Calibrating Savings on a Steam-to-Hot-Water Retrofit
A 60-unit affordable property has 12 months of utility data: 78,000 therms/year total gas, of which the owner believes 55% is space heating and 30% is domestic hot water, with the remainder miscellaneous. The building has a central atmospheric boiler at 78% steady-state efficiency, uninsulated distribution in unconditioned crawl spaces, and a two-pipe distribution system. The proposed scope is a 92% condensing boiler, pipe insulation on 1,200 feet of bare mains, and a setpoint reset. The owner's consultant has proposed a 38% gas savings figure based on "boiler efficiency improvement plus pipe losses." The utility's weather-normalized baseline differs from the raw 12-month total by about 6%. The client needs a defensible savings number for a program application.
1. Critique the 38% figure. What is wrong with the consultant's method, and what savings would you defend?
Read the worked answer
The 38% figure is almost certainly inflated because it compounds savings that overlap and ignores interactive effects. Boiler efficiency improvement from 78% to 92% is a combustion-efficiency gain, but it applies only to the fuel that actually reaches the boiler. If distribution losses are large, the boiler is already firing more to compensate, and insulating the pipes reduces the load at the boiler. You cannot apply the full efficiency gain to the full baseline and then add pipe savings on top; the pipe savings reduce the input to which the efficiency gain applies.
A defensible approach: start from weather-normalized space-heating gas, not raw totals. Apply the 6% normalization adjustment to the baseline. Then estimate distribution loss as a fraction of heating input, typically 10-25% for uninsulated mains in unconditioned space, and apply the pipe insulation as a reduction of that loss, not of total gas. Then apply the efficiency gain only to the remaining boiler input, and account for part-load: a condensing boiler only reaches 92% with low return-water temperatures, which requires reset and adequate radiation. Without verified low return temperatures, a realistic seasonal efficiency might be 85-88%, not 92%.
A defensible range is roughly 15-22% total gas savings, with the DHW portion largely unaffected unless the DHW is also upgraded. The application should state assumptions, the normalization method, and the interactive-effect treatment explicitly.
2. The utility's weather-normalized baseline differs from the raw total by 6%. How do you handle this in the savings claim, and what data would you request?
Read the worked answer
The difference indicates the 12-month period was not an average weather year. Savings claims should be made against a normalized baseline so that a mild or severe year does not distort the result. The correct handling is to use the normalized baseline for the savings calculation and to disclose the method and the degree-day data source.
Data to request: at least 24-36 months of gas and electric billing to establish a stable baseline; daily or monthly HDD from the nearest representative station; the actual meter read dates, since estimated reads create step changes that look like savings; occupancy and vacancy history, because a 60-unit affordable property can swing 5-10% in occupancy; and any major equipment or envelope changes during the baseline period. Also request the boiler's firing rate and run hours if a burner log exists, and the DHW setpoint and storage temperature.
If only 12 months are available, the claim should be presented as a range with the normalization uncertainty stated, and the program should require post-installation measurement and verification rather than paying on modeled savings alone. A 6% baseline shift is larger than many of the individual measures being claimed, so it must be resolved before the application is submitted.