An Incentives Planning Framework for Residential Energy Retrofits: A Life Cycle Thinking-Based Analysis under Uncertainty

Abstract

Building energy retrofits can reduce emissions and increase cost savings. Some retrofits that can deliver higher emissions savings are not popular due to a lack of economic justifications. Financial incentives can be used to change buyer perception around such retrofits. This study proposes a framework to identify the best-performing retrofit strategies for a given building cluster and the optimal incentive amounts to promote the chosen strategies, accounting for uncertainties, stakeholder priorities, and budget constraints. The proposed framework was demonstrated using a case study complemented with policy insights. Life cycle cost savings and capital cost significantly impact retrofit purchase decisions. Case study results showed that retrofitting houses heated with electricity can produce significant cost savings. However, adopting energy-conscious behaviours in houses heated with natural gas and injecting renewable natural gas into the gas supply can produce two times more emissions savings achieved by any retrofit strategy applied to an electrically heated house. This indicates the need for adopting performance-based incentives over the prescriptive approach to reward occupant efforts in addition to asset performance. Despite potential life cycle cost savings, incentives must be complemented with low-interest loans to promote retrofit strategies carrying higher capital costs.

1. Introduction

Climate change is not identified as a future crisis anymore. Leading scientific bodies such as NASA acknowledge that climate change’s impacts are already affecting the world [1]. The inter-governmental panel on climate change (IPCC) indicates that the associated environmental impacts would be beyond repair if the global temperature increases 1.5 °C over pre-industrial temperatures [2]. IPCC advocates curbing anthropogenic greenhouse gas (GHG) emissions to avoid the predicted climate adversities and identifies the building sector as a key area that can support this as it accounts for 12% of anthropogenic GHGs and 40% of the energy use in the world [2,3,4].

In response to these rising challenges, climate leaders such as British Columbia (BC), Canada developed innovative policy tools such as BC Energy Step Code to reduce the environmental impacts of new constructions [2]. However, improving the performance of existing buildings constructed to less stringent energy efficiency standards remains a challenge [3].

Retrofits are commonly used to improve the energy performance of existing buildings. Energy retrofits lead to economic, environmental, and indoor environmental quality (IEQ) performance enhancements [4]. Occupants can benefit from operational cost savings and improved IEQ [5]. In addition, building owners can benefit from the property value increases. In the bigger picture, retrofitting projects can support national climate change mitigation initiatives and local economies through emissions savings and job creation [6]. However, financial pressures such as high life cycle costs (LCC) and capital costs (CC) and lack of economic justifications can reduce the community acceptance of retrofits [7,8]. Therefore, building owners have to be not only guided and educated but also need to be provided with an economic rationale to adopt these interventions.

Carbon taxing and emissions capping are used to minimize the use of energy generated using higher emissions sources [9]. However, regionally implemented penalties such as provincial carbon taxes can pose challenges to local industries competing in the global market due to increased production costs [10]. Moreover, energy affordability issues resulting from penalties can create social predicaments that could result in a backlash against government-led clean energy promotion and climate change mitigation initiatives [11,12]. Therefore, bringing forward the penalties as the first step can hinder the success of energy retrofit promotion campaigns. Moreover, emissions penalties and caps fail to promote energy conservation opportunities associated with cleaner energy sources. Therefore, promoting performance improvements with financial incentives is more appropriate to drive industries and the public towards energy-efficient technologies and cleaner energy sources in the initial stages [13,14]. When the community is ready for more clean energy uptake and equipped with energy efficiency practices, penalties can be slowly introduced in an incremental approach. Considering the importance of incentives/rebates in promoting building energy retrofits (BER), this article focuses on financial incentives planning for energy retrofits.

A Compendex database search in January 2022, using “residential”, “building”, “energy”, “retrofit”, and “incentive” as keywords yielded seventy-three research articles. Diakaki et al. developed a retrofit selection algorithm considering annual energy use, annual emissions, and capital cost employing a multi-objective solution methodology [15]. Bonomolo et al. proposed an approach to identify the cost-optimal lighting retrofits accounting for daylighting opportunities [16]. Zang et al. proposed a Pareto optimization approach for residential retrofit selection [17]. Ruparathna et al. developed an economic evaluation approach for retrofit evaluation [18]. All the reviewed literature contributed to improving retrofit selection processes. However, the current approaches did not address either one or a few of the following aspects of community retrofit planning: life cycle costing and emissions, addressing the uncertainties, contradicting decision priorities, and promotion of retrofits through financial incentives.

Findings from the literature review motivated the authors to investigate the need for a comprehensive strategy to identify the best retrofit strategies and optimal incentive levels to promote the best-performing retrofit packages/strategies under uncertain conditions. In addition to the literature review, the authors communicated with experts from local municipalities, construction companies, building management agencies, utility providers, and Natural Resources Canada to establish the need for an energy retrofits and incentives planning mechanism. All stakeholder groups echoed the need for a decision support tool enabling standardized incentives and energy retrofits planning strategy to achieve the climate action goals of the existing buildings sector. Identified research gaps related to building energy retrofits and incentives planning are summarized below:

(1)

No comprehensive models have been developed for quantitatively determining incentive amounts for building energy retrofits, and the multi-stakeholder nature of the problem is overlooked [13,19,20].

(2)

Life cycle emissions (LCE) are overlooked in retrofits and financial incentives planning. Embodied emissions are often neglected due to the singular focus on operational emissions [21].

(3)

Uncertainties related to building retrofits and the external environment have been largely neglected in past studies in retrofit strategy and incentives planning [13,22]. Moreover, the sensitivity of the residential energy retrofit project outcomes to the operational conditions and macroeconomic parameters has not been sufficiently investigated.

This study addresses the research gaps discussed above regarding its academic merit by proposing an incentive/rebate planning strategy for BERs to maximize environmental and economic benefits under uncertainty while accounting for budgetary constraints and stakeholder priorities. More importantly, it provides a much-needed quantitative decision-making model that can be used by different levels of government and non-governmental organizations to identify the optimal incentive levels with the limited budget in hand. Moreover, the proposed retrofit selection framework can be used by individual building owners/contractors to determine the best retrofit strategies under uncertainty for their buildings.

The proposed approach was demonstrated using a Canadian case study. The sensitivity of different retrofit strategies to uncertainties in energy supply and building operational conditions, macroeconomic parameters, and alternate decision priority scenarios was evaluated based on the case study results. However, the proposed approach can be used worldwide by altering the inputs. Inputs to the model include the energy-saving potential of different retrofit strategies (generated from the energy simulation software), energy cost and emission factors, retrofit-related cost factors (e.g., capital and maintenance costs), decision priorities, and constraints. These inputs can be selected according to the local conditions and the proposed retrofits selection and incentives planning algorithm can be applied irrespective of the location or building type.

