Energy efficiency in residential buildings amid COVID-19: A holistic comparative analysis between old and new normal occupancies

Graphical abstract

Abstract

Stringent lockdowns have been one of the defining features of the COVID-19 pandemic. Lockdowns have brought about drastic changes in living styles, including increased residential occupancy and telework practices predicted to last long. The variation in occupancy pattern and energy use needs to be assessed at the household level. Consequently, the new occupancy times will impact the performance of energy efficiency measures. To address these gaps, this work uses a real case study, a two-story residential building in the Okanagan Valley (British Columbia, Canada). Further, steady-state building energy simulations are performed on the HOT2000 tool to evaluate the resiliency of energy efficiency measures under a full lockdown. Three-year monitored energy data is analyzed to study the implications of COVID-19 lockdowns on HVAC and non-HVAC loads at a monthly temporal scale. The results show a marked change in energy use patterns and a higher increase in May 2020 compared to the previous two years. Calibrated energy models built on HOT2000 are then used to study the impacts of pre-COVID-19 (old normal occupancy) and post-COVID-19 (new normal occupancy) on energy upgrades performance. The simulations show that under higher occupancy times, the annual electricity use increased by 16.4%, while natural gas use decreased by 7.6%. The results indicate that overall residential buildings following pre-COVID-19 occupancy schedules had higher energy-saving potential than those with new normal occupancy. In addition, the variation in occupancy and stakeholder preferences directly impact the ranking of energy efficiency measures. Furthermore, this study identifies energy efficiency measures that provide flexibility for the decision-makers by identifying low-cost options feasible under a range of occupancy schedules.

Nomenclature

3E

Energy, Environment, and Economy (–)

ASHRAE

American Society of Heating, Ventilating, and Air Conditioning Engineers (–)

CV(RMSE)

Coefficient Of Variation of Root Mean Square Error (%)

dC_EOL

Difference in end-of-life cycle costs ($)

dC_I

Difference in initial investment costs ($)

dC_o

Difference in operational costs ($)

DHW

Domestic Hot Water (–)

dLCC_t

Difference in the total life cycle costs ($)

E_dmy

Sum of hourly median values (kWh)

EEMs

Energy Efficiency Measures (–)

EF_j

Emission Factor (–)

E_i

Median energy used in an hour (kWh)

E_j

Energy consumed (GJ)

I

Indicator for measuring energy variation

FN

Foundation (–)

GHG_annaul

Greenhouse Gas Emissions (kgCO₂eq.)

HVAC

Heating, Ventilation, and Air Conditioning (–)

ICF

Insulated concrete form

ILM

Intrusive Monitoring Methods (–)

LCC

Life Cycle Cost ($)

MBE

Mean Bias Error (%)

MCDA

Multi-Criteria Decision Analysis (–)

m_k

Measured data points (kWh)

NNS

New Normal Situation (–)

N_p

Number of data points (#)

ONS

Old Normal Situation (–)

OSB

Oriented Strand Board

PV

Solar Photovoltaic (–)

RF

Roof (–)

s_k

Simulated data points (kWh)

TOPSIS

Technique for Order of Preference by Similarity to Ideal Solution (–)

WL

Wall (–)

WN

Windows (–)

Subscript

k

Instance (#)

j

Type of fuel (electricity or natural gas)

t

Period under study (year)

t

Period under study (year)

1. Introduction

World Health Organization declared COVID-19 a global pandemic on March 11th, 2020 [1]. In order to maintain the spread of this disease, massive lockdowns came into effect throughout the world. Most countries blocked their borders, restricted the movement of their citizens, and even confined citizens in quarantine for weeks [2]. As a consequence of these lockdowns, disruptions occurred across sectors. In the economic sector, the pandemic reformed the economy and interrupted business by creating two significant changes [3]. The first change is the total or partial pause of travel, arts and entertainment, and personal services. The second is the extensive shift from in-office work, i.e., physically attending the workplace, to working remotely from home. Many organizations, including governments across the globe, encouraged or instructed people to work remotely. Consequently, many people have worked from home for the last two years. According to Statistics Canada [4], from March 22–28, about 39.1 % of people were working from home. These trends continued in 2021, and a report observed that 32 % of the employees (15–19 years of age) worked from home [5].

