Thermal System and Net-Zero-Carbon Least-Cost Design Optimization of New Detached Houses in Canada

Abstract

This study focused on optimizing a model house for different locations and types of thermal systems to understand better how heating system type affects thermal envelope design. The study investigated six different thermal system configurations in separate optimizations for five locations. Optimization implies reducing energy consumption, minimizing greenhouse emissions (GHG), lowering operational costs, ensuring regulatory compliance, enhancing resilience, and improving occupant comfort and health. The Pareto front, multi-objective optimization, is used to identify a set of optimal solutions, considering multiple goals that may conflict with each other. In determining the least-cost building design envelope, the design balances costs with other goals, such as energy efficiency and environmental impact. The optimizations determine the life-cycle cost versus operational GHG emissions for a single-detached house in Canadian locations with varying climates, emissions factors, and energy costs. Besides natural gas, the study evaluated four electricity-heated options: (a) an air-source heat pump, (b) a ductless mini-split heat pump, (c) a ground-source heat pump, and (d) an electric baseboard. A net-zero-carbon design with grid-tied photovoltaics was also optimized. Results indicate that the heating system type influences the optimal enclosure design. In each location, at least one all-electric kind of design has a lower life-cycle cost than the optimized gas-heated model, and such designs can mitigate the majority of operational GHG emissions from new housing in locations with a low carbon electricity supply.

1. Introduction

In Canada, over half of the homes are single-detached, single-family houses, which also account for 35% of new residential constructions [1,2]. Residential buildings consume 17% of the country’s energy and produce 14% of the annual greenhouse gas emissions (GHG). Single-detached homes, which are the focus of this study, produce 6.39% of the total GHG emissions [3]. The residential emissions come from both off-site electricity generation using fossil fuels and on-site combustion of fossil fuels for space and domestic hot water (DHW) heating. Natural gas is the most used heating fuel in Canada, accounting for 76% of households. Even though 80% of Canadians reside in homes connected to a relatively low carbon electric grid with an average annual emissions factor of 50 gCO₂e/kWh or less, the majority of GHG emissions from residential buildings come from on-site combustion of fossil fuels, primarily natural gas [4,5]. The prevalent low-emitting alternatives to natural gas and oil furnaces include electric resistance baseboard heaters and water heaters. The relatively high cost of electricity per unit of energy usually renders electric resistance heating economically punitive. On the other hand, the inconvenience of air pollution, including GHG emissions associated with gas furnaces and wood stoves, makes these options not ideal for suburban and urban environments.

Canada’s 2020 national model codes aim to build all new buildings to net-zero-energy-ready standards by 2030 [6]. New and innovative design approaches were proposed for net-zero-energy building standards to gain public and industry acceptance and become mandatory. The new strategy should maintain initial and operational costs over the existing norm. Additionally, they should be achievable and scalable in practice. Net-zero energy buildings are energy-efficient buildings equipped with state-of-the-art futuristic technologies (heat pumps, heat recovery, air-tight envelope design, etc.) and provisions to integrate renewable energy generation to balance energy demand and carbon emissions [7]. This approach is based on the fact that electricity must be generated from low and non-emitting sources, such as nuclear and renewable energy.

In order to reduce GHG emissions and the operating costs of the residential buildings, energy standards have tended to become more stringent over time. Building standards prescribe minimum levels of energy performance for specific building envelopes to control annual space heating demand and peak heating load [8,9]. To support these goals, building standards specify minimum efficiencies for heating systems, heat recovery ventilators (HRVs), DHW heaters and storage tanks, air-conditioners, and building envelopes. Electric heat pumps, with an efficiency ranging from 200% to 500% depending on the application, are a promising solution for a sustainable and economical heating transition. However, a more comprehensive study is necessary to evaluate the relative effectiveness of all of these options in the Canadian context, taking into account multiple variables [10,11].

Space heating accounts for about 60% of the energy used in residential homes in Canada. In order to minimize energy consumption and greenhouse gas emissions, the design of the building envelope should consider the local climate as well as factors such as the type of system, fuel type, and fuel price structure. Building energy modelling software (e.g., EnergyPlus, TRNSYS, eQuest, HOT2000, WUFI Plus, etc.) is a powerful design tool and extensively used in house design and performance evaluation [12,13]. The software falls into two general categories: evaluative tools and guidance tools. Evaluative tools rely heavily on expertise and trial and error [14,15]. The recommended tool for low-rise residential buildings, the case of this research, is the U.S. DOE’s BEopt program. BEopt allows users to create a building energy model and choose from various design options. BEopt uses the “modified sequential search technique” algorithm, which has been enhanced with three-speed techniques to skip certain design variants, thereby reducing overall computation time.

Bucking et al. [16,17] studied the EcoTerra house in Montreal by using BEopt. The optimization function selects energy models with a high probability of being nearer to the Pareto front. There are 25 variables available for adjustment within this optimization including PV generation. Bucking et al. have suggested future work to gain a better understanding of initial and construction costs as well as to identify the most appropriate incentives to reduce either initial or lifecycle costs.

The aim of this building design optimization study is to evaluate building performance by comparing operating costs, energy consumption, and environmental impact. The study compares a model house’s costs and GHG emissions when using electric heating options versus natural gas heating systems that comply with building codes in five major Canadian cities. These cities cover a range of geographical and climatic variations across the country. The optimization of house design takes into account factors such as natural gas and electricity prices, electricity emissions factors, local climates, and construction costs to determine the most efficient designs and the requirement of solar PV electricity to make it net-zero carbon (NZC).

