Fundamental Performance Verification Approaches
IPMVP-style M&V offers four option families. Option A (retrofit isolation, key parameter measurement) measures the parameters that most affect savings and stipulates the rest; Option B (retrofit isolation, all parameter measurement) measures all relevant parameters continuously or periodically. Option C (whole facility) uses whole-building utility meters and typically a regression model. Option D (calibrated simulation) uses a calibrated energy model of the whole facility or system.
Selection depends on: expected savings magnitude relative to M&V cost, uncertainty tolerance, budget, and whether interactive effects matter. A common rule of thumb is that M&V cost should be a small fraction of the savings - often cited as roughly 1–10% of project savings, with larger projects tolerating lower percentages. Higher rigor (B, D) is justified when savings are large, disputed, or when performance contracting requires it.
Traps: assuming Option A is always cheaper and acceptable - it requires justification of which parameters are stipulated and evidence they won't change. Confusing Option B with Option C: B isolates the measure with dedicated meters, C uses the whole-facility meter. Believing Option D requires no field data - calibration against measured data is essential. Also, questions may ask which option to use when interactive effects (e.g., lighting affecting cooling load) are significant: whole-facility or calibrated simulation captures them better than isolation.
Worked example: A chiller retrofit with large expected savings and significant interaction with building cooling load, in a building with 12 months of hourly whole-building data. Option C with a weather-normalized regression is defensible and low-cost; Option B with dedicated BTU metering is more rigorous but costlier. If savings are small relative to total building load, Option C may lack resolution, pushing toward B.