The impact of changing lifestyles due to COVID-19 had a profound effect on the energy sector [6]. The time and amount of energy consumption have substantially changed [7]. Under mass lockdowns, countries across the globe saw a reduction in national energy consumption (and associated emissions), with an overall global decrease of 6 % in the year 2020 [8]. However, these energy changes were not uniform and varied with the level of lockdown (full or partial), local infrastructure, socio-economic conditions, environment, and cultural practices [9]. Buechler et al. [10] determined the impact of pandemic lockdowns in 58 countries. The results showed stricter lockdowns and low transport mobility related to higher energy savings. Higher energy savings were recorded at the time of the first lockdown; for example, electricity demand in Australia fell by 6.7 % and for Italy by 3–4 % [11], [12]. Likewise, Prol et al. [13] work showed a decrease in electricity by 3–12 % for the first few months of the 2020 lockdown. Hence the impacts of lockdowns were not uniform, and there is uncertainty associated with the change in energy use at regional levels. Similarly, the existing literature indicates that though lockdown decreased energy and emissions related to the transport sector [14], there was an increase in residential sector energy consumption [15].

A number of studies related to COVID-19 and its effect on the energy sector have been carried out to date. However, the majority of the research is geared toward the whole sector, such as grid energy demand [13], transportation [16], public buildings [17]. In the residential sector, surveys both at the demand and supply sides have been performed to identify the effects of the pandemic on energy use and carbon [18]. US indicated 30 % increase [19], Spain up to 15 % [20] and the UK up to 2.1 % [21]. This change was also apparent in the type and amount of appliances usage. Studies also indicate that the usage time for television and computing instruments increased [22]. Krarti and Aldubyan [19] reported increased air-conditioning systems, appliances, and lighting in residential buildings. They suggested using energy star-rated equipment, smart thermostats, and photovoltaic systems to reduce the electricity cost. Likewise, an extensive study on 225 housing units revealed that heating, cooling, and air conditioning loads increased even under similar temperature settings. In short, little work has been done on individual residential building demand and potential increase due to working-at-home policies [7].

With the advent of the COVID-19 pandemic, remote working or telecommunicating was observed by more than 50 % of the population [23]. The increased occupancy impacted both residential HVAC and non-HVAC loads. The HVAC loads increased as people spent more time at home and needed to maintain indoor temperatures for comfort [24]; non-HVAC loads increased as many of the school and offices, and related activities shifted at home [7]. This implies that people had to combine living and working patterns giving rise to new occupancy profiles and peak times [9]. In addition, people started adopting indoor activities as a precaution or as a result of government lockdowns. A study showed an almost 47 % decrease in physical activity with lockdown compared to pre-pandemic times [25]. Even outdoor activities such as park mobility were decreased [26]. Hence, a new form of flexibility to work from home came into effect, and the pre-pandemic nine to five jobs, five days a week, may no longer be the norm for the majority of the population. According to a recent simulation-based study in Canada, people could spend 80 % more time at home compared to pre-pandemic times [27]. Hence, the impact of these higher occupancy times needs to be realized. Though recent trends in COVID-19 research have led to a proliferation of studies in the energy sector, most of the work is based on survey studies. Though survey and grid level studies provide variation in energy use, they fail to provide a change in occupancy patterns at individual building levels [28].

Energy grids can run smoothly without disruptions if the new peak loads and energy usage related to new occupancy patterns under the COVID-19 lockdown are accurately determined. Likewise, the longer occupancy times imply more energy costs for the occupants [7]. The new occupancy patterns under the COVID-19 lockdown must be accurately understood to ensure the energy grid runs smoothly under new peak loads and times. These higher energy costs will directly affect low-income households, so the energy incentives policies must be tailored for conditions under lockdowns [29]. Therefore, as the new occupancy patterns will likely continue after the pandemic [30], [31], these changes must be appropriately analyzed and quantified. High-resolution data based on detailed monitoring of individual buildings can identify the change in occupancy patterns and variations in direct energy use.