2. Methodology

The combustion of fossil fuels, both on-site and off-site, leads to GHG emissions. However, it is important to note that not all energy sources have the same carbon intensities. Therefore, energy use does not always correlate to GHG emissions. On the other hand, excessive energy use can result in high energy costs, but again, different energy sources have varying costs, which may not correlate with energy use. Other equally important factors to consider include energy independence, peak electric demand implications, and the possibility of renewable electricity use.

The space heating demand intensity (SHDI) metric is of interest as a proxy measure for the overall thermal performance of the building enclosure, including envelope thermal resistance, window performance, orientation and area, air barrier system performance, HRV thermal efficiency, and occupant density. Another important metric is the Energy Utilization Intensity (EUI). The EUI is used to compare a building’s actual energy consumption with its optimal energy consumption. These metrics were selected for this research because of their critical roles in affecting the energy consumption of a building [18,19]. However, the building envelope design optimization is not solely based on energy demand (EUI and SHDI), but also other factors, which will be discussed in this study (Figure 1):

• Determine the life cycle cost (LCC) and GHG emissions of low-cost electric and zero-carbon building designs using fundamental economics, building physics, and the best available heating technologies, without subsidies or a carbon price.
• Analyze the correlation between heating system type/efficiency and optimal enclosure design (SHDI). Also, investigate the relationship between energy standard performance metrics (SHDI and EUI), homebuyer/ratepayer primary concern metrics, and electricity emission factors (EFs).

The study analysis is based on the whole-house optimized life cycle least-cost design. The resultant designs will be compared to an optimized gas-heated house design. The reference design in all locations is a gas-heated climate-specific minimum code-compliant design based on the National Building Code (NBC) 9.36 [20,21]. The optimization results have included location-specific factors such as climate, energy prices, construction costs, and electricity EFs.

2.1. Building Energy Performance Optimization

The National Renewable Energy Laboratory (NREL) created BEopt, a residential front-end to EnergyPlus [13]. This software can optimize the design of a typical single-detached house by creating various parallel building energy models. The optimized design is determined by comparing the developed models’ lifetime cost and annual GHG emissions. The objective is to achieve the global minimum, which is the minimum cost. However, for net-zero-carbon optimizations, the design with the lowest lifetime cost to achieve a Greenhouse Gas Intensity (GHGI) of 0 is manually selected from the Pareto front.

2.2. Thermal System Optimization

The study focused on optimizing a model house for different locations and types of thermal systems to understand better how heating system type affects thermal envelope design. The study investigated six different thermal system configurations in separate optimizations for each location (Figure 2). The type of heating system selection can affect GHGI, EUI, and operational costs based on the input energy source and system efficiency, which can be energy-source-dependent. For example, only heat pumps (which use electricity) can have a Coefficient of Performance (COP) greater than 1. Some configurations can also provide significant capital savings, such as those that do not require ducting or those that use a single piece of equipment to provide multiple services.

All models have a dedicated outdoor air ventilation system, which allows for the isolation of the thermal system as a set of interacting components, including space heating (both main and supplementary), space cooling, space conditioning delivery (ducting), and DHW production. By isolating this system, it becomes easier to compare it with other systems on a capital, operational, and LCC basis, as well as its effect on optimal enclosure design. The following equations illustrate the relationship between the thermal envelope, the thermal system (heating), and the purchased energy that governs the annual heating cost (operational only) [22].

Thermal Envelope:

$$AnnualHeatingDemand\left({\displaystyle \frac{{kWh}_Q}{year}}\right)=SHDI\left({\displaystyle \frac{{kWh}_Q}{m^2·year}}\right)\times FFA\left(m^2\right)$$

(1)

Mechanical Heating System:

$$PriceofHeat\left({\displaystyle \frac{\$}{{kWh}_Q}}\right)=PurchasedEnergyPrice\left({\displaystyle \frac{\$}{{kWh}_{\$}}}\right)÷\eta \left({\displaystyle \frac{{kWh}_Q}{{kWh}_{\$}}}\right)$$

(2)

Passive-Active Relationship:

$$AnnualHeatingCost\left({\displaystyle \frac{\$}{year}}\right)=AnnualHeatingDemand\left({\displaystyle \frac{{kWh}_Q}{year}}\right)\times PriceofHeat\left({\displaystyle \frac{\$}{{kWh}_Q}}\right)$$

(3)

where

${kWh}_{\$}$ = purchased energy unit (electrical or chemical)

${kWh}_Q$ = thermal energy unit

$\eta$ = system thermal efficiency (or COP in the case of heat pumps)

2.3. Building Enclosure Thermal Performance

The aim is to find the best whole-house designs by exploring various options for building enclosure components such as thermal resistance, air leakage, window type, and windows-to-walls ratio. This is done separately for each thermal system type and location. NBC 9.36 sets the lower limit for each component, while increments and upper limit options are chosen based on realistic alternatives and engineering judgment. The range and granularity of component options have been carefully selected to achieve the true Pareto front while keeping computation time to a minimum. Building enclosure option inputs remain the same for each thermal-system-type optimization.

2.4. Electrification of Heating

Only the gas optimization and the reference model utilize natural gas, while all other models are fully electric. Given the uncertainties surrounding the GHG accounting of renewable energy generation offsets, this study prioritizes direct, onsite GHG reduction that can be accurately calculated. Indirect, offsite GHG reduction is given less priority. In other words, renewable generation capacity to achieve NZC is added only to all-electric designs. This approach is motivated by the low electricity EFs in the most Canadian provinces. The relative merit of prioritizing full electrification is assessed based on location, where electricity price and EFs play a crucial role.