1.1. Promoting Building Energy Retrofits

Opportunities and challenges associated with BERs can be better understood using the eco-efficiency classification presented in

Table 1. Eco-efficiency classification [13].

	LCE Savings	LCC Savings
Priority retrofits (PR)	Yes	Yes
Secondary retrofits (SR)	Yes	No
Tertiary retrofits (TR)	No	Yes/No

.

Retrofits that require energy source switching from a low-emission source to a high-emission source and retrofits that create higher embodied emissions than operational emissions savings fall under tertiary retrofits. Nevertheless, some TRs can produce LCC savings.

PRs simultaneously produce emissions and cost savings [13]. If the payback periods are sensible, building owners can be easily convinced to implement PRs, making them effective options for government emission reduction initiatives. High capital costs can be a barrier for PRs, preventing building owners from adopting them [23]. In such situations, a loan scheme can be introduced to assist the building owners in implementing such retrofits [17,24]. However, stakeholders are generally reluctant to adopt loan schemes due to the high administrative workload and pre- and post-performance uncertainties [25]. Administrative issues can be mitigated by involving energy performance contracts and energy services companies (ESCO).

Emissions reduction potential alone will not likely motivate building owners to adopt retrofits. Therefore, a business case needs to be created by financially incentivizing energy-efficient technologies and/or penalizing the cheaper alternatives that depend on emissions-intensive energy sources to make a business case for SRs [9,26].

1.2. Incentivizing Energy Retrofits

Incentivizing energy retrofits can be done following either a prescriptive or a performance-based approach, and each approach has unique positives and negatives [12]. In the prescriptive approach, the implementation of a prescribed retrofit is incentivized [27]. In practice, the incentive provider checks a predefined document, such as the bill of sale, and provides the incentive based on that. Prescriptive incentives are easier to implement compared to performance-based incentives [28]. From the receivers’ perspective, this boosts confidence in retrofit investments, as they are certain about the incentive to be expected once the retrofit is installed. This is easy to implement from the incentive issuers’ perspective as less administrative work is involved. Importantly, the actual energy and emissions performance of a retrofit’s post-installation phase highly depends on locational parameters, including weather and energy prices [13]. Moreover, it can significantly change with behavioural practices such as setting the heating and cooling temperatures and occupancy schedule. Therefore, if the prescriptive approach is followed, the outcomes must be evaluated based on local conditions because provincial or national-level generalizations may promote unsuitable energy retrofits for a given municipality or region. On the other hand, the prescriptive approach does not have the leverage to motivate building occupants to adopt energy-conscious behaviour during the post-retrofit period, as the incentives are issued upfront. This challenge can be addressed to a certain level with awareness programs.

In the performance-based approach, incentives are offered for achieving a certain goal concerning energy use/intensities or emissions savings/intensities based on a predefined performance benchmark. In this approach, the building owners are given the flexibility to reach the goal by applying any retrofit combination rather than being limited to predefined options. This flexibility for innovative decision-making is the strength of the performance-based approach. Moreover, it pays attention to the behavioural component of building energy performance. In contrast to the prescriptive approach, the owners are compelled to adopt energy-conscious behaviour to reach better performance levels to secure higher incentive amounts [29]. However, assessing the pre- and post-retrofit performance of an existing building and predicting the future outcomes of design decisions can be quite challenging when issuing this type of incentive [30].

1.2.1. Performance Assessment

There are two main methods commonly used for building energy performance assessments. The first approach uses energy bills for performance evaluations [19]. When employing this approach, it should be noted that energy demand can significantly vary with other factors such as weather, occupancy schedule, and occupant behaviour, which are independent of the energy efficiency of the building. This method needs at least one-year worth of pre- and post-retrofit energy records to evaluate the performance of the building/buildings under evaluation. Collecting energy records and accommodating missing utility bills can be quite challenging [31]. Alternatively, energy simulations can be used to evaluate the operational performance of retrofits [13]. The simulation assumptions can significantly affect the accuracy of the predicted retrofit performance, leading to many uncertainties [32]. ASHRAE energy simulation guidelines for existing buildings or local standards can be followed to minimize the errors created by assumptions [33,34].

1.2.2. Incentives Planning

Compared to the challenges associated with the penalties, rebates and subsidies are deemed better suited to retrofit incentivization initially [35]. Therefore, this study investigates incentive planning considering rebates and subsidies. However, incentive programs need viable financial models to keep the money flowing for continued operations and future sustainability. Therefore, penalties can also be introduced in the following phases to ensure the longevity of incentive programs.

Incentivizing energy retrofits aims to reduce the emissions associated with buildings. However, local governments and private sector organizations such as utility companies have only a limited budget for incentive programs. Therefore, identifying the best split of available funding among different retrofit options needing incentives is essential to maximize the emissions savings within the allocated budget from the incentive providers’ perspective. Parallelly, building owners and contractors need to be provided with tools to determine the most suitable retrofit options for their situations to meet the required performance targets and maximize incentive benefits.

In both situations discussed, selecting the best retrofit strategies and retrofits that need to be incentivized is challenging due to associated uncertainties and the exhaustive number of potential retrofit strategies that can be developed by combining alternative retrofit options [13]. Individual performance of retrofits can vary with the other retrofits combined with them. Moreover, the condition of the building and components needing replacement and the available budget greatly influence the building owner’s decision-making. Therefore, the incentive providers cannot prescribe a single retrofitting strategy for a house. Ideally, incentive amounts proposed for individual incentives should be able to promote the top-ranked retrofit strategies (combinations) applicable for a given building or building cluster.

2. Materials and Methods

This study proposes a framework for selecting retrofit strategies and financial incentives to promote environmentally friendly BERs under uncertain conditions. Therefore, this study did not analyze tertiary retrofits (which deliver no environmental benefits). Figure 1 summarizes the proposed framework and methodology.

Decision priorities and constraints act as inputs to the retrofit selection and incentives optimization process. Stakeholder consultation sessions are recommended to identify the decision priorities and constraints associated with the incentives planning exercise for any community planning to employ this framework.

The next step of the proposed framework is to conduct a sensitivity analysis to identify the critical parameters that influence the performance of BERs. Results from the sensitivity analysis help model the uncertainties associated with the BER performance employing fuzzy numbers. The fuzzy-based BER performance information, decision priorities, and constraints feed into a fuzzy-multi-criteria options ranking algorithm to identify the best retrofit strategies that meet the decision priorities and constraints. This algorithm outputs the best retrofit strategies (combination of retrofits) for the given community.

The selected retrofit strategies move forward to the proposed incentive optimization algorithm. This algorithm determines the optimal financial incentive amounts needed to promote the chosen strategies. Both retrofit selection and incentives planning algorithms can be flexibly used at the individual building or community level by changing the inputs and constraints.