Since residential energy consumption is directly linked with occupant patterns, it is essential that the emerging occupancy patterns and variations in energy demand are incorporated into building energy simulation models. Energy efficiency measures (EEMs) improve buildings' energy conservation and carbon mitigation potential. The change in occupancy profile under COVID-19 adds uncertainty to EEMs energy and environmental performance and hence cost viability. Studies to date have concentrated on the impact of the pandemic on a building's overall energy consumption, and few studies have explored the feasibility of energy efficiency upgrades under COVID-19 occupancy schedules [32]. Building energy simulation models are the most common method to predict the performance of EEMs. However, the default occupancy profiles pre-built in these tools are not representative of new occupancy patterns emerging due to COVID-19 lockdowns. Occupancy profiles based on monitored data of the residential sector are still lacking [33]. Hence to assess the realistic performance of EEMs, these models need to be updated to consider the implications of new occupancy patterns (or new normal) under COVID-19. As a general rule, it can be assumed that under increased occupancy times with the COVID-19 lockdown, the energy-saving potential of these upgrades will be increased. However, to efficiently utilize these measures and related economic resources, a quantitative measurement of EEMs performance is essential. Therefore, as successive COVID-19 lockdowns continue or extended occupancy at homes becomes a norm, it is crucial to estimate the impacts of new occupancy on the EEMs [34].

Overall, there is a dearth of real-time data and its analysis to predict emerging occupancy profiles and energy demand related to how these lockdowns impact residential buildings [35]. Likewise, the research has started moving towards the impact of lockdowns. However, no research has determined the energy, environment, and economy (3E) potential of EEMs under the new normal occupancies.

To fulfill the gaps mentioned above and limitations this paper aims:

• (i)To determine changes in energy consumption and usage patterns due to COVID-19 lockdowns using monitored data;
• (ii)To identify impacts of pre-COVID and post-COVID occupancy on energy efficiency measures using building energy simulations; and
• (iii)To evaluate the impacts of stakeholders' perspectives on energy efficiency measures selection using a multi-criteria decision analysis method.

To achieve these objectives this, work first performs an in-depth study on monitored energy data of three years (2018–2020) on a single-family detached home in Okanagan Valley, British Columbia (Canada). It determines the variation in energy demand and occupancy patterns at monthly temporal scales for both HVAC and non-HVAC loads for the three years. The study then determines the energy, environmental, and economic savings potential on EEMs under pre-pandemic occupancy and full-lock down. Lastly, the study explores the impact of stakeholders’ preferences on the selection of EEMs using a multi-criteria decision analysis method. Here feasible EEMs are identified and ranked for stakeholders under the pre-pandemic occupancy and full-lockdown. Hence, the research goes beyond the existing trends of focusing on energy demand after COVID-19 and quantifies the impact of increased occupancy on EEMs from different stakeholders’ perspectives. Since the new occupancy patterns are likely to co-exist even as lockdowns are relaxed, this research can benefit both short- and long-term planning of future energy policies for residential buildings. The work provides unique insights for policy planners, utility providers, and home buyers to look into more feasible strategies to mitigate climate change under emerging occupancy scenarios. Fig. 1 provides an overview of paper flow from data collection and monitoring towards analysis and EEMs selection.

Fig. 1.

Paper Flow.

2. Materials and method

A multi-phase methodology is adopted to achieve the objectives mentioned above. The four main phases of the research are illustrated in Fig. 2 and explained in the following sections.

Fig. 2.

Research Methodology.(Where: EEMs: Energy Efficiency Measures; ONS: Old Normal Situation; NNS: New Normal Situation; MCDA: Multi-Criteria Decision Analysis).

2.1. Data collection and monitoring

2.1.1. Case study

The base case building is a two-story detached house located in the Okanagan Valley (50.05 N, 119.44 W), British Columbia, Canada, and occupied by four inhabitants (two working adults and two teenagers). The house complies with the BC building code 2012 [36] and is a representative of standard construction practices of the Okanagan region. The occupancy of the house represents 53.6 % of the total private dwellings in Canada and around 44.1 % in British Columbia in 2016 [37]. Weather profiles of the Okanagan valet indicate that the heating season is long and accompanied by temperatures falling below zero degrees [38]. Dry bulb temperatures vary between 24° and −26 °C, relative humidity falls between 76 % and 88 %, and average monthly precipitation is 35 mm, while the average monthly snow is 19.5 mm. Electricity and natural gas are the dominant energy sources in the region, while space and water heating are the primary energy uses.