2.5. Onsite Renewable Energy Generation

The energy generated through renewable sources, solar photovoltaic (PV) panels mounted on rooftops, is distributed. However, there are limitations with the electricity input EFs in BEopt, which means that a separate marginal EF for PV generation cannot be inputted. As a result, the NZC optimizations are performed by comparing the LCC versus the site energy generation. The optimal design is chosen from the Pareto front based on the PV generation versus gross consumption ratio necessary to achieve NZC for the electric grid.

3. Building Energy Model

3.1. Model House Description

A single-detached, single-family house was modelled. The house represents an archetypal new suburban house in Canada, based on Ottawa’s NRC CCHT twin test houses [23]. The complex geometry includes a slab-on-grade garage, with a simplified gable style roof to allow for rooftop PV if needed (Figure 3). The design complies with NBC 9.36, except for the floor plan. The modelling software considers the number of floors and house size for modeling and the cost of ducts, DHW piping, and ventilation. The occupancy is determined using the Building America House Simulation Protocol [12]. The house includes an unfinished basement within the thermal envelope and air control layer, but it is not directly conditioned or ventilated due to software limitations. Whole-slab insulation for basements is not required by NBC 9.36. The house’s back has more glazing than the front due to a south-facing sliding glass patio door. This unique orientation allows better winter passive solar heating and optimal rooftop PV generation. The author chose this setup to explore the feasibility of increasing the south glazing area for passive solar heat gain. The dimensions of the model house are as follows: Footprint Area—142.1 m²; Finished Floor Area—244.5 m²; Floors—2; Floor Height—2.74 m; Conditioned Volume—670.8 m³; Bedrooms—4; Bathrooms—2.5; Unfinished Attic and Basement. (Please check Appendix A and Appendix B for reference model inputs and sources).

3.2. Thermal System Models

Each thermal-system-specific energy model is labelled by the space heating method (Figure 2). Each model contains a suite of space heating (main and supplementary) and cooling production and delivery, DHW production and delivery, and, in the case of the NZC model, solar PV generation. Each model is intended to provide a comparable level of service, including space cooling and backup electric resistance heating for all air-source heat pumps for reliability during very low temperatures. Four all-electric models, one gas-heated model, and one NZC model were developed at each location.

In most cases, the DHW production system is the same as the one used for space heating. For instance, the gas model uses gas-heated DHW and the heat pump models are equipped with an HPWH, while the electric model has a resistive heating electric hot water tank. HPWH results in space cooling during the cooling season, which is beneficial, but it can be detrimental during the heating season.

3.2.1. Modelled Thermal Systems Summary

As previously stated, the gas optimization uses the same system type as the reference model, but with more efficient options. All the models undergo a whole-building optimization except for the reference model.

Table 1. Thermal system model inputs summary. “NA” means “Not Applicable”.

Model Name	DHW	Cooling	Heating	Backup Heating	Delivery
Reference	Natural gas, 40 gal; Energy Factor 0.59	Central AC; SEER 15	Natural gas furnace; 80% AFUE	NA	Ducted air
Gas	Natural gas (tank/tankless options)	Central AC (multiple options)	Natural gas furnace (multiple options)	NA	Ducted air
ASHP	HPWH 80-gal; Energy Factor 2.3	Central ASHP; SEER 22	Central ASHP; 10 HSPF	Integrated electric	Ducted air
MSHP	HPWH 80-gal; Energy Factor 2.3	Ductless MSHP; SEER 22	Ductless MSHP; 11 HSPF	Baseboard electric	4 Indoor units
GSHP	HPWH 80-gal; Energy Factor 2.3	Central GSHP; EER 20.2	Central GSHP; COP 4.2	Integrated electric	Ducted air
Electric	Electric tank, 50 gal; Energy Factor 0.95	Window AC; EER 10.7	Electric baseboard	NA	NA

provides a summary of the five thermal systems described in this section, as well as the reference system. A water heater’s efficiency is determined by its energy factor which is based on the amount of hot water produced per unit of fuel consumed. A higher energy factor means a more efficient water heater.

3.2.2. Thermal Envelope Performance (SHDI)

Thermal envelope performance refers to a building’s overall thermal performance. SHDI value expresses it, taking into account thermal resistance, air leakage, solar gain, and internal heat gains. It varies by location due to climate and differences in the NBC 9.36 reference enclosure design. SHDI value is not much affected by different thermal system types, but by minor variations in internal heat gain and interactions with HPWH. Figure 4 shows SHDI and EUI values for all locations.

3.2.3. Purchased Energy Consumption (EUI)

Figure 5 provides some insight into the energy end-use breakdown for the reference model in Toronto, and how each of the all-electric thermal systems perform in the same building enclosure and location. The natural gas furnace and water heater in the reference model are low–medium efficiency. Furthermore, the fact that there are two types of purchased energy included in the chart makes it impossible to conclude anything about the operational cost or GHG emissions of the reference building relative to the other buildings.

4. LCC Optimization

4.1. Location, Climate, and Emissions Factors

This study includes Canadian cities that are heating-dominated. Vancouver is the warmest city among the selected cities, while Edmonton is the coldest. Latitude determines solar PV potential, with Edmonton being the northernmost, and Toronto the southernmost. Seasonal electricity generation fluctuation is not critical to a customer with a grid-tied net metering contract. However, it helps to determine the aggregate value/burden of distributed solar generation to the electric grid.

Table 2. Climate, location data, electric grid annual average, and marginal EFs.

