By

In the last edition of the magazine, we explored why heat loss calculations are critical for heat pump retrofits — highlighting emerging empirical heat loss calculation methods that use utility billing data as a simple option for contractors.

The logic is straightforward: energy use rises and falls with a home’s heat loss, so billing data can be used to infer heat loss. This follow-up article reviews ongoing research validating these methods, practitioner insights, and early case study results from real homes.

A national validation effort for empirical heat loss methods is already underway. Jérémie Leger, project lead with

(NRCan) local energy efficiency partnerships (LEEP) team, explains, “Leveraging real-world data for heat loss and heat gain calculations presents a significant opportunity to improve the sizing and selection of heat pumps in Canada’s retrofit market. We have recognized its potential and are now conducting a large-scale research study on Canadian homes to assess the accuracy of these methods. The objective is to provide insights to the CSA F280 standards committee and, if methods are proven reliable and accurate, incorporate these emerging approaches into the next versions of the standard.”

The research will take time, but contractors and energy advisors are already using empirical methods in the field, and their experience offers early insights.

Michelle Hjort, managing partner at

, has worked on more than 300 retrofits in the Greater Toronto Area (GTA). She uses both EnerGuide audit design heat loss (DHL) values and empirical estimates from utility bills. Hjort finds heat loss estimates from empirical methods to be consistently lower and relies on them for heat pump sizing, while using EnerGuide DHL values for backup system sizing, “… a significant barrier to heat pump uptake is lack of confidence in correct sizing. The empirical heat load sizing methods are a game changer for contractors and HVAC coaches to quickly and easily determine heat load based on real-world data.”

Sarah Grant, founder of

, has audited over 1,500 homes, with more than 1,000 heat pump installs. Her main tool is Volta SNAP, which produces a CSA F280-compliant DHL calculation from EnerGuide audit data. Where relevant, she adds empirical calculations from utility bills. “Sometimes the empirical and F280 data align while other times they are grossly different with the empirical methods coming in lower. It seems that most often the differences occur with homes that aren’t well insulated.”

Grant notes that uneven temperature distribution in older homes may be a contributing factor to the discrepancy. Both experts see empirical methods as simpler for contractors but stress understanding their limitations.

Overall, a key insight from the field is that empirical methods produce lower heat loss estimates. This bears out in early observations from case studies.

Supported by

(TAF), the

(TRCA)

(STEP) tested its free empirical heat loss tool, ThermalPoint, in six GTA homes using data from a previous heat pump study, which is titled Replacement of Air-Conditioners with Cold Climate Air-Source Heat Pumps. Data included pre- and post-retrofit utility bills, EcoBee thermostat runtime data, EnerGuide audits, and equipment specs.

Homes were typically medium-sized, 1960s to 2000s construction, detached or semi-detached, with cold-climate heat pumps (two to three tons) backed up by existing furnaces. Two homes had small wood stoves (Homes 5 and 6); others had no additional heat sources. Indoor setpoints ranged from 19 C to 23 C. Details on homes and equipment are in Figure 2.

Four values were compared for each home, including empirical heat loss, which was estimated by ThermalPoint using pre-retrofit gas bills, actual heat loss calculated from furnace runtime data, EnerGuide audit DHL, and furnace maximum capacity.

Furnace runtime was treated as the closest approximation to the actual real-world heat loss and an independent benchmark to compare against. The runtime analysis went as follows. Taking Home 2 as an example (data in Figure 2), a single-stage Lennox G41UF-36B-070 furnace (60 kBtu/h output) ran 34 per cent of the time to maintain the indoor temperature near 23 C on a day with an average outdoor temperature of -17 C. The furnace provided the equivalent of 20 kBtu/h (34 per cent times 60 kBtu/h), and this was assumed to be equal to the heat loss. The heat pump was off during this time.

Results were adjusted to the design temperature assumed in the EnerGuide audit (-18.3 C). For two-stage furnaces, a range was calculated since the operating stage wasn’t logged. The calculation was repeated for both low- and high-stage outputs, although low-stage was likely closer to reality because the furnaces were typically oversized.

Results are shown in Figure 3. The empirical heat loss prediction based on the pre-retrofit gas bills tends to agree with the actual heat loss determined independently from the furnace runtime data. The EnerGuide audit DHL was greater than the actual heat loss to varying degrees, sometimes by a factor of two times. In all cases, furnaces were on the scale of two to three times larger than needed.

The larger heat loss determined from the EnerGuide audits can impact heat pump sizing. For example, considering Home 1, NRCan’s air-source heat pump sizing and selection web app predicts that furnace back-up is required for outdoor temperatures below -3 C, if assuming the EnerGuide heat loss. If assuming the empirical heat loss, it’s -12 C.

The difference is large and makes the heat pump appear less capable than it is in practice. Furthermore, the post-retrofit thermostat data showed that the heat pump could indeed meet the load in the colder temperatures predicted by the empirical heat loss.