IPMVP Options A, B, C and D Explained

Key takeaways

• Options A and B isolate the affected equipment; Option C uses the whole-facility meter; Option D uses a calibrated computer model.
• Option A measures key parameters and estimates the rest; every estimate must be documented and its effect on savings shown.
• Option B measures all the parameters that drive the equipment's energy use and is often the best fit for chiller and HVAC-R measures.
• FEMP recommends Option C only when savings exceed about 10% to 15% of metered consumption, with at least 12 months of baseline data.
• Option D needs calibration to measured data; ASHRAE Guideline 14 tolerances are commonly used.

The International Performance Measurement and Verification Protocol (IPMVP), published by the nonprofit Efficiency Valuation Organization (EVO), gives four ways to determine savings, labeled Options A through D. The current edition is IPMVP Core Concepts 2022. The federal government's M&V procedures, the FEMP M&V Guidelines Version 5.0, use the same four options. If a vendor says its results follow IPMVP, the first question is: which option?

The options differ mainly in where the measurement boundary is drawn. Options A and B are retrofit-isolation methods that put the boundary around the affected equipment. Option C puts it around the whole facility meter. FEMP notes that Option D can be used either way but is usually applied at whole-facility level.

The four options at a glance

Option A: key parameter measurement

Under Option A, savings come from field measurement of the parameter or parameters that define the energy use of the affected system. EVO says parameters not measured are estimated from historical data, manufacturer specifications or engineering judgment, and that the source of each estimate must be documented. EVO also requires the plan to show the range of possible savings across the range of plausible values for the estimates. FEMP calls an unmeasured value a stipulated value, agreed between the parties.

The classic Option A example in both EVO and FEMP is a lighting retrofit: measure fixture wattage, estimate operating hours. FEMP notes that while every technology can be verified with Option A, a simple lighting retrofit will typically be estimated more accurately with it than a more complicated chiller retrofit.

For HVAC-R: Option A can work when the efficiency change can be captured by a spot test and the load is reasonably predictable. Its weakness is that cooling loads vary with weather and occupancy, so a single spot test of kW per ton may not represent the whole year. EVO adds two Option A plan requirements: justification of all estimates, and periodic inspections to confirm the equipment is still in place and operating as assumed. FEMP's 4.0 edition notes that ASHRAE Guideline 14 has no direct parallel to Option A.

Option B: all parameter measurement

Option B measures energy use (or reliable proxies for it) and the variables that drive it, for the affected equipment only. EVO's example is a variable-speed drive on a pump motor, with a kW meter reading power every minute: in place for a week during the baseline to confirm constant loading, then throughout the reporting period. FEMP says Option B is typically more precise than Option A, though it can be more difficult and costly.

For HVAC-R, Option B is often the natural fit. FEMP 5.0 gives a chiller replacement evaluated with Option B: because chiller load depends heavily on outdoor weather, a model correlating outdoor air temperature with chiller kW is used to make routine adjustments. For chiller plant improvements, FEMP 5.0 recommends Option B with continuous measurement of key parameters, plant efficiency in kW per ton recorded during the baseline, post-installation and performance periods, and that efficiency normalized for load and weather.

FEMP lists conditions where Option B suits well:

• Independent variables that affect energy use are not too complex or expensive to monitor.
• Operating data on the equipment are available through control systems.
• Submeters already exist for the systems in question, such as a separate HVAC submeter.

An equipment-level treatment, such as one that changes heat transfer inside coils, maps naturally onto Option B: meter the treated equipment, measure the conditions that drive its load, and compare like with like.

Option C: whole facility

Option C uses the utility meter (or a sub-facility meter) and statistical analysis, usually regression against weather and other variables, to separate savings from everything else affecting the bill. It is attractive because the utility bill is what the owner pays. FEMP 5.0 is cautious about it, noting that changes in the treated area can be obscured by changes elsewhere, that adjustments for occupancy, mission and plug loads grow over time, and that whole-facility analysis is "not always a very accurate method of estimating savings."

FEMP 5.0 recommends Option C only when the project meets these criteria:

• Savings are predicted to be greater than about 10% to 15% of the consumption on the meter, on a monthly basis.
• At least 12 (and preferably 24 or more) months of pre-installation data are available.
• At least 9 (preferably 12) months of performance-period data are used to calculate annual savings.
• Good data on independent variables exist, and changes to the facility are unlikely or easily quantified.

FEMP warns that, unless the model is very highly predictive, smaller percentage savings risk being "lost in the noise" with monthly data. It also suggests Option C may suit a short-term use of two to three years, after which M&V can switch to Option A or B. In DOE's Uniform Methods Project, the whole-building protocol is consistent with Option C and relies on billing (consumption) analysis, often across many buildings with comparison groups.

For HVAC-R: Option C is a good choice when a project is large relative to the building's total use, for example a plant-wide program in a building where cooling dominates the bill. It is a poor choice for verifying a measure on a handful of units in a large mixed-use facility.