Location (NBC 2015)	Elevation (m)	Latitude (N)	Design Temperature				Degree Days below 18 °C	NBC Zone	EF g CO2e/kWh
			January		July 2.5%
			2.5% °C	1% °C	Dry °C	Wet °C			Ave	Mar
Vancouver	120	49°15′	−6	−8	28	20	2925	4	10.1	517
Edmonton	645	53°32′	−30	−33	28	19	5120	7A	960	469
Toronto	175	43°42′	−20	−22	31	24	3760	5	40	394.3
Montreal	45	45°30′	−23	−26	30	23	4270	6	1.9	330.9
Halifax	55	44°38′	−16	−18	26	20	4000	6	710	712

summarizes location data. The average and marginal emission factors (EFs) for electricity for the chosen cities are also shown in Table 2. The EF of onsite natural gas combustion is constant for all locations at 179.1 g CO₂e/kWh [24].

4.2. City Cost Index (CCI)

The material and labour costs for all BEopt components are from National Residential Efficiency Measures Database (NREMDB) and represent the U.S. national average in USD. Costs for rapidly developing technologies, such as LED lighting and solar PVs, are updated frequently. To adjust these costs for each city, the RSMeans “City Cost Index” (CCI) has been applied. For Canadian cities, the CCI simultaneously adjusts for local construction costs and currency conversion from USD to CAD [25,26]. By using the 2017 base CCI values for Canadian cities and fuel price inputs in CAD, all cost outputs from BEopt are in CAD (the 2017 average currency conversion was 0.77 CAD/USD). There is a separate CCI for materials and labour. The BEopt cost multiplier input corresponds to the CCI/100. See

Table 3. RSMeans (2017) CCI [25,26].

City	Material	Labour/Inst.
Vancouver	115.6	93.9
Edmonton	119.8	98.6
Toronto	114.5	105.6
Montreal	115.7	91.5
Halifax	116.1	85

for the CCI values used in the BEopt models.

4.3. Purchased Energy Prices

The purchased energy investigated in this study are electricity and natural gas. The term “purchased energy” is necessary because electricity is not considered a fuel, and it is precisely that energy with a cost that a homeowner seeks to minimize for reducing the building’s operational cost. Such a statement may appear obvious; however, it is useful to emphasize, since building enclosure designers and energy modelers sometimes prioritize a reduction in the thermal energy demand. This approach neglects the thermal conversion efficiency of the mechanical system and the unit price of the input purchased energy, which can vary greatly by location. SHDI may provide little useful information about the operational space heating cost of a building if the heating system efficiency and type are not known. A potential reason for such an economically suboptimal design approach is an ongoing disconnect between the civil/architectural disciplines and the mechanical discipline.

Table 4. Electricity and natural gas prices used in the modelling [27,28].

Location	Electricity		Natural Gas
	Marginal (CAD/kWh)	Fixed (CAD/Month)	Marginal (CAD/GJ)	Fixed (CAD/Month)
Vancouver	$0.1101	$8.26	$7.19	$12.75
Edmonton	$0.0785	$30.10	$6.55	$49.50
Toronto	$0.1048	$34.89	$7.95	$22.60
Montreal	$0.0670	$14.25	$14.76	$18.98
Halifax	$0.1582	$11.37	$20.97	$25.15

summarizes the purchased energy input prices and the sources from which they were retrieved or derived.

4.4. Financial Model

4.4.1. Capital Costs

The initial cost of a building covers energy-related components and equipment, excluding items such as concrete foundation, interior finishing, and miscellaneous electrical appliances. The NREMDB [29] includes additional parameter options, with material and labor costs determined using linear extrapolation. For the cost of the air barrier, a default equation in BEopt is used. Equation (5) defines this function and is based on empirical data from the national average.

$$Air Sealing Cost\left(\mathrm{\$}\right)=\left(0.65\times FFA\right)-(0.205\times \mathrm{l}\mathrm{n}(ACH50)\times FFA)$$

(4)

where

FFA = building’s finished floor area in (ft²)

ACH50 = building’s airtightness in air changes per hour at 50 Pascal depressurization

Space conditioning systems have a fixed cost and a capacity cost that correlates to the design peak heating load of the building.

Table 5. Space conditioning system input capital costs.

Design Name	Space Conditioning System	Price Metrics (CAD)
		Heating Main		Supplementary Electric	Cooling (AC)		Delivery (Ducting)
		Capacity [Output] ($/kBtuh)	Fixed ($/Unit)	Capacity ($/kBtuh)	Capacity ($/kBtuh)	Fixed ($/Unit)	Area ($/ft2)	Total ($/Unit)
Reference	Gas 80% AFUE + AC SEER 15	$3.09	$2157		$48.09	$2698.83	$5.56	$5266.99
Gas	Gas 92.5% AFUE + AC SEER 16	$4.78	$2630.60		$48.09	$2840.81	$7.04	$6669.00
ASHP	ASHP 10 HSPF; SEER 22	$48.09	$4395.00				$7.04	$6669.00
MSHP	MSHP 11 HSPF + baseboard	$92.75	$1856.62	$41.78
GSHP	GSHP 4.2 COP; EER 20.2	$564.25	$1833.36				$7.04	$6669.00
Electric	Baseboard electric + window AC	$41.78			$50.95	$38.02

lists these costs for each thermal system (and excludes DHW cost).

