Methodology: Billing-Aligned Energy Performance Analysis

Version: 1.2.1 Last Updated: January 2026

Overview

This document describes the Fully Billing-Aligned methodology used to reconcile asynchronous utility meter read dates with calendar-year weather normalization and performance tracking. This approach addresses a fundamental challenge in residential energy auditing: utility billing cycles rarely align with calendar months, creating ~15-day phase shifts that corrupt weather-normalized efficiency calculations.

---

The Billing Misalignment Problem

Standard Industry Practice (Flawed Approach)

Most residential energy analyses simply assign utility bills to calendar months:

Utility Bill: Dec 15, 2024 → Jan 14, 2025 (846 kWh)
Standard Assignment: "January 2025: 846 kWh"
Problem: Bill includes 16 days of December consumption!

Consequences:

• Weather normalization uses wrong HDD values (January HDD vs. actual Dec 15-Jan 14 HDD)
• Year-over-year comparisons become meaningless if billing dates shift
• Cannot accurately isolate seasonal HVAC loads

Billing-Aligned Solution

Our methodology allocates consumption based on actual meter read dates and calculates daily rates from adjacent billing periods for pro-ration.

---

Core Methodology Components

1. Billing Period Decomposition

Process:

For each utility bill, extract:

• Meter Read Date (Start): Previous bill end date
• Meter Read Date (End): Current bill end date
• Consumption: kWh or CCF for the period
• Duration: Days between read dates

Daily Rate Calculation:

Daily Rate = Total Consumption ÷ Days in Billing Period

Example:

Bill Period: Dec 15, 2024 → Jan 14, 2025 (31 days)
Consumption: 846 kWh
Daily Rate: 846 ÷ 31 = 27.3 kWh/day

2. Calendar Month Allocation

Process:

For each calendar month, sum consumption from all overlapping billing periods using pro-rated daily rates:

January 2025 Consumption =
  (Dec 15-31, 2024 bill × 14 days) +
  (Jan 1-14, 2025 from same bill × 14 days) +
  (Jan 15-31, 2025 from next bill × 17 days)

Calculation:

Jan 2025 = (27.3 kWh/day × 14 days) + (next bill daily rate × 17 days)

3. Weather Normalization Alignment

HDD Calculation for Billing Periods:

Once consumption is allocated to calendar days, match with corresponding HDD values from NOAA daily weather data:

HDD65 for Jan 2025 = Σ(max(0, 65°F - T_avg)) for Jan 1-31

Key Advantage: HDD values now precisely match consumption periods, enabling accurate calculation of heating intensity:

Heating Intensity = Space Heating CCF ÷ (HDD ÷ 1,000)

---

Space Heating Isolation

Method 1: Low-HDD Regression (DHW Baseline)

Objective: Determine DHW consumption independent of space heating

Process:

1. Select Low-HDD Billing Periods: April-October when HDD < 200/month
2. Calculate Daily Gas Consumption: CCF/day for each low-HDD period
3. Linear Regression: Regress daily CCF against daily HDD to find intercept
4. DHW Baseline: Y-intercept represents zero-HDD consumption (pure DHW)

Example:

Low-HDD Periods (2025):
  Apr: 187 HDD, 1.02 CCF/day
  May: 98 HDD, 0.87 CCF/day
  Jun: 12 HDD, 0.68 CCF/day
  Sep: 45 HDD, 0.71 CCF/day

Regression: Daily CCF = 0.533 + (0.0087 × HDD/day)
DHW Baseline = 0.533 CCF/day

Annual DHW Calculation:

Annual DHW = 0.533 CCF/day × 365 days = 194.5 CCF
Billing-Aligned Total (2025): 188 CCF (accounts for billing period boundaries)

Method 2: High-HDD Allocation (Space Heating)

Objective: Subtract DHW baseline from total gas to isolate space heating

Process:

1. Calculate Monthly DHW: 0.533 CCF/day × days in month
2. Subtract from Total Gas: Space Heating = Total - DHW
3. Sum Annual Space Heating: Aggregate across all months

Example (January 2025):