The proposed framework was demonstrated using a case study. Two representative buildings were selected as the base buildings for the case study. Commonly used BERs were identified based on published literature and other resources such as manufacturer data. The energy-saving potential of BERs was evaluated with HOT2000 energy simulation software (future users can choose the energy simulation tools applicable to their jurisdiction or application). All possible combinations of the retrofits with life cycle emissions reduction potential were considered in the evaluation. The evaluation results produced the best retrofit strategies and optimal incentive amounts to promote the selected strategies in the case study location.

2.1. Sensitivity Analysis

BER performance predictions are sensitive to input parameters and assumptions. Therefore, a sensitivity analysis was conducted to understand the variations in the energy model outputs with the input parameters, including operational conditions and energy supply-related factors.

2.1.1. Sensitivity of Energy Demand to Operational Conditions and User Behaviour

Percentage energy demand change (PEDC) was against a percentage or an absolute change in input parameters, including the number of occupants, time at home, hot water demand and temperature, and space cooling and heating setpoints. The sensitivity analysis was conducted using the weather conditions in Kelowna, Canada. The sensitivity of energy demand to different input parameters can be more or less pronounced in harsher or milder climates compared to Kelowna. The results of the sensitivity analysis are presented in Section 4.1.1.

2.1.2. Sensitivity of Operational Cost and Emissions to Energy Price and Emission Factors

Depending on the grid energy mix, electricity can be cleaner than natural gas, but natural gas is generally cheaper in Canada. Energy cost and emission factors can be decisive parameters for source-switching decisions related to building heating and hot water systems. Therefore, operational cost and emissions were plotted against the energy sources’ unit energy price and emission factor variations, respectively, to understand the potential impacts. Two energy price and emission scenarios were considered for this. Scenario 1 assumes that the electricity price and emission factor stay constant while the natural gas (NG) price and emission factor vary. Scenario 2 assumes that NG’s price and emission factor remains constant while the price and emission factor of the electricity supply vary.

The operational emissions were calculated using Equation (1)

$$OE=\sum Deman{d}_j\times E{F}_j$$

(1)

where

• $OE$—Operational emissions (kgCO_2e)
• $Deman{d}_j$—Demand for j^th energy source (kWh)
• $E{F}_j$—Emissions factor of j^th energy source (kgCO_2,eq/kWh)

Similarly, the operational cost was calculated using Equation (2)

$$OC=\sum Deman{d}_j\times UE{P}_j$$

(2)

where

• $OC$—Operational cost ($)
• $UE{P}_j$—Unit energy price of j^th energy source ($/kWh)

Detailed results of this analysis are presented in Section 4.1.2.

2.2. Retrofit Performance Evaluation

The goal of the retrofit performance evaluation was to identify the best retrofit strategies to maximize the environmental and economic benefits. LCC and LCE were employed as economic and environmental performance indicators [13]. Retrofit performance can significantly vary with the uncertainties introduced by occupant behaviour, building operational conditions, energy price, and emissions factors [36,37,38]. Previous studies discussed the pros and cons of using fuzzy-based and probabilistic methods to address uncertainties [39]. However, most uncertain parameters in building-related decision-making do not have sufficient data to develop probability distributions [19,39]. Moreover, fuzzy numbers can handle qualitative and quantitative inputs [40,41]. Using fuzzy numbers allows the proposed framework to be extended to include qualitative parameters such as user satisfaction, occupant health, and thermal comfort in future research. Considering the discussed advantages, the framework was developed using fuzzy-based methods, and fuzzy sets were used to model uncertainties in the model parameters [19,39].

Previous studies have evaluated the performance of the retrofits considering individual retrofits [13]. However, the performance of individual retrofits can change significantly due to performance interactions when combined with other retrofits. Therefore, all possible combinations of the retrofits that can produce emission savings (priority and secondary retrofits) were considered when evaluating the retrofit performance. The performance values of individual retrofits under these variable conditions and retrofit combinations were brought forward to the incentive optimization algorithm. The detailed retrofit evaluation process is discussed below.

2.2.1. Energy Performance

HOT2000 is currently used for energy simulations of small residential buildings in BC Energy STEP Code and EnerGuide [8,29]. HOT2000 (Version 11.3) was used in this study as it is the recommended software for Canadian residential energy studies. Natural Resources Canada (NRCan) developed the Housing Technology Assessment Platform (HTAP) to support energy analysts in handling large numbers of energy simulations [42]. Using the strength of HTAP and HOT2000, all possible combinations of the considered retrofit options were simulated. The simulation results were used to calculate the operational cost and emissions savings potential of retrofits in the subsequent steps.

Fuzzy numbers were used to represent the energy performance of the base buildings and the retrofit upgrades. The key parameters to which the energy use was sensitive were used to characterize the aforementioned fuzzy number. Energy simulations were conducted considering three operational conditions to define the Lower and Upper Bounds and the Likely Values of the said fuzzy numbers. The Likely Value was determined considering the nominal operational conditions from the study by Prabatha et al. [13]. The Lower and Upper Bounds were defined by varying the nominal operating conditions. Operational energy savings under each fuzzy value (i.e., Lower Bound, Likely Value, Upper Bound) were calculated using Equation (3).

$${OES}_{i,j,k}={PreDemand}_{j,k}-{PostDemand}_{i,j,k};k∊[\mathrm{L}\mathrm{o}\mathrm{w}\mathrm{e}\mathrm{r},\mathrm{L}\mathrm{i}\mathrm{k}\mathrm{e}\mathrm{l}\mathrm{y},\mathrm{U}\mathrm{p}\mathrm{p}\mathrm{e}\mathrm{r}]$$

(3)

where

• ${OES}_{i,j,k}$—Operational energy savings of j^th energy source by i^th retrofit scenario
• ${PreDemand}_{j,k}$—Pre-retrofit energy demand for j^th energy source
• ${PostDemand}_{i,j,k}$—Post-retrofit energy demand for j^th energy source by the i^th retrofit combination

The fuzzy form of the operational energy savings is given below.

$${\underline{OES}}_{i,j}=[{OES}_{i,j,Lower},{OES}_{i,j,Likely},{OES}_{i,j,Upper}]$$

(4)

This fuzzy number was combined with the emission factors and energy prices to evaluate the operational performance of the retrofit strategies.