Option D: calibrated simulation

Option D uses building energy simulation software, calibrated against measured data, to estimate baseline and post-change energy use. EVO says it is suited to cases where baseline or reporting-period data are unreliable or unavailable, and that it requires considerable skill. FEMP's example is a comprehensive retrofit with several interacting measures in a large building, with the model calibrated to at least 12 months of utility billing data.

Calibration is judged statistically. FEMP 5.0 reproduces the ASHRAE Guideline 14-2023 tolerances:

MBE shows whether the model runs consistently high or low. CV(RMSE), the coefficient of variation of the root mean square error, shows how far individual periods scatter from actual values. EVO's Option D requirements add that the plan should name the software and version, provide the input, output and weather files, and report the calibration accuracy achieved.

How to choose: a decision guide

FEMP lists the factors behind the choice: the value of the measure, its complexity, the number of interrelated measures, and the risk that savings will not be achieved. It also notes that unproven technologies may warrant additional attention. That is a fair expectation for any product category where buyers have reason to be skeptical, including refrigerant additives.

Operational verification comes first

Whatever the option, EVO says operational verification should come before savings are counted: confirming the measure is installed, commissioned and doing what it is meant to do. EVO's approaches range from visual inspection and sample spot measurements to short-term performance testing and trending data through the building automation system, typically over a few days to a few weeks.

"IPMVP-adherent" versus "IPMVP-certified"

EVO states that IPMVP does not currently provide formal certification of project-specific M&V plans. What exists is adherence: a plan is adherent if it meets EVO's published criteria, which cover the facility overview, measure intent, chosen option and boundary, baseline, reporting period, basis for adjustment, calculation method, energy prices, meter specifications, monitoring responsibilities, expected accuracy, budget, report format and quality assurance. EVO also notes that savings can be called IPMVP-adherent only for reporting periods that actually used adherent procedures. Projecting future savings from a past result does not count. The full checklist appears in how to read an M&V report.

Applying this to CryogenX4

CryogenX4 is a one-time treatment for existing HVAC-R equipment, installed while the system runs. According to the company, its effect is on heat transfer and lubrication inside the treated equipment, which is why a retrofit-isolation approach is the logical way to verify it. The field method the company describes measures minute-level energy use, chilled-water or air flow, indoor and outdoor temperature and humidity at the equipment, before and after treatment, and compares periods with similar weather and load. That is close in structure to Option B. The plan for any specific site should still state the option, the boundary, the adjustment method and the expected accuracy in writing. See what a CryogenX4 pilot measures for the data points.

Next step

List the equipment you want to evaluate, note which already has power metering or BAS trends, and estimate what share of your utility bill it represents. That short exercise usually settles the choice between Option B and Option C. To discuss an equipment-level pilot on your systems, contact CryogenX4.

Frequently asked questions

Which IPMVP option is most common for chiller projects?

For chiller plant improvements, FEMP 5.0 recommends Option B with continuous measurement of key parameters and kW per ton recorded during the baseline, post-installation and performance periods, normalized for load and weather.

Can Option A be used for HVAC measures?

Yes, but it relies on estimated values for anything not measured, such as annual operating hours or load. EVO requires each estimate to be justified and its effect on the savings range shown. It is more accurate for simple, steady loads than for variable cooling loads.

Why do whole-facility methods need so much data?

Regression models must capture a full seasonal cycle. FEMP recommends at least 12 months, and preferably 24 or more, of baseline data and at least 9 months of reporting data, and recommends building models from whole-year sets so seasonal effects are not overstated.

What does CV(RMSE) mean in an M&V report?

It is the coefficient of variation of the root mean square error, a measure of how far a model's predictions scatter from measured values. Lower is better. For calibrated simulation, ASHRAE Guideline 14 tolerances are 15% with monthly data and 30% with hourly data.

Sources

1. IPMVP Generally Accepted M&V Principles (October 2018) — Efficiency Valuation Organization (EVO)
2. International Performance Measurement and Verification Protocol (IPMVP) — Efficiency Valuation Organization (EVO)
3. Release of the new IPMVP Core Concepts 2022 — Efficiency Valuation Organization (EVO)
4. M&V Guidelines: Measurement and Verification for Performance-Based Contracts, Version 5.0 (September 2024) — U.S. Department of Energy, Federal Energy Management Program
5. M&V Guidelines: Measurement and Verification for Performance-Based Contracts, Version 4.0 (November 2015) — U.S. Department of Energy, Federal Energy Management Program
6. Uniform Methods Project, Chapter 8: Whole-Building Retrofit with Consumption Data Analysis Evaluation Protocol (NREL/SR-7A40-68564, 2017) — U.S. Department of Energy / National Renewable Energy Laboratory
7. Guideline 14-2023: additional material (contents, retrofit-isolation metering and whole-building model forms) — ASHRAE Journal