4.4.2. Mortgage

In total, 79% of Canadians choose high-ratio mortgages with a maximum 25-year loan repayment period, but some prefer conventional mortgages for the longer 30-year repayment period and no need for CMHC mortgage default insurance [30]. With a conventional mortgage, you need a 20% down payment and monthly payments are fixed over the 30-year repayment period.

$$MonthlyPayment=P\left[{\displaystyle \frac{m{(1+m)}^n}{(1+m{)}^n-1}}\right]$$

(5)

where

P = Principal amount on loan after down payment

n = Number of mortgage payments (i.e., 12 × 30 = 360)

m = Monthly interest rate (i.e., 0.039/12 = 0.00325)

4.4.3. Economic Parameters

This information comes from the U.S. Department of Energy’s EERE Building Technologies Program report titled “Methodology for Evaluating Cost-Effectiveness of Residential Energy Code Changes” [31]. The report justifies this method by explaining that it is a reliable way to calculate the real discount rate.

The real discount rate is thus calculated as follows:

$$R_{real}=\left[{\displaystyle \frac{{(1+R}_{nominal})}{(1+i)}}\right]-1$$

(6)

where

R_real = real discount rate

R_nominal = nominal discount rate

i = inflation rate

A summary of the mortgage and financial project inputs for BEopt model is presented in

Table 6. Mortgage and financial project inputs.

Economics		Mortgage
Project Analysis Period (yr.)	30	Down Payment	20%
Inflation Rate	1.54%	Mortgage Interest Rate	3.90%
Discount Rate (Real)	2.33%	Mortgage Period (yr.)	30
Discount Rate (Nominal)	3.90%

. Real fuel escalation rates have been kept at zero—that is, electricity and natural gas prices rise with inflation.

4.4.4. Cash Flows and LCC

Cash flows for the financial project include mortgage down payment in year zero, annual mortgage payments, annual utility charges (purchased energy use), annual PV generation payments (if applicable), equipment replacement costs, and the building residual value credit at year thirty. The residual value of all building components is calculated based on linear asset value depreciation over time. This is determined based on the material input data for each component, which includes both cost and lifetime.

All cash flows except mortgage payments (principal and interest) are inflated according to the year they occur. Annual energy use is based on the annual simulation and does not change over the project’s life. Inflated future costs are calculated using Equation (7).

$${Cost}_{year=k}={Cost}_{initial}{(1+i)}^k$$

(7)

where

i = inflation rate

Cost_initial = Cost of component/payment in year k = 0

k = the year the costs occur

To understand LCC for a given model, it helps to visualize an annual cash flow diagram. Output from BEopt for the optimal TO-MSHP model results in the cash flow shown in Figure 6. Only replacement events for HVAC and DHW equipment are labelled. Other events include ceiling fans and major appliances and do not change for all models in this study.

All cash flows for a given year are then summed to obtain a net total cost (TC) for each project year. These future costs are then converted into present costs using the nominal discount rate and summed to obtain the LCC Equation (8).

$$LCC=\sum _{k=0}^N({TC}_{year=k}{(1+d_n)}^{-k})$$

(8)

${TC}_{year=k}$ = total costs in year = k

$d_n$ = the nominal discount rate

4.5. Design Input Options

This section refers to design input categories for which there is more than one parameter option selected for the optimization run. Note that some of the fixed parameters have been reset from the reference model (

Table 7. Optimization design input options—gas model only.

Group	Category	1	2	3	4
Space Conditioning	Central Air Conditioner	SEER 15	SEER 16 (2 stage)	SEER 18	SEER 24.5
	Gas Furnace	80% AFUE	92.5% AFUE	95% AFUE	98% AFUE
Water Heating	Gas Water Heater	Energy Factor: 0.59	Condensing; Energy Factor: 0.82	Tankless; Energy Factor: 0.82	Tankless condensing; Energy Factor: 0.96

). This has been done for the categories of “interzonal wall” and “interzonal floor”. These are the envelope components separating the building living space from the garage, and they constitute a small portion of the total enclosure area. To reduce computational intensiveness, these components were not separately optimized. However, to ensure that they did not “pull down” the resultant optimal R-value of the rest of the enclosure, the interzonal floor and walls were reset to an effective R-value of 36.6 and 38.6, respectively.

The gas thermal system model was the only one for which multiple heating, cooling, and DHW system efficiency options were explored. The water heater options that use a storage tank have a 40-gallon tank. However, multiple DHW and space conditioning delivery system efficiencies were explored for all applicable models.

To create an option for a Drain Water Heat Recovery (DWHR) in BEopt, a custom option for the “hot water fixtures” category was created that has a shower DHW use reduced by 53% (the efficiency of the unit), resulting in a total DHW use reduction of 25%. The standard 3” diameter, 60” long unit selected has a cost of $622 + $80 (material and labour in USD), and a lifetime of 40 years.

Since the reference model varies for some parameters by location (wall, ceiling, and foundation R-values, and window U-value), the lower limit of the option range varies for these categories, though intermediate options and the upper limit do not. Only design options were explored that do not comply with NBC 9.36 for the location in question (i.e., all options explored meet or exceed the reference model parameter for that category by location). The only exception to this rule is the possibility to downgrade from a U-0.27 window to a U-0.29 window for Edmonton, Montreal, and Halifax.

Table 8. Optimization design input options for all models.