Total Gas (Jan 2025): 112 CCF
DHW Baseline: 0.533 × 31 days = 16.5 CCF
Space Heating: 112 - 16.5 = 95.5 CCF

---

Electricity Decomposition

Baseload Determination

Method: Average consumption during minimum-load periods (spring/fall shoulder seasons)

Process:

1. Identify Low-Load Months: April, May, October (minimal HVAC operation)
2. Calculate Daily Average: kWh/day for each month
3. Baseload Value: Mean of shoulder-season daily rates

2025 Calculation:

April: 11.2 kWh/day
May: 9.8 kWh/day
October: 8.1 kWh/day
Baseload = (11.2 + 9.8 + 8.1) ÷ 3 = 9.7 kWh/day

HVAC Load Isolation

Cooling Load (Summer):

Monthly Cooling kWh = Total kWh - (Baseload × days) - Furnace Blower kWh

Heating Season Blower:

Blower kWh = Furnace Runtime Hours × Blower Power Draw

Example (July 2025):

Total July kWh: 892 kWh
Baseload: 9.7 × 31 = 301 kWh
Furnace Blower: 0 kWh (no heating)
Cooling: 892 - 301 = 591 kWh

Residual Category

Definition: Consumption not attributed to Baseload, AC, or Furnace Blower

Calculation:

Residual = Total Annual kWh - Baseload - AC - Blower - Modeled Dehumidifier

Diagnostic Value: Residual is concentrated in April-October, suggesting unmeasured seasonal load (likely dehumidifier excess beyond modeled baseline).

---

Building UA Calculation

Input Parameters

1. Total Delivered Heat (MMBTU):
   • Furnace: Space Heating CCF × BTU/CCF × AFUE
   • Fireplace: Independently measured via fuel consumption × efficiency
   • Historical values below use 100k BTU/CCF × 0.96 AFUE; going-forward standard is 103,700 BTU/CCF × 0.95 AFUE (submittal-verified)

2. Heating Degree Days (HDD): Weather-normalized demand

3. Balance Point: Temperature below which heating is required

Calculation Steps

Step 1: Calculate Delivered Heat

Furnace Delivered = 599 CCF × 100k BTU/CCF × 0.96 AFUE = 57.5 MMBTU
Fireplace Delivered = 3.6 MMBTU (measured)
Total Delivered = 57.5 + 3.6 = 61.1 MMBTU

Step 2: Adjust HDD for Balance Point

HDD65 (2025) = 6,270
Balance Point = 59°F (determined via regression)
HDD59 (2025) = 5,294 (recalculated from daily NOAA data)

Step 3: Calculate UA

UA = Total Delivered Heat ÷ (24 hr/day × HDD)
UA = 61.1 MMBTU × 1,000,000 BTU/MMBTU ÷ (24 × 5,294)
UA = 61,100,000 ÷ 127,056
UA = 481 BTU/hr-°F ≈ 480 BTU/hr-°F

Corrections (February 2026 audit):

• The AFUE value above (0.96) reflects the series marketing rating (“up to 96%”). The tested AFUE per manufacturer submittal (S9X1C100U-SUB-1D-EN) is 0.95 (95.0%).
• The BTU/CCF value above (100,000) is the simplified therm convention. The precise energy content of pipeline natural gas is 103,700 BTU/CCF.
• With corrected values (103,700 × 0.95 + 3.6 MMBTU fireplace), the equivalent UA is 493 BTU/hr-°F. Going forward, all Home Assistant sensor formulas use these corrected values. Historical calculations in this document are preserved as-published for reproducibility.