2.2.2. Environmental Performance

Global warming potential is commonly used to indicate environmental impact in climate change mitigation activities [22,43]. Life cycle emissions have been previously used as an indicator of global warming potential and environmental performance [13,44,45]. Therefore, life cycle emissions (kgCO_2e) were selected as the environmental performance indicator. Three main components, including embodied, operational, and disposal emissions, were considered in calculating the life cycle emissions of retrofits [13]. Embodied emission data from literature was adopted for this study [13]. Equation (5) was used to calculate the life cycle emission savings associated with each retrofit scenario.

$$\underline{LC{ES}_i}=-\sum _{\delta =1}^{N}E{E}_{i,\delta }+\sum _{j=1}^2\underline{O{ES}_{i,j}}\times \underline{E{F}_j}-\sum _j^N{DE}_{i,\delta }$$

(5)

where

• $\underline{LC{ES}_i}$—LCE savings resulted from i^th retrofit scenario; (in fuzzy form)
• $N$—Number of retrofits implemented in the i^th retrofit scenario
• $E{E}_{i,\delta }$—Added embodied emissions by the $\delta$^th retrofit in the i^th retrofit scenario
• $\underline{{OES}_{i,j}}$—Operational energy savings of the i^th retrofit scenario related to j^th fuel (NG, E); (in fuzzy form)
• $\underline{{EF}_j}$—Emission factor of the j^th fuel (NG, E); (in fuzzy form)
• ${DE}_{i,\delta }$—Disposal emissions of the $\delta$^th retrofit in the i^th retrofit scenario

2.2.3. Economic Performance

Life cycle cost (LCC) was used to evaluate the economic performance of different retrofit combinations accounting for capital, operational, and disposal costs [46]. The cost of equipment and installation was accounted for under the capital cost. RSMeans Building Construction Cost Database and Online vendor prices were employed in the costing. Equations (6)–(11) were adopted from Prabatha et al. for LCC calculations [13].

$$I{C}_{envelop,i}=\sum _{\forall \theta }IC\left(p,\theta \right)\times A_{\theta };\mathsf{\theta }\in [Walls,Roof,Windows,Doors,Foundation]$$

(6)

where

• $I{C}_{envelop,i}$—Initial cost of the envelope of the i^th strategy
• $A_{\theta }$—Area of the envelope component
• $IC(p,\theta )$—Initial cost to implement p^th modification option under the $\theta$^th envelope component

$$I{C}_{EnergySystem,i}=\sum _{\forall q}{ICE}_q$$

(7)

where

• $I{C}_{EnergySystem,i}$—Initial cost of i^th system retrofit strategy
• $IC{E}_q$—Initial cost of q^th energy system retrofit component

$$C{C}_i=I{C}_{envelop,i}+I{C}_{EnergySystem,i}$$

(8)

where

• $C{C}_i$—Total capital cost of i^th retrofit strategy

$${OCS}_i=\sum _{\forall j}OE{S}_{i,j}\times C_j$$

(9)

where

• $OC{S}_i$—Annual operational cost savings achieved by i^th retrofitting strategy
• $C_j$—Average unit cost of j^th energy source

$${DC}_i=\left(C_{LF}+F\times T\times C_T\right)\sum _{\forall \delta }{W}_{\delta }$$

(10)

where

• $D{C}_i$—Disposal cost associated with i^th retrofit strategy
• $W_{\delta }$—Amount of disposal waste (kg) produced by $\delta$^th retrofit option removed
• $C_{LF}$—Landfill cost per kg of waste
• F—Number of trips to the landfill
• $T$—The double distance between the landfill and the building (km)
• $C_T$—The cost of transportation per kg of waste per km

$$\underline{LC{CS}_i}=-C{C}_i+\left(\frac{{\left(1+r\right)}^t-1}{r{\left(1+r\right)}^t}\right).\underline{{OCS}_i}-D{C}_i$$

(11)

where

• $\underline{LC{CS}_i}$—LCC savings resulted from i^th retrofit scenario; (in the fuzzy form)
• $r$—Discount rate
• $t$—Project period

2.3. Performance-Based Retrofits Planning

LCCS and LCES were considered the key performance indicators (KPI) of retrofit performance, and they were calculated as fuzzy numbers to account for uncertainties. The relative importance of the LCCS and LCES can change depending on the priorities of the stakeholders. Policymakers and building owners are more interested in the LCES and LCCs, respectively. Importantly, a retrofit project is unlikely to proceed if the building owners’ expectations are unmet. Therefore, it is essential to establish the relative importance of the KPIs based on a stakeholder survey in practical implementation. The relative importance of LCCS (W₁) and LCES (W₂) was changed from 0–100% to explore the impact of decision priorities on the outcomes (ensuring that the sum of the two weights is always 100%). The overall performance score of each retrofit combination was calculated as shown in Equation (12).

$$\underline{F_i}=W_1\times \underline{LC{CS}_i}+W_2\times \underline{LC{ES}_i}$$

(12)

where

• $\underline{F_i}$—Overall performance score of the i^th retrofit scenario

$\underline{F_i}$ was calculated as a fuzzy number for all the retrofit strategies from 1 to n.

$$\underline{F_i}=\left\{x|f_{F_i}\left(x\right)>0\right\};i=\mathrm{1,2},\dots ,n$$

(13)

Maximizing and minimizing set method has been used for fuzzy number ranking in several previous studies [47,48]. This technique identifies the option closest to the best performance level and the furthest from the worst performance level as the best option. Based on this criterion, it ranks the retrofit strategies. This method is close to the fuzzy TOPSIS method, which is considered a reasonable approximation of the human decision-making process [43]. The normalizing method employed makes it possible to compare a large number of alternative solutions against each other. It also adapts the TOPSIS concept in a generalizable manner to a fuzzy environment, reducing the loss of fuzzy information [49]. Maximizing set (A) and minimizing set (B) are defined as shown in Equations (14) and (15), respectively. The membership of x to A and B is represented by $f_A\left(x\right)$ and $f_B\left(x\right)$, respectively [48].

$$A=\left\{\left(x,f_A\left(x\right)\right)|x\in R\right\}$$

(14)

where

$$\begin{gathered} f_A\left(x\right)=\left\{\begin{array}{c}\frac{x-x_{min}}{x_{max}-x_{min}}\hspace{1em},\hspace{1em}{x}_{min}\le x\le x_{max}\\ 0\hspace{1em},\hspace{1em}otherwise\end{array}\right. \\ B=\left\{\left(x,f_B\left(x\right)\right)|x\in R\right\} \end{gathered}$$

(15)

where

$$f_B\left(x\right)=\left\{\begin{array}{c}\frac{x-x_{max}}{x_{min}-x_{max}}\hspace{1em},\hspace{1em}{x}_{min}\le x\le x_{max}\\ 0\hspace{1em},\hspace{1em}otherwise\end{array}\right.$$

Parameters used in ${\mathrm{f}}_{\mathrm{A}}\left(\mathrm{x}\right)$ and ${\mathrm{f}}_{\mathrm{B}}\left(\mathrm{x}\right)$ calculations are defined below.

$$x_{min}=\inf \left(S\right);x_{max}=\sup \left(S\right);S=\bigcup _{i=1}^n{F}_i$$

The right, left, and total utility values for each retrofit scenario were calculated as seen in Equations (16)–(18), respectively.

$$U_A\left(F_i\right)=\mathit{sup}\left(f_{F_i}\left(x\right)\bigcap f_A\left(x\right)\right),i=\mathrm{1,2},\dots ,n$$

(16)

$$U_B\left(F_i\right)=\mathit{sup}\left(f_{F_i}\left(x\right)\bigcap f_B\left(x\right)\right),i=\mathrm{1,2},\dots ,n$$

(17)

$$U_T\left(F_i\right)=\frac{U_A\left(F_i\right)+{1-U}_B\left(F_i\right)}{2}$$

(18)

The retrofit combination that produces the highest ${\mathrm{U}}_{\mathrm{T}}\left({\mathrm{F}}_{\mathrm{i}}\right)$ is considered the best-performing retrofit strategy. Building owners and policymakers can use this method to identify the best retrofit strategies for a given building or a building cluster. If the requirement is to determine the best retrofit strategy for a given building, the effectiveness of alternative retrofit options will be modelled and ranked following the approach discussed above.