The house was monitored using three types of sensors: temperature, humidity, and power sensors installed on both floors (See Fig. 3 .) from its occupancy in 2017 until August 2020. These sensors monitor variations in internal and external air temperatures, the temperature of the water heater, air humidity, and power usage (associated with electrical appliances and plug-in loads) every 5 min interval. Specifications of the sensors used are provided in Appendix A. Unlike previous studies largely dependent on non-intrusive monitoring methods, this study uses Intrusive Monitoring Methods (ILM). The ILM technique monitors energy usage for individual devices, i.e., sensing monitors were placed throughout the premises and communicated using a central hub. This method provides accurate data and assists in generating better energy models [39]. The house was used in previous studies by the authors to assess the effectiveness of incentives and household sizes [40], [41]. Here the work uses monitored data to evaluate the impact of COVID-19 on energy use and assess the implications of changing occupancy times on EEMs selection.

Fig. 3.

Plans of the case study home.

2.1.2. Occupancy profiles

The study used data from the literature review and field to develop the two occupancies.

Under COVID-19, the time spent at home has increased, resulting in increased energy use. Furthermore, the schedules typically adopted in an energy modeling tool do not represent the actual occupancy. Therefore, as people have started/will start spending more time at home due to COVID-19 or a similar emergency, the performance of EEMs needs to be evaluated to ensure their financial viability in generating low carbon emission buildings. Previous studies have considered three occupancies (i.e., pre-COVID lockdown, during COVID-19, and post-COVID-19) or more occupancies related to the stringency of COVID-19 quarantine measures [42]. Despite numerous studies on energy consumption under COVID-19 lockdowns, no work has considered the impact of these longer occupancy times on energy upgrade selection considering stakeholders’ perspectives. This research uses two occupancy situations called (i) Old Normal Situation (ONS) and New Normal Situation (NNS) to assess the impact of decision makers on EEMs selection. The term “new normal” has been extensively used in planning and development about anticipated changes due to COVID-19 [43], [44]. In this research context, the “new normal situation” defines the occupancy schedules that have become more common post-COVID.

Old normal situation (ONS): ONS represents the common occupancy profiles pre-COVID-19. The energy modeling guidelines for Canada recommend 50 % occupancy, which is the case adopted for ONS [45]. Here, the occupants spend half their day outside the home, such as in workplaces and schools.

New normal situation (NNS): NNS represents the occupancy profiles when occupants stay at home for a larger portion of the day. This profile emerged during COVID-19 lockdown times. The 100 % occupancy is an extreme case that would happen under the most stringent lockdown or an event that forces occupants to spend all their time at home. This situation is not expected to last over long periods. However, the analysis between a normal occupancy present pre-COVID and a 100 % occupancy does cover the whole range of intermediate situations. In addition to lockdown, other reasons (related to age, health, and job) can force people to spend large time indoors [46].

These different occupancies will give rise to different occupancy profiles. Hence, the ultimate goal of considering 100 % occupancy is to provide a holistic picture of future energy upgrade recommendations that would be feasible under all occupancy schedules.

2.1.3. Energy efficiency measures

The appropriate selection of energy efficiency measures (EEMs) will ensure lower energy input without compromising the comfort of the occupants. Passive, active, and renewal energy are three basic categories for energy performance upgrades. Improvement of thermal mass, thermal insulation, and windows upgrades are in passive energy upgrades since they does not include mechanical techniques [47]. Upgrading heating, cooling, and ventilation systems and domestic water heaters are characterized as active energy upgrades. Using ground source heat pumps and roof photovoltaic systems refers to renewable energy upgrades [48]. In Canada, non-heating energy upgrades such as lighting and appliances account for a small portion of upgrade expenses in residential buildings. Since the scope of this study was restricted to the energy upgrades in residential buildings; hence, only upgrades related to the space conditioning and water used were investigated. In this study, seven components—wall (WL), roof (RF), heating, ventilation, and air conditioning (HVAC), windows (WN), solar photovoltaic (PV), domestic hot water (DHW), and foundation (FN) have been selected for energy upgrades. The selected EEMs are readily available for residential buildings in the Okanagan Valley. The neighborhood developer agreed and validated these choices. Appendix A represents the technical specification of the house characteristics and the EEMs selected for this study. The main costs associated with these EEMs were collected from data provided by the local developer and RSMeans residential cost database [49]. The maintenance and replacement costs were determined using factors from existing literature, while operational energy costs were based on local utility rates [50]. The cost parameters and utility costs are presented in Table 1 and Appendix A, respectively.