Step 4: Area Normalization

UA/ft² = 480 ÷ 2,440 sq ft = 0.197 BTU/hr-°F-ft²

Balance Point Determination

Method: HDD Neutral Band Optimization

Process:

1. Calculate UA at multiple balance points (55°F, 57°F, 59°F, 61°F, 63°F)
2. Select balance point that minimizes year-over-year UA variance
3. Validate against theoretical expectations (55-62°F typical for high-efficiency homes)

Result: 59°F balance point yields most stable UA across 2022-2025 dataset

---

Statistical Validation

Coefficient of Variation (CV)

Definition:

CV = (Standard Deviation ÷ Mean) × 100%

Application:

Used to quantify baseline stability across 4-year dataset:

Heating Intensity (CCF/1k HDD):
  Mean = 89.1
  Std Dev = 6.2
  CV = (6.2 ÷ 89.1) × 100% = 7.0%

Interpretation:

• CV < 5%: Exceptional stability
• CV 5-10%: Good stability
• CV > 10%: Investigate systematic variance

Regression Analysis

Applications:

1. DHW baseline determination (CCF vs. HDD regression)
2. Balance point optimization (UA stability vs. balance point)
3. AC power verification (kWh vs. cooling degree days)

Standard Metrics Reported:

• R² (coefficient of determination)
• Slope and intercept with 95% confidence intervals
• Residual analysis for outlier detection

---

Data Quality Controls

1. Billing Period Continuity Check

Validation:

End Date (Bill N) = Start Date (Bill N+1) - 1 day

Purpose: Detect missing billing periods or data entry errors

2. Consumption Reasonableness Bounds

Electricity:

• Daily baseload: 5-15 kWh/day
• Summer peak: 20-40 kWh/day
• Winter peak: 15-30 kWh/day

Natural Gas:

• Daily DHW: 0.4-0.7 CCF/day
• Peak winter: 3-8 CCF/day

Action: Flag any billing period outside 2× normal range for manual review

3. HDD Cross-Validation

Method: Compare NOAA KBDL data against:

• Nearby weather stations (Bradley Field, Tweed-New Haven)
• Home Assistant local temperature sensors
• Expected HDD from historical climate normals

Tolerance: ±5% variance acceptable; >5% triggers investigation

---

Limitations and Uncertainty

Known Sources of Error

1. DHW/Space Heating Separation: ±10-15% uncertainty due to seasonal inlet temperature variation
2. Fireplace Heat Contribution: ±5% uncertainty in efficiency assumption (65-70%)
3. Balance Point Determination: ±1°F uncertainty in optimal value
4. Dehumidifier Modeled Load: ±20% uncertainty (unmetered equipment)

Confidence Intervals

Heating Intensity (95% CI):

2025: 95.5 ± 8.3 CCF/1k HDD
Range: 87.2 to 103.8 CCF/1k HDD

Building UA (95% CI):

2025: 480 ± 35 BTU/hr-°F
Range: 445 to 515 BTU/hr-°F

---

Software and Tools

Data Processing Pipeline

1. Raw Data Extraction: Utility PDFs → CSV (manual or OCR)
2. Weather Data: NOAA API → daily temperature and HDD
3. Billing Alignment: Python pandas (custom script)
4. Regression Analysis: Python scipy.stats / R
5. Visualization: Matplotlib / ggplot2

Reproducibility

All analysis scripts and raw data CSV files are available in the /data directory of this repository (if applicable). Key dependencies:

pandas >= 1.5.0
numpy >= 1.23.0
scipy >= 1.9.0
matplotlib >= 3.6.0

---

Future Enhancements

Planned Improvements (v1.3.0)

1. Circuit-Level Monitoring Integration: Replace modeled dehumidifier load with actual CT measurements
2. Fireplace Efficiency Calibration: Install flue gas temperature sensor for real-time efficiency calculation
3. DHW Recirculation Loss Isolation: Independent metering of recirculation pump runtime
4. Multi-Zone Temperature Logging: Quantify cathedral ceiling stratification effects on UA

---

References

1. ASHRAE Fundamentals (2021): Chapter 19 - Energy Estimating and Modeling Methods
2. BPI Technical Standards: Building Performance Institute, Inc. - Standard 2400-S-2013
3. RESNET Standards: ANSI/RESNET/ICC 301-2019 (Energy Calculation Methodology)
4. NOAA Climate Data: National Centers for Environmental Information (NCEI) API documentation

---

Document Prepared By: William K. Collis Methodology Review Status: Validated against 4-year historical dataset (2022-2025) Last Calculation Update: January 11, 2026