If the requirement is to identify the best retrofit strategies to be promoted in a community, a set of representative buildings will be selected. Then, the best retrofit strategies for those buildings will be identified using the proposed approach. Top-ranked retrofit strategies will indicate the best retrofit options and the building archetypes, providing the best opportunities to realize cost and emissions savings for the community.

2.4. Incentives Optimization

If incentives were assigned considering only the best retrofit strategy, some retrofit strategies having closer performance would be overlooked. Moreover, it is unlikely for all renovation projects across the community to choose the same retrofit strategy. Therefore, instead of incentivizing only one strategy, a set of top-performing retrofit strategies selected using the retrofits selection framework was considered for incentives planning. However, incentivizing retrofit strategies (combination of retrofits) is not practical as that increases the administrative burden on the incentive issuers and reduces the design flexibility for the incentive receivers (homeowners).

It was decided to incentivize individual retrofit options based on their appearance in top selected retrofit strategies to address the aforementioned challenges. The goal of the proposed incentive optimization approach was set as “to identify the best incentive levels that need to be assigned to individual retrofits to promote the top-ranked retrofit strategies”.

As explained by Equations (12) through (18), the top-ranked retrofit strategies achieve the highest performance scores (${\mathrm{U}}_{\mathrm{T}}\left({\mathrm{F}}_{\mathrm{i}}\right)$). The total utility of all retrofit strategies was compiled in a row matrix $\mathrm{U}$ (1 × n), where “n” is the number of strategies selected as suitable for incentivizing in the retrofit selection process discussed in Section 2.3.

Some retrofit strategies may not contain certain retrofits. The availability of individual retrofits in each retrofit strategy was represented in an availability matrix (A). A is a (m × n) matrix where “m” is the total number of retrofit options available to choose from. The elements of the availability matrix were defined as shown in Equation (19).

$$A_{i,\delta }=\left\{\begin{array}{c}1;\hspace{1em}if retrofit “\delta ” is available in scenario “i”\\ \hspace{1em} 0;\hspace{1em}if retrofit “\delta ” is not available in scenario “i”\end{array}\right.$$

(19)

If the retrofits were evaluated just based on the number of occurrences in the chosen retrofit strategy set, then a retrofit that appeared in a few top-ranked retrofit strategies would be incentivized similar to a retrofit that just made a similar number of appearances but in lower-ranked retrofit strategies. Therefore, to avoid the aforementioned issue, a performance index matrix P (m × n) was defined by elementwise multiplication of U and A, as shown in Equation (20).

$$P_{i,\delta }{=A}_{i,\delta }.U_i$$

(20)

The incentive amount assigned to individual retrofit options is the optimization variable and was represented by a (m × 1) matrix indicated by “$IN$” ($IN ϵ R$). The objective function was defined as shown in $ϵ R$ Equation (21).

$$Z=sum\left\{P^T.IN\right\}$$

(21)

Optimization problem:

The optimization problem was a maximizing problem with two constraints, as presented below. The incentive amounts that can maximize Z were selected as the optimal solution.

$$Z^{*}=\max \left\{Z\right\}$$

Bound to the following constraints:

• The incentive amount shall not exceed the capital cost of the retrofit ${IN}_i\le C{C}_i$
• Total incentives should not exceed the maximum budget allocation for incentivizing retrofits of the given building archetype

$$sum\left\{IN\right\}\le IncentiveBudget$$

The optimization problem was modelled using the MATLAB linprog solver.

3. Case Study

Kelowna, Canada (Climate zone 5, heating degree days: 3000–3999) was selected as the case study location for demonstration purposes. NG and electricity are Canada’s most common fuel types for space and water heating. Therefore, the two base buildings proposed by Prabatha et al. [13] were employed in this study to represent the different heating energy sources in Canadian single-family detached houses. BB-1 and BB-2 represent houses in which the space and water heating energy demand are catered by electricity and NG, respectively. In addition to the energy supply sources, all other characteristics of BB-1 and BB-2 are similar. Characteristics of the base buildings are summarized in Appendix A.

The retrofit performance and sensitivity analysis were done with reference BB-1 and BB-2. Even though these two representative buildings in Kelowna, BC were selected for demonstration purposes, the proposed framework can handle any number of building archetypes deemed necessary by the decision makers to sufficiently represent a given building cluster or community in any location (i.e., the application of the proposed framework is not limited to Kelowna or the two building archetypes involved or residential building sector). For demonstration purposes, total incentives offered for any retrofit strategy (hence, for a given building) were capped at CAD 10,000. The decision makers can adjust this constraint as applicable to their communities, depending on the budget availability.

Capital cost and embodied emission data of the retrofits are summarized in Appendix B. Space and water heating systems that required source switching from electricity to NG did not show any emissions-saving potential under any of the future scenarios considered. Therefore, those options were omitted from the BB-1 retrofits options list. However, both NG and electric space and hot-water heating upgrades were considered for BB-2. The current energy costs and operational emissions were used as the baseline for the comparison. A summary of the current energy source details of BC is presented in

Table 2. Emission factors and price for energy sources [13].

Source	Emission Factor	Energy Price
Electricity	0.0129 kgCO2/kWh	11.62 cents/kWh
Natural Gas	49.58 kg/GJ 0.1784 kg/kWh	7.36 $/GJ 2.65 cents/kWh

.

The annual energy use of the base buildings is summarized in

Table 3. Annual energy use of the base buildings [13].

Building	Electricity	Natural Gas
BB-1	56,767 kWh	-
BB-2	5804 kWh	6015 m3, 236 GJ, 65,556 kWh

.

3.1. Operational Uncertainties

The energy use of a building can significantly vary with occupants’ practices and operational conditions. Therefore, uncertainties introduced by these factors were modelled, as shown in

Table 4. Operational Uncertainties.