Group		Category	1	2	3	4	5	6	7	8	9
Walls	1	Wall sheathing	None	R-5 XPS	R-10 XPS	R-15 XPS	R-20 XPS	R-25 XPS	R-30 XPS	R-35 XPS	R-40 XPS
Ceilings/Roofs	2	Unfinished attic	R-38 (39.6)	R-49 (50.6)	R-60 (61.6)	R-70 (71.6)
	3	Radiant barrier	None	Double-sided foil
Foundation/Floors	4	Unfinished basement	R-12 polyiso (13.3)	R-18 polyiso (19.3)	(22.6)	(28.1)	(33.4)	(38.6)
Windows and Doors	5	Window areas	back: 23%	Back: 33%	Back: 43%
	6	Windows	U-value 0.32	U-value 0.3	U-value 0.29	U-value 0.27	U-value 0.3 (triple)	U-value 0.29 (triple)	U-value 0.21 (triple)	U-value 0.18 (triple)
	7	Overhangs	None	Back windows: 2′	Back windows: 3′
Airflow	8	Air leakage	2.5 ACH50	2 ACH50	1 ACH50	0.6 ACH50
	9	Mechanical ventilation	HRV, SRE: 70%	ERV, SRE: 70%, TRE: 0.48%
Space Conditioning	10	Duct insulation	None	R-4	R-8
	11	Ceiling Fans (four fans)	45 W/fan	20 W/fan
Hot Water	12	Pipe insulation	None	R-2	R-5
	13	Shower DWHR	None	53% efficiency

Blue = VN and TO only, Green = VN only, Yellow = category not available for electric and MSHP models. Bracketed values indicate effective R-values for assembly (US units).

shows a complete list of design input options used for all models.

5. Optimization Results and Discussion

5.1. Gas Model Optimization

This section showcases the optimization results for the gas model, which represents a whole-building optimization using the same thermal system type as the reference model. It can be used as a benchmark design to compare the economic feasibility of all-electric and NZC design alternatives. The optimal LCC gas design reduces GHG emissions and energy-related building LCCs compared to the reference model. Figure 7 illustrates the Toronto (Gas-TO) optimization results, and

Table 9. Resultant lowest LCC gas designs by location. In this Table, VN stands for Vancouver, ED for Edmonton, TO for Toronto, ML for Montreal, and HX for Halifax.

Group	Category	Ref-TO	Gas-VN	Gas-ED	Gas-TO	Gas-ML	Gas-HX
Walls	Wall sheathing	None	None	None	None	R-15 XPS	R-20 XPS
Ceilings/Roofs	Unfinished attic	R-38 (39.6)	R-38 (39.6)	R-49 (50.6)	R-38 (39.6)	R-60 (61.6)	R-60 (61.6)
	Radiant barrier	None	None	None	None	None	None
Foundation/Floors	Unfinished basement	R-18 polyiso (19.3)	R-12 polyiso (13.3)	R-18 polyiso (19.3)	R-18 polyiso (19.3)	R-18 polyiso (19.3)	R-18 polyiso (19.3)
Windows and Doors	Window areas	Back: 23%	Back: 23%	Back: 23%	Back: 23%	Back: 23%	Back: 23%
	Windows	U-value 0.32	U-value 0.32	U-value 0.27	U-value 0.32	U-value 0.27	U-value 0.27
	Overhangs	None	None	None	None	None	None
Airflow	Air leakage	2.5 ACH50	2.5 ACH50	0.6 ACH50	1 ACH50	0.6 ACH50	0.6 ACH50
	Mechanical ventilation	HRV, SRE: 70%	HRV, SRE: 70%	HRV, SRE: 70%	ERV, SRE: 70%, TRE: 0.48%	ERV, SRE: 70%, TRE: 0.48%	HRV, SRE: 70%
Space Conditioning	Duct insulation	None	R-8	R-8	R-8	R-8	R-8
	Ceiling fans (four fans)	45 W/fan	45 W/fan	45 W/fan	45 W/fan	45 W/fan	45 W/fan
Hot Water	Pipe insulation	None	R-2	R-2	R-5	R-5	R-5
	Shower DWHR	None	None	None	None	DWHR	DWHR
Gas model-specific options
Space Conditioning	Central air conditioner	SEER 15	SEER 15	SEER 16 (2 stage)	SEER 16 (2 stage)	SEER 16 (2 stage)	SEER 15
	Furnace	80% AFUE	92.5% AFUE	92.5% AFUE	92.5% AFUE	92.5% AFUE	92.5% AFUE
Water Heating	Water heater	Energy factor: 0.59	Tankless; energy factor: 0.82	Tankless; energy factor: 0.82	Tankless; energy factor: 0.82	Tankless; energy factor: 0.82	Tankless condensing; energy factor: 0.96

Grey = upgrade from reference, green = downgrade from reference. Bracketed values indicate effective R-values for assembly (US units).

summarizes the lowest LCC gas design options for each location. Increasing insulation can improve energy/emissions performance, but the capital costs do not outweigh the energy cost savings. Mechanical system upgrades were generally more economically desirable than thermal envelope upgrades.

5.2. LCCs versus GHG Optimization Results

5.2.1. Results Summary

The following plots (Figure 8) show the reference design plus the lowest-LCC design from each optimization for the chosen cities. For comparison, all y-axis minor gridlines are shown in $1000 CAD increments.

Some electric models have a lower LCC than the gas model. MSHP has the lowest overall cost of design in every location except Halifax. The high EF of the Alberta (Edmonton) and Nova Scotia (Halifax) electric grids results in gas being the lowest-emitting design without PVs. In locations with high-EF electricity, PVs substantially impact emission reductions more than thermal envelope and system efficiency improvements combined. PVs are the most affordable in Halifax. The electrification of heating is the most advantageous economical measure in reducing emissions in locations with low-EF electricity.

