible Power Supplies dataset ↗

Worked example: 50 kW output, 95% versus 97% efficiency

Assume a hypothetical constant 50 kW output load, continuous operation for 8,760 hours and an energy rate of $0.12 per kWh. Compare two otherwise equivalent operating points at 95% and 97% efficiency. These are teaching assumptions, not measured products, a utility quote or a promise of achievable savings.

At 95%, loss = 50 × (1/0.95 − 1) = 2.632 kW. At 97%, loss = 50 × (1/0.97 − 1) = 1.546 kW. The difference is about 1.085 kW, or 9,506 kWh over the assumed year. At the assumed rate, that is approximately $1,141 per year in UPS loss electricity.

This calculation compares the difference in UPS losses. It excludes cooling, battery recharge events, demand charges and other project costs. It is useful precisely because its boundary is explicit: another reviewer can replace the assumptions and reproduce the result.

Replace one load assumption with your actual load profile

A facility rarely needs to be modeled as one unchanging load merely because the example is simple. Build a table of operating periods, expected output and applicable efficiency. Calculate loss energy for each period, then add the results. Do not take an unweighted average of percentages and expect it to reproduce the same cost.

For a hypothetical two-period model, use 30 kW for 4,000 hours and 50 kW for 4,760 hours. Obtain efficiency at each relevant point for each candidate. If you have only a few published test points, identify the gap and ask the manufacturer for the appropriate curve or calculation rather than treating a guess as measured performance.

EPA guidance explains that proportional and fixed losses can make partial-load behavior important. Use that as a reason to investigate the curve, not as a universal claim that one architecture wins at every load. Your comparison should retain the evidence for every efficiency input.

Sources: ENERGY STAR: Reduce energy loss from UPS systems ↗

Compare the same mode and required power-quality behavior

Manufacturers may publish different efficiency figures for double-conversion and higher-efficiency operating modes. Vertiv describes this distinction in its operating-mode guidance. An attractive percentage from one mode should not silently be compared with a different mode from another supplier.

Ask each bidder to name the proposed normal mode, provide its relevant performance documentation and identify which operating conditions change the mode. Have the engineer responsible for the protected load evaluate the behavior. The commercial comparison should then use the mode that will actually be approved and used.

Sources: Vertiv: High-efficiency modes of operation ↗

Make the cost boundary visible before discussing payback

Present the electricity estimate as one line in the ownership comparison. Separately record installed cost, service, battery replacement assumptions and any other costs supported by quotations. If a number is unknown, leave it unknown rather than applying a convenient market average.

If an alternative has a hypothetical $5,000 incremental installed cost and only the $1,141 annual loss-electricity reduction above, simple payback would be about 4.4 years. That is a mathematical illustration with flat prices and unchanged load; it excludes financing, discounting and differences in maintenance or replacement. It is not a return estimate for either manufacturer.

Avoid multiplying the result by a generic cooling factor and presenting the expanded total as verified savings. Ask the facility team to establish an appropriate method and boundary if cooling impacts belong in the business case. Utility demand charges likewise need the actual tariff and demand profile.

Request an efficiency comparison you can audit

A good bidder response should let you reproduce its savings claim. Ask for the source of the input data as well as the final chart. This gives the purchaser a practical way to identify a favorable assumption without arguing about whose presentation looks more polished.

• Exact configured UPS and installed power modules used for the quoted performance.
• Operating mode and efficiency curve at the proposed load range.
• Current and growth load assumptions with hours assigned to each period.
• Energy-rate source, date and treatment of time-varying charges.
• Explicit inclusion or exclusion of cooling, recharge and demand charges.
• Installed-price difference and the supporting scope for any payback statement.
• Sensitivity cases showing what happens when growth or operating hours differ.

• Compare load-specific values in the catalog ↗
• Review the rest of the UPS specification ↗

Common questions