M&V Baseline Adjustment for a Chiller Retrofit
A university retrofits a 1,200-ton chilled-water plant with variable-speed drives and new controls under an energy performance contract. The baseline period had 1,000,000 ton-hours and 1,200,000 kWh, giving 1.20 kWh/ton-hour. The first post-retrofit year has 1,150,000 ton-hours and 1,050,000 kWh. Campus enrollment grew 12%, and the M&V plan uses IPMVP Option B with a baseline regression of plant kWh against ton-hours and cooling degree days; the regression has an R² of 0.91 and a 10% uncertainty at 90% confidence. The contract guarantees 20% savings and requires annual verification. The owner wants to claim savings using simple before-after kWh difference.
1. Compute the simple before-after savings and the savings normalized to post-retrofit ton-hours using the baseline specific consumption. Explain why they differ.
Read the worked answer
Simple before-after: 1,200,000 - 1,050,000 = 150,000 kWh, or 12.5% of baseline. This assumes the campus cooling load was unchanged.
Baseline specific consumption = 1,200,000 / 1,000,000 = 1.20 kWh/ton-hour. Expected post-retrofit energy at the actual 1,150,000 ton-hours, if efficiency had not changed, = 1.20 × 1,150,000 = 1,380,000 kWh. Normalized savings = 1,380,000 - 1,050,000 = 330,000 kWh, or 23.9% of adjusted baseline (330,000 / 1,380,000).
They differ because the simple method ignores the 15% increase in ton-hours. The campus did more cooling in the post year; the simple claim credits the retrofit only for the net kWh drop while attributing the added load to the retrofit. Under Option B, savings must be baseline-adjusted for measurable independent variables such as ton-hours and weather.
2. Does the normalized result satisfy a 20% savings guarantee? What role does the stated 10% uncertainty play?
Read the worked answer
The guarantee is 20% of baseline: 0.20 × 1,200,000 = 240,000 kWh/year. The normalized savings estimate of 330,000 kWh exceeds that by 90,000 kWh. With ±10% uncertainty at 90% confidence, the interval is roughly 297,000–363,000 kWh if the 10% applies to the savings estimate; even the low end exceeds 240,000 kWh. So the point estimate supports the guarantee, but the verifier should not simply declare victory.
The uncertainty statement must be tied to the regression, not to a hand calculation. Check the regression standard error, whether ton-hours and CDD are independent enough, and whether the 10% is absolute or relative. If the guarantee is based on measured savings after adjustments, report the adjusted value with its confidence interval and follow the contract's risk-sharing rules. If the guarantee is a fixed 20% of baseline, the reported value matters mainly for the guarantee settlement and any shortfall payment.
3. List the verification checks you would perform before accepting the 330,000 kWh result.
Read the worked answer
- Confirm the metering boundary: plant kWh and ton-hours must cover the same chillers, pumps, and cooling towers in both periods.
- Verify baseline regression form and validity: coefficients, R², autocorrelation, and whether CDD or ton-hours are double-counted.
- Check post-period data quality: missing intervals, meter calibration, and any operational changes such as new building loads or setpoint changes.
- Confirm non-routine adjustments: enrollment growth should show up in ton-hours; do not adjust for it twice.
- Review the Option B plan for independent variable measurement and uncertainty method.
- Reconcile with utility bills and verify no fuel switching or added chillers outside the boundary.
- Apply the contract's baseline and reporting rules, then document the adjusted savings and confidence interval.