5.2.2. NZC Models

This subsection explains in more detail how the NZC/NZE optimizations were conducted that yielded the results shown in Figure 9. This method has been used to determine the amount of renewable generation required as a proportion of total energy consumption to achieve NZC status. The NZC model applies PV to the resultant lowest-cost all-electric optimization model, which for every location was the MSHP model. However, rather than add PV to the resultant design, the model is re-optimized with all of the original options with PV included. Figure 9 is an example of using the Edmonton case for a lowest-LCC NZE design. Being the only location where the average EF exceeds the marginal EF (Table 2), it would require a net-positive energy building to achieve NZC status. However, since net-metering programs do not reward annual generation in excess of annual usage, this design was optimized for site-specific NZE rather than NZC. A preliminary optimization iteration was conducted to determine the appropriate cost-optimal range of the PV system size. The resultant design uses 12.6 MWh of electricity, requiring 10 kW of PV for this site.

While it is an interesting exercise to conduct a whole-building optimization that includes only the required quantity of PV to achieve NZC, the question of rooftop PV economic feasibility can be answered in isolation. If it is profitable, the rational decision would be to install the maximum amount allowed, which would be the lesser of what can fit on the roof and what will attain NZE levels of generation. In such a case, conducting a whole-building LCC vs. site energy optimization would still be beneficial in determining the optimal overall design. Depending on the profitability of PV, this can yield a less efficient building design that enables more PV. Though this could benefit the homebuyer, it would result in the electrically heated building having a higher peak winter electricity demand (when PV is not generating), negatively impacting the utility (and ultimately all ratepayers).

5.3. Initial Cost and Design

The initial cost of each of the optimal designs for Toronto displayed in Figure 10 is broken down by building component category. In each case, 20% of the total cost is paid as the down payment, with the remaining 80% constituting the mortgage principal, which is paid over the 30-year mortgage period. Note that the building components with the greatest cost variation are the walls/air-barrier and the space conditioning system. Also note that, while the reference design has the lowest initial cost, it has the highest LCC in the charts above (Figure 10). The MSHP design is initially 1.0% costlier than the optimal gas design, whereas the electric design is 2.2% cheaper.

The resultant optimal designs for each of the Toronto models are listed in

Table 10. Resultant lowest-LCC designs for all thermal system types—Toronto (TO).

Group	Category	Ref-TO	Gas-TO	ASHP-TO	MSHP-TO	GSHP-TO	Electric-TO
Walls	Wall sheathing	None	None	None	R-20 XPS	None	R-20 XPS
Ceilings/Roofs	Unfinished attic	R-38 (39.6)	R-38 (39.6)	R-60 (61.6)	R-60 (61.6)	R-38 (39.6)	R-60 (61.6)
	Radiant barrier	None	None	None	None	None	None
Foundation/Floors	Unfinished basement	R-18 polyiso (19.3)	R-18 polyiso (19.3)	R-18 polyiso (19.3)	R-18 polyiso (19.3)	R-18 polyiso (19.3)	R-18 polyiso (19.3)
Windows and Doors	Window areas	Back: 23%	Back: 23%	Back: 23%	Back: 23%	Back: 23%	Back: 23%
	Windows	U-value 0.32	U-value 0.32	U-value 0.32	U-value 0.32	U-value 0.32	U-value 0.29
	Overhangs	None	None	None	None	None	None
Airflow	Air leakage	2.5 ACH50	1 ACH50	0.6 ACH50	0.6 ACH50	2 ACH50	0.6 ACH50
	Mechanical ventilation	HRV, SRE: 70%	ERV, SRE: 70%, TRE: 48%	HRV, SRE: 70%	HRV, SRE: 70%	HRV, SRE: 70%	HRV, SRE: 70%
Space Conditioning	Duct insulation	None	R-8	R-8	NA	R-8	NA
	Ceiling fans (four fans)	45 W/fan	45 W/fan	45 W/fan	45 W/fan	45 W/fan	45 W/fan
Hot Water	Pipe insulation	None	R-5	R-5	R-5	R-5	R-5
	Shower DWHR	None	None	DWHR	DWHR	DWHR	DWHR

Grey = upgrade from reference. Bracketed values indicate effective R-values for assembly (US units).

.

The analysis found that in Toronto, all models had improvements in airtightness and DHW piping insulation. Electric optimization resulted in the most significant thermal envelope upgrades. The NZC design is the same as the MSHP design, with 1 kW of PV added. In Edmonton and Halifax, higher quantities of PV had a more significant impact on cost and emissions results, influencing building designs.

5.4. Operational Energy, Cost, and Emissions

5.4.1. Purchased Energy Use

Energy end-use for each of the Toronto designs is shown in Figure 11 and utility costs by fuel type for the same designs can be seen in Figure 12. The comparisons in Figure 13 are analogous to the EUI, since the building geometry is unchanged for all designs. Figure 13 provides a breakdown of GHG emissions by fuel type for the same designs.

The models that utilize heat pumps to provide space heating and DHW have the lowest energy consumption. However, somewhat surprisingly, they also have a lower operational cost than the gas design. The electric design consumes less energy, but the energy cost is higher. Taken together, these three charts illustrate how total energy use does not necessarily provide any useful information regarding either energy cost or energy-related emissions, especially when comparing buildings with different heating system types. For the gas-heated designs, the natural gas proportion of the total cost is much lower than it is for total energy, but it accounts for almost all of the GHG emissions. Furthermore, these interrelationships can vary greatly by location, as the electricity/gas price ratio changes, as well as the electricity EF.