Parameter	Lower Bound of the Fuzzy Number	Average User (Likely Value of the Fuzzy Number)	Upper Bound of the Fuzzy Number
Number of adults	avg + 1	Three adults [13]	avg-1
Percentage time inside the house	60%	50% [13]	40%
Appliance, lighting, and other loads	90% of avg.	Base load information is summarized in Appendix C. [13]	110% of avg.
Domestic hot water consumption and temperature	197$l$, 53 °C	247$l$, 55 °C [13]	297$l$, 57 °C
Daytime heating temperature	20 °C	21 °C [13]	22 °C
Nighttime heating temperature	17 °C	18 °C [13]	19 °C
Setback duration	9 h	8 h [13]	7 h

.

3.2. Environmental Uncertainties

As seen in Equation (5), operational emissions are directly correlated to the emission factors of the energy sources. Therefore, the retrofit evaluation process considered potential variations of the NG and electricity emission factors. BC natural gas providers plan to make the domestic NG supply greener by injecting renewable NG (RNG) into the gas network. RNG has a smaller emission factor (0.2932 kgCO_2e/GJ) as opposed to the much higher emission factor of conventional NG (49.87 kgCO_2e/GJ). A major NG provider in BC predicts that RNG could cater to approximately 25–46% percent of the NG demand by 2036 with the technology developments in the biogas sector [10]. Two scenarios were considered assuming a gradual growth of RNG inclusion in the gas network to reach 25% and 46% by 2036. The same predicted NG inclusion growth rate was also assumed to continue from 2036 to 2040.

The average emission factor for the project period was represented using a fuzzy number accounting for alternative energy supply realities aligning with the predictions, as shown in

Table 5. Fuzzy conditions used to represent emission factors.

	Electricity Emission Factor	NG Emission Factor
Lower Bound	Stays at the present value	Reduces by 28.5% from the present value
Likely Value	Increases by 250% from the present value	Reduces by 15.5% from the present value
Upper Bound	Increases by 500% from the present value	Stays at the present value

These fuzzy numbers were used to represent electricity and NG emission factors.

. The Upper Bound of the emission factors was defined considering that the electricity emissions factor will stay the same while the NG emissions factor reduces due to RNG inclusion in the supply. Under the current conditions, renewable energy sources account for approximately 92% of the BC electricity grid’s installed capacity [50]. As of 2020, the installed generation capacity of the province is 18,286 MW [51]. According to the literature, the installed capacity is expected to increase at an approximate annual growth of 1% until 2040 [52]. In the Upper Bound (from Table 5), a 500% increment of the electricity emission factor was used to represent a future where the approximate demand growth (20% from current demand) in the next two decades is supplied by NG-based electricity generation stations. Likely Value presents a reality where NG-based generators supply 50% of the excess demand.

3.3. Economic Uncertainties

Energy prices of NG and electricity can affect the effectiveness of retrofit performance. Therefore, fuzzy numbers were developed to represent energy prices similar to the emission factors. Electricity price is linked with the grid energy mix and many other factors, including government policies [53,54]. BC’s current grid electricity price is 11.62 cents/kWh [13]. The current price was taken as the reference energy price for the study. Electricity price is expected to increase by 8.1% [54] in the coming five years, and it could be approximately represented by an annualized increment of 1.16% each year compared to the previous year. It was predicted to grow by 13.7% from 2013 to 2024 [54], and this prediction can be represented as an annualized value of 1.15%. If an average price hike of 1.155% was assumed for the project period (2020–2040), then the average electricity price is 12.96 cents/kWh. The aforementioned growth predictions are closely falling on each other. Therefore, the anticipated uncertainty is minimal. A variation of ±1% was introduced to the electricity price to account for uncertainties due to price fluctuation.

NG cost in the region considered is 7.89 CAD/GJ [55]. Approximately 5.5 CAD/GJ is charged for delivery, storage, and transport, while the cost of gas is 2.28 CAD/GJ. Over the past decade, gas prices have varied between 2–5 CAD/GJ approximately [56]. Following the great financial recession in 2008–2009, gas prices have hiked up to 10 CAD/GJ. The fuzzy number of the NG price was developed, assuming such price extremities will not be present in the coming two decades, and the NG cost will vary between 2–5 CAD/GJ range with an average of 3.5 CAD/GJ. Charges other than the gas cost were assumed constant in the project time. With these assumptions, the average gas price is 9 CAD/GJ. The Upper and Lower Bounds are 10.5 and 7.5 CAD/GJ, respectively.

4. Results and Discussion

The proposed research methodology was demonstrated through a case study. The detailed results and outcomes of the study are discussed in the following sections.

4.1. Sensitivity Analysis

The results of the sensitivity analysis are discussed below.

4.1.1. Energy Demand Sensitivity to Operational Conditions

PEDC against the operational conditions was observed. It was observed that PEDC variation with the increasing input and output values was symmetric (i.e., reductions in input parameters will create a mirrored pattern of what is seen in Figure 2 in the third and fourth quadrants of the graph). Moreover, two buildings (BB-1 and BB-2) considered in the study indicated similar trends in the sensitivity analysis. Therefore, the results related to BB-1 with increasing operational parameter values are presented in Figure 2 to examine the sensitivity analysis results. Detailed sensitivity analysis results are attached in Appendix D.

Figure 2 indicates that variations in the parameters related to the heating system, including the heating set points and setback duration, can significantly change the operational energy demand. This indicates the importance of smart controllers for temperature and setback duration control in residential buildings. In addition to smart controllers, occupants can adopt energy-conscious behaviours such as wearing an additional layer of clothing instead of increasing the heating setpoints.

The hot water temperature, HW demand, and major appliance load variations create comparable impacts on the PEDC. Even though these parameters do not create as high an impact as heating system parameters, demand savings achieved by altering these parameters can be easily achieved by adopting less intrusive energy efficiency solutions. (e.g., adopting water-efficient faucets, purchasing energy-efficient appliances when replacing the older ones, and adjusting the average water temperature in day-to-day use).

In general, the energy demand of a house is responsive to the number of occupants and the time spent at home parameters due to changes in energy demand for lighting, hot water, appliances, heating, and cooling. Figure 2 shows that when the number of occupants and time inside the house increase, the energy demand reduces. This energy demand reduction happens due to the energy savings from the space heating system with the increasing passive heat gains from the occupants. This observation is understandable as the heating energy demand dominates the building energy use in most parts of BC and Canada. Importantly, in HOT2000, baseload values do not automatically update with occupancy parameters. Therefore, the observed variation does not include the impact on energy use due to baseload variations resulting from the occupancy patterns. Therefore, baseload parameters must be manually adjusted when altering the number of occupants in HOT2000 models.

4.1.2. Operational Cost and Emission Sensitivity to Source Conditions

A 50% price variation was introduced to energy prices following the scenarios discussed in Section 2.1.2. The results are presented in Figure 3.

