Floating Head Pressure — M&V Regression Framework & Sensitivity Analysis

Abriliam Consulting — Industrial Energy Management

This notebook demonstrates regression-based measurement and verification (M&V) for floating head pressure savings using simulated pre/post compressor power data, aligned with IPMVP Option B/C methodology.

Key outputs:

Pre/post regression model comparison
Verified annual savings calculation
Sensitivity analysis on key assumptions

1. System Parameters (Same as Notebook 01)

We reuse the same system configuration for consistency.

2. Simulate Pre-Retrofit Baseline Data

We simulate 12 months of hourly compressor power data under fixed head pressure control. The simulation adds realistic measurement noise and minor load variation to represent field conditions.

The regression specification follows IPMVP:

$$W_{\text{comp}} = \alpha + \beta \cdot T_{\text{amb}} + \varepsilon$$

3. Simulate Post-Retrofit Data (Tier 2 — EEV Float)

Post-retrofit data simulates floating head pressure with an EEV floor at 18°C. The compressor power is lower at cool/cold outdoor temperatures because the condensing pressure follows ambient downward.

4. Fit Pre/Post Regressions

Ordinary least squares regression of compressor power against outdoor temperature for both periods. The key diagnostic is:

$R^2 > 0.75$ for adequate model fit
Post-retrofit slope $\beta' < \beta$ (less sensitive to ambient)

5. Figure 3 — Pre/Post Regression Lines (M&V Demonstration)

6. Verified Savings Calculation

Verified savings are calculated as the area between the pre-retrofit and post-retrofit regression lines, integrated over actual post-implementation weather:

$$\Delta E_{\text{verified}} = \sum_{j=1}^{M} \left[(\alpha + \beta \cdot T_{\text{amb},j}) - (\alpha' + \beta' \cdot T_{\text{amb},j})\right] \cdot \Delta t$$

7. Residual Analysis

Check regression residuals for systematic patterns that would indicate confounding variables or model specification issues.

8. Sensitivity Analysis

Examine how savings vary with key assumptions:

Minimum condensing temperature floor
Electricity rate
Condenser approach temperature

9. Summary & Conclusions

The regression-based M&V framework demonstrates:

Pre/post regression isolates floating head pressure savings from other variables
R-squared > 0.80 is achievable for constant-load cold storage facilities
Savings are most sensitive to the minimum condensing temperature floor — each 3°C reduction adds approximately 5-8% savings
Simple payback under 1 year for Tier 1, 2-3 years for Tier 2 at Ontario industrial electricity rates

IPMVP Compliance Checklist

Pre-retrofit baseline regression with R² > 0.75
Post-retrofit regression with comparable data quality
Verified savings calculated as area between regression lines
Residual analysis confirms no systematic bias
Sensitivity analysis quantifies uncertainty bounds