The Chiller Retrofit With a Moving Baseline
A 22-year-old, 650-ton centrifugal chiller serving a 180,000 ft² office tower is being replaced with a new 600-ton variable-speed magnetic-bearing chiller. The owner wants IPMVP Option B (retrofit isolation, key parameter measurement) and a 12-month savings report. Available data: 15-minute interval whole-building kWh for 3 years, 15-minute chiller kW submetering for 2 years, condenser water supply/return temperatures, and monthly fuel bills for the gas boiler. During the pre-install year the building ran 24/7 for a tenant that has since vacated, and occupancy is now 6 a.m.–7 p.m. weekdays. The new chiller has a different rated capacity and a different part-load curve. Budget allows either continuous kW metering of the new chiller plus a Btu meter on chilled water, or periodic spot measurements with a portable ultrasonic flow meter and clamp-on thermocouples.
1. Which measurement boundary and metering package best satisfies Option B here, and what do you do about the pre-install occupancy change?
Read the worked answer
The retrofit-isolation boundary is the chiller plant (chiller plus its auxiliaries as defined), not the whole building. Because the new machine has a different capacity and part-load curve, the baseline must be expressed as a function of load, not as a fixed annual kWh - you need the independent variable (tons or Btu) to survive the capacity change. That argues for continuous chilled-water Btu metering plus continuous kW at the new chiller, and reconstruction of the pre-install load-kW relationship from the 2 years of interval chiller data, not from the whole-building series.
The 24/7-to-weekday occupancy shift is a baseline adjustment issue, not a savings number. Options: (a) restrict the baseline to periods with operating profiles comparable to the post-install period, (b) fit a model with a schedule/occupancy indicator, or (c) adjust the baseline using the tenant load. Document the choice. Spot measurements cannot capture a variable-speed machine's part-load behavior across the year and will not support a defensible Option B claim - they are a fallback for a much smaller, constant-load system.
2. The new chiller's measured kW/tons at design is worse than the vendor's curve. Does that invalidate the savings claim?
Read the worked answer
No. M&V measures savings against the adjusted baseline, not against the vendor's performance curve. If the new chiller underperforms its curve, that reduces measured savings relative to a curve-based expectation, but the IPMVP savings claim is baseline energy minus reporting-period energy with routine and non-routine adjustments. The right move is to report the measured kW/tons and flag the discrepancy to the owner as a commissioning/warranty item, separate from the M&V report. Mixing a performance guarantee into the savings calculation without saying so is a common error - keep them as two documents.