According to Figure 3, it is evident that even after a 50% increment in NG price or a 50% decrement in electricity price, the electrically heated house has a higher operational cost under the current conditions of the two buildings. However, BB-1 is only equipped with electric baseboards for heating. Suppose an electric heat pump was used for heating BB-1. In that case, there is a possibility of operating BB-1 at a lower cost compared to the NG-heated house (BB-2) when the electricity price drops to the lowest levels presented above.

In Figure 4, NG and electricity emission factors were varied following scenarios 1 and 2 from Section 2.1.2. In scenario 1, the NG emission factor was varied by ±50%. BC electricity demand is growing due to numerous factors, including population and industrial growth. On top of that, some government initiatives aggressively motivate the conversion of fossil fuel-based systems (such as building space heating and transportation) to electricity. With this new trend, alternative energy sources such as solar photovoltaic arrays, wind energy, and NG might be integrated into the grid if hydroelectricity becomes insufficient to cater to the provincial demand. However, according to a demand forecast by a major electricity supplier in the province, the total electricity demand growth from 2020 to 2040 is less than 20% [52]. A 500% increment in the electricity emission factor will be observed in an unlikely future scenario where the total provincial energy demand growth (20% from current demand) in the next two decades is supplied by NG-based electricity generation. As renewable energy sources already cater to the majority of the BC electricity supply, the potential for emission factor reduction is minimal. However, a 50% emission factor reduction was considered to study the potential operational savings in a scenario like that. In scenario 2, the electricity emission factor varied from −50% to 500%.

As seen in Figure 4, electrically heated houses (BB-1) produce much lower operational emissions under both scenarios. The operational emissions’ performance of the electrically heated house is better even after reducing the emission factor of NG by 50% and increasing the emission factor of electricity to 500% of the current value. It is important to note that the operational performance of an electrically heated house can be further improved by introducing heat pumps instead of the baseboards used in BB-1.

4.2. Retrofit Performance Outcomes

The priority assigned to LCES (W₂) was varied from 0% to 100% (i.e., the weight of LCCS (W₁) was varied from 100% to 0%) to understand the effect of the weights assigned to decision priorities on the percentage occurrence of building component upgrades in the best 100 ranked retrofit strategies. Figure 5 indicates the variation in the percentage inclusion of building component upgrades in top-ranked retrofit strategies vs. the relative importance of LCES. Here, the top-ranked 100 retrofit strategies (combinations) were explored in depth to quantify the number of times a particular retrofit type appeared in the solutions. For example, Figure 5 highlights that the HWU appeared in 80–100% of the solutions under all decision scenarios.

When W₂ is below 50%, all the upgrades in the top-ranked 100 strategies were proposed to be applied on BB-1. Similarly, when W₂ is above 50%, all the upgrade scenarios ranked in the top 100 options were proposed to be applied on BB-2. When LCES and LCCS were given equal weights, most upgrade scenarios in the first 100 ranks were proposed for BB-1, while 12% were proposed for BB-2. Notably, this is the only decision scenario (W₂ = 50%) where upgrades for both BB-1 and BB-2 were included in the best 100 options at the same time. The first two observations confirm the BC provincial government’s interest in retrofitting and source-switching initiatives focusing on houses heated with NG from an emissions point of view.

From Figure 5, it is evident that space heating upgrades are commonly selected under any weighting scenario. The HWU upgrades closely follow the same trend. However, HWU upgrade percentage inclusion reduces slightly below 85% when the relative importance of LCES is below 50%, yet being the second most used upgrade. Wall upgrades and air tightness are the third and fourth best-performing retrofits, irrespective of the decision priorities. Ceiling upgrades are included 50–70% of the time in the best 100 retrofit strategies, regardless of the decision priority. However, when the relative importance of LCES grows beyond 70%, 80%, and 90% (i.e., LCCS becomes less important), window upgrades become a common inclusion in the top-ranked retrofit strategies compared to ceiling, air tightness, and wall upgrades, respectively. Windows are disadvantaged compared to other retrofits due to the high capital cost. It is important to note that at 100% importance for LCES (i.e., 0% importance for LCCS), windows are included in all of the best 100 retrofit strategies, indicating that window upgrades have a significant emissions reduction potential. A detailed summary of the upgrades included in the best 100 ranked options under varying decision priorities is presented in Appendix E.

4.3. Incentive Performance

Incentive performance was evaluated under varying decision priorities. Only the top-ranked retrofit strategies were considered for incentivizing as the goal of the incentivizing is to promote the best. The incentive amounts applied to individual retrofits vary with the number of retrofit strategies (e.g., 50, 100, 500) shortlisted for incentivizing. The number of strategies incentivized can be decided based on the marginal abatement cost, LCCS, and LCES benchmarks established by the decision makers. For demonstration purposes, the incentive amounts applicable for individual retrofits were determined considering the best 50, 100, and 500 retrofit strategies out of the 2736 retrofit strategies considered. Detailed incentive optimization results are presented in Appendix F.

4.3.1. Optimal Incentive Allocation

In this discussion, the incentive amounts that can be optimally allocated to promote the top 100 retrofit strategies were analyzed for demonstrative purposes. In the top 100 retrofit strategies, when the LCES weight exceeds 50%, all the incentives focus on retrofits applied on BB-2. Similarly, when LCES weight is less than 50%, the incentives are focused on retrofits applied on BB-1. Optimal incentive amounts for each decision priority scenario are presented in Figure 6. The legend contains the capital costs of each retrofit option within brackets.

According to Figure 6, the Electric-HP HWU upgrade and airtightness upgrades are incentivized under all decision scenarios, and the incentives are proposed to match the capital cost of the retrofit. Upgrading walls to R-31 (5.459RSI) is the other retrofit proposed to be incentivized under all decision scenarios. Multi-split ASHP is also a retrofit proposed to be incentivized under any decision priority except when the only decision criterion is LCES. When the focus is more biased towards LCCS than LCES, the multi-split ASHP is incentivized to match its capital cost. These are the decision scenarios where the wall upgrade is not incentivized to match its capital cost. If a higher priority is given to LCES, then the proposed incentives cover the entire capital cost of the wall upgrades instead of multi-split ASHP. When 100% focus is on LCES, two-pane windows were incentivized instead of multi-split ASHP. This is understandable as the window upgrades cannot compete with the other retrofits in financial aspects due to their high initial cost.

4.3.2. Incentive Effectiveness

The incentive effectiveness can be evaluated considering the average LCES and LCCS reduction per dollar incentive. The incentive effectiveness variation with the changing weights (representing the variation in the relative importance of the key decision criteria) for the best 100 retrofit strategies is summarized in Figure 7.