5.4.2. Enclosure Thermal Performance

The SHDI metric is intended here as a means of characterizing the thermal performance of the building enclosure, including heat recovered by the HRV/ERV. In addition to representing the enclosure performance, it indicates how much heat is required to be provided by any given mechanical space heating system. However, for models containing an HPWH, the space heating and space cooling demands are “distorted” by the cooling effect it provides. To understand the magnitude of this effect, the optimal MSHP design has been run separately, with and without the HPWH. When the building includes an HPWH, the space heat demand, or SHDI, for the building increases by 3.3 kWh/m², or 12%. Additionally, the space cooling demand intensity decreases by 0.7 kWh/m², or 15%. For the ASHP and GSHP models, the SHDI increases by 3.5 and 3.6 kWh/m², respectively, when the HPWH is included. To achieve a fair comparison of enclosure thermal performance, the SHDI values illustrated in Figure 14 have been determined without the distorting effect of the HPWH. To understand how much heat is delivered by the heating system for the heat pump models, add approximately 3.5 to the value shown below for those models.

Whether thermal system type has a substantial impact on optimal thermal envelope design is inconclusive based on these data alone. There are many factors, including both the capital and operational price of heat. Ideally, one would attempt to correlate an enclosure’s SHDI and the life-cycle price of heat ($/J). However, since the capital investment in the heating source is interlinked with the equipment for cooling and DHW, and in different ways for different thermal system types, it is difficult to isolate.

However, at least for the electric heating sources, there appears to be an inverse correlation with the operational price of heat. From the simulated heating COP values, we can see that the price of heat from the GSHP is the lowest, followed by the ASHP, the MSHP, and finally, the electric baseboard. It is therefore logical that as a heat source becomes more expensive, it is optimal to invest more in a higher-performing thermal envelope to reduce the SHDI. This is a fact that is not recognized by NBC 9.36, or either of the Passive House standards, and only very simply by OBC SB-12. Furthermore, when multiple heating system types are considered, as is the case here, a lower SHDI guarantees neither reduced GHG emissions nor a lower-LCC building, as can be seen by the lack of correlation with LCCs in Figure 8.

5.4.3. Mechanical Heating

Simulated hourly heat delivered by the MSHP to the building for the MSHP-TO optimal design is illustrated in Figure 15. The maximum capacity limit of the heat pump, which falls with the outdoor temperature, can be seen at the top left of the scatter. Beyond this point, supplementary electric baseboard heating meets the hourly load. However, supplementary heating only accounts for 0.7% of the annual delivered heat.

If large numbers of buildings begin switching to electric sources of heating, it will be important to mitigate peak electric demand on the grid. Part of the way this can be done is through higher performance enclosures to minimize the peak heating load. The largest spikes in demand are due to the electric supplementary heating, which, for the MSHP-TO design, can result in a whole building’s electric demand of up to 7 kW. However, as illustrated in Figure 16, the building electric demand rarely exceeds 4 kW, and the heat pump never exceeds 2.3 kW. On the other hand, heating with a heat pump can reduce peak electric demand if the alternative is baseboard electric heating (see Figure 17).

5.4.4. EUI versus GHGI Relationship

The EUI does not necessarily correlate to energy cost or LCC. Figure 18 plots site energy use against GHG emissions for all building variants simulated by the Gas-TO and MSHP-TO (Toronto) optimizations. In each case, there is very good positive correlation. In the case of the gas model, this is likely due to a combination of the fact that the vast majority of emissions are from natural gas consumption and because for the gas models, electricity consumption varies little for all design variants. In the case of the MSHP, there should be perfect correlation, since there is only one purchased energy type, with a fixed carbon intensity. The imperfect correlation here is simply caused by rounding errors, which are due to BEopt’s insufficiently small GHG-accounting increments for very low-emitting buildings. However, when building designs across thermal system types are compared, the correlation is less clear. To illustrate this, the procedure was repeated for Edmonton, but this time considering all points as a single data series (Figure 19).

In Figure 19, the left cluster is the MSHP variants and the right is the gas variants. Here, again, there is good correlation within a thermal system type, but very poor correlation across system types. In this location, for these two system types, there is a slightly negative correlation between EUI and GHGI.

6. Conclusions and Recommendations

Optimization studies were conducted to find the best designs for five types of thermal systems plus a NZC model for five major Canadian cities. Based on optimized designs of all-electric thermal systems, the ones that contain MSHP are the most affordable option in every location. In Halifax and Montreal, the optimal design of all-electric types is cheaper than the optimal gas design. In Toronto, ASHP and electric models are also cheaper than gas, and in Edmonton, the GSHP and electric models are more cost-effective. In every location, except Edmonton, the NZC design is cheaper than the gas design. Electric heating systems emit substantially less GHG than gas systems in Vancouver, Toronto, and Montreal. In Edmonton, the gas system is the least-emitting thermal option. Passive and active energy conservation measures, along with rooftop PV, can reduce emissions.

Building energy standards use total building EUI and SHDI as the most common performance metrics to control the use of purchased energy and minimize heating demand. The intent is to reduce costs and emissions, as well as provide consumer and environmental protection. This study found that EUI is strongly correlated to energy cost when only one type of purchased energy is used (e.g., all-electric), but there is reduced correlation for buildings that use two types of purchased energy with different prices. Additionally, when comparing buildings with different thermal system types, such as all-electric and gas-heated buildings, the EUI is no longer a reliable metric. Optimal SHDI is related to the initial cost of the heating system, and further research could be informative. While GHGI is tightly correlated to EUI for buildings that use one or two types of purchased energy, there is no longer a correlation when comparing emissions from buildings with different thermal system types. Building energy standards should mandate heating system type, not just minimum efficiency levels. This will benefit both consumers and the environment. Economic analysis of the impact of heating electrification on the electricity grid are needed.