The solid lines in both Figure 7a,b indicate the average performance that can be expected. The dotted lines indicate the possible range within which the performance can vary due to uncertainties such as emission factors, energy prices, and occupant behaviours. As seen in Figure 7, the decision priorities significantly impact per dollar LCES and LCCS potential of incentives. It is observed that LCES shows a rapid increment, and LCCS experiences a dip when the priority for LCES exceeds 50%. This can be mainly attributed to the characteristics of the base building to which the retrofits were applied. When the LCES priority is above 50%, all the top retrofit strategies are related to BB-2. When LCES priority is less than 50%, all the top retrofit strategies are proposed for the BB-1 house. This observation highlights that more LCES per dollar of incentives can be achieved by retrofitting houses with NG heating, while more LCCS savings can be achieved from the retrofits applied on BB-1.

Retrofitting BB-1 can produce 1-10kgCO_2e LCES per dollar incentive. Even though this is not matching the LCES levels of BB-2, this shows a greater potential for community penetration with the LCCS to building owners. If energy-conscious behaviours can be promoted, BB-1 can reach the upper threshold of emissions and cost savings (i.e., energy-conscious behaviour can approximately double the emission savings achieved by retrofitting an electrically heated house compared to energy-intensive occupant behaviours).

Retrofitting houses heated with NG makes more sense from an emissions-saving point of view because it can generate LCES as high as seven times the savings of a similar building heated with electricity. However, implementing retrofits on BB-2 with the view of emission savings can create life cycle financial losses in the range of CAD 0–25,000 even with the proposed incentives. Therefore, promoting energy retrofits to BB-2 owners is still challenging from an economic point of view.

When determining the financial viability of retrofit projects, capital cost should be considered in addition to the LCC figures, as high capital costs tend to drive away building owners from implementing retrofits. The case study indicates that the average CC of the chosen retrofit strategies notably increases with the increasing priority assigned to LCES, as the focus is mainly on emissions reduction rather than cost. When the importance given to LCCS exceeds 50% (i.e., LCES weight ≤ 50%), the average incentive level approximately covers more than 25% of the average CC of the retrofit strategies considered, and all the strategies are recommended for electrically heated houses. If higher LCCS are supplemented with richer incentives, retrofitting electrically heated houses can be easily promoted in the community compared to houses heated with NG.

In summary, it is safe to say that there is a limit to promoting emissions savings solely by overcoming the associated capital cost barriers via incentives. However, low-interest loans can significantly support building owners to overcome capital cost barriers in high capital cost options.

5. Conclusions

This study proposed a framework to find optimum incentive levels for individual energy retrofits to promote the best-performing retrofit strategies (combination of retrofits). The framework considers the life cycle emissions and cost-saving potential in selecting the best retrofit strategies for a given jurisdiction and the optimal incentive levels to promote them under uncertain conditions. The uncertainty-capturing capability of the framework allows the decision makers to identify the strategy, which is robust under future energy supply condition, weather, and macroeconomic changes. Accounting for uncertainties is crucial in climate policy development, given the highly volatile nature of the global economy and climate conditions.

The framework was demonstrated using two case study buildings in Kelowna, BC, with identical characteristics except for the heating energy source (electricity vs. NG). According to the case study, retrofitting the house with electric space and water heating (BB-1) showed greater LCCS potential with relatively lower LCES than retrofitting the house with NG space and water heating (BB-2). On the other hand, any retrofit applied on BB-2 led to higher LCES and lower LCCS compared to BB-1. Despite the high LCES potential of retrofit strategies applied on BB-2, associated net positive LCC (i.e., negative LCCS), particularly with source switching (from NG to electricity) make promoting them in communities challenging even with the proposed incentives. In such situations, other merits of electric retrofits can be leveraged to encourage source switching (e.g., heat pumps can produce cooling for summer while catering to the heating requirements in the winter as opposed to NG-heating counterparts; heat pumps can produce other benefits such as better indoor air quality).

Despite the economic challenges, switching from NG to electricity can significantly reduce the emissions of buildings in BC, which is crucial for meeting provincial emissions reduction targets. However, source switching can potentially increase the utility bills for the building owners and create reliability issues for the heating supply when a majority of the buildings depend on one energy source for heating in a country with extreme winters. With stepwise carbon tax increases, the financial disparity between NG and electric heating is reducing in BC. Even though carbon tax effectively reduces the reluctance for source switching by closing the economic performance gap between the sources, it does not reduce the actual cost of electricity usage for the occupants. Therefore, high energy costs could exacerbate other issues around housing affordability and energy security, which British Columbians already experience in the face of a predicted economic recession. Therefore, policy decisions around source switching extend beyond the climate change mitigation discussion. Understanding these challenges, the provincial government supports the community in source switching with general and income-qualified rebate programs with special attention to vulnerable populations.

Without any major physical retrofitting in houses, the combined effect of injecting RNG into the NG supply and adopting energy-conscious behaviour patterns by building occupants over energy-intensive practices can result in LCES over 100,000 kgCO_2e and LCCS closer to CAD 20,000 per house heated with NG, over the twenty-year project period considered in the study. The life cycle emissions saving achieved by increasing the RNG fraction in the NG supply and simply adopting energy-conscious behaviour in the NG-heated houses is almost two times the maximum emission savings achieved by any retrofit strategy applied to the house heated with electricity. This highlights the importance of raising community awareness regarding energy-conscious behaviour and the merits of employing performance-based incentives that can also reward energy-conscious behaviours, in addition to asset efficiency, to reduce emissions. Considering the discussed benefits of performance-based incentives and widespread usage of prescriptive incentives in BC, a mix of both approaches can expedite the realization of provincial residential emission reduction targets.

Limitations and Future Work

This paper only considered two base buildings representing Canadian averages for demonstration purposes. Nevertheless, the proposed incentive optimization strategy can comfortably handle multiple building archetypes and designs to cater to city-level planning. A comprehensive building cluster analysis and modelling are recommended for better accuracy in a city-level analysis.

The lack of locally applicable life cycle impact data for BERs, especially for HVAC and hot water systems, was identified as a major gap in the current body of knowledge. The accuracy of the life cycle emissions assessments can be improved with accurate embodied impact data.

The proposed framework can determine the optimal incentive levels to be assigned to individual retrofits to promote the desirable retrofit strategies for a given community, depending on the decision priorities. However, suppose the available budget is limited, and the incentive amount assigned to a given retrofit is not sufficient to change a purchase decision. In that case, the program uptake in the community can be lower. Therefore, the authors suggest conducting stakeholder engagement sessions and surveys to understand the minimum incentive levels needed to motivate the building owners to consider a given retrofit. The survey information can be integrated into the optimization algorithm to avoid assigning insufficient incentive amounts to impact purchase decisions and use those funds to strengthen the incentives for other retrofits that can potentially change the choice of building owners with the increased financial support.

Some energy-efficient and environmentally friendly retrofits (e.g., heat pumps) can be expensive compared to their counterparts. However, these retrofits can produce additional benefits, such as occupant comfort and health, which can be leveraged to promote them in communities. Therefore, the authors recommend the extension of the proposed framework to capture those aspects in future studies.