Table 1.

Technical specifications of energy efficiency measures.

Energy efficient measures (EEMs)	Energy efficiency measures		Expected service life (Yrs)a	Annual maintenance (%)a
Space heating and cooling (HVAC)	Payne (PG92SCS) 17.58 kWh AFUE 92.1 % Natural Gas & Payne (PA14NC) 9.96 kWh 14 SEER Split A/C	Base Case	15	5
	5 series (500A11) – Geothermal c/w ECM variable speed blower (Heating 4 COP, Cooling 5.6 COP)
Domestic hot water (DHW)	Standard DHW system, 227 L [50 Gallons], EF 1.901 Standard	Base Case	15	5
	Hybrid electric water heater, 303 L [80 Gallons], EF 3.14, Electricity-based
Wall insulation (WL)	USI 0.28 [RSI 3.57] Batt	Base Case	30+	–
	9.525 mm [3/8″]EPS Styrofoam & USI 0.28 [RSI 3.57] Batt
Roof insulation (RI)	USI 0.26 [RSI 3.85] Batt, USI 0.14 [RSI 7.14] blown	Base Case	30+	–
	USI 0.28 [RSI 3.57] Batt, USI 0.11 [RSI 9.09] blown
Foundation insulation (FN)	USI 0.26 [RSI 3.85] Batt	Base Case	30+	–
	Insulated concrete form (ICF) blocks
Windows insulation (WN)	Vinyl double glazed windows c/w 180 low-E	Base Case	40+	2.8
	Vinyl triple glazed windows c/w 366 low-E
Solar photovoltaic system (PV)	–	Base Case	25	1.5
	10 panels (16.3 m2) azimuth 15′ and slope 25′

^aThe expected life time of EEMs and annual maintenance costs (as % of investments costs) are based on [54], [55], [56].

2.1.4. Energy model calibration

The base case energy model was constructed in the HOT2000 tool [51] from architectural drawings and equipment, appliances, and systems specifications provided by the building developer. These models were than updated with the occupancy data provided by the occupants and air-tightness values calculated from the blower door test [52]. This model was calibrated using real-time monitored data collected on-site. The accuracy of the calibrated model was checked through two benchmarks provided by the American Society of Heating, Ventilating, and Air Conditioning Engineers (ASHRAE) Guidelines [53]. These benchmarks are Mean Bias Error (MBE) (%) and the Coefficient of Variation of Root Mean Square Error (CV(RMSE)%). The smallest timescale at which energy results can be obtained from HOT2000 is at a monthly level. Hence, the base case simulation model was calibrated on a monthly scale. Both benchmarks are used, to sum up the number of errors and acceptable limits. If the quantity of error falls within these limits, the energy model is acceptable.

Otherwise, more calibration is performed through the iteration process until acceptable limits are reached. The two governing benchmark models are calculated as:

$$\mathit{M}\mathit{B}\mathit{E}\left(\mathbf{\%}\right)=\frac{{\sum }_{\mathit{k}=1}^{{\mathit{N}}_{\mathit{p}}}{\mathit{m}}_{\mathit{k}}-{\mathit{s}}_{\mathit{k}}}{{\sum }_{\mathit{k}=1}^{{\mathit{N}}_{\mathit{p}}}{\mathit{m}}_{\mathit{k}}}$$ (1)

$$\mathit{C}\mathit{V}(\mathit{R}\mathit{M}\mathit{S}\mathit{E})\left(\mathbf{\%}\right)=\frac{\sqrt{\sum _{\mathit{k}=1}^{{\mathit{N}}_{\mathit{p}}}{\left({\mathit{m}}_{\mathit{k}}-{\mathit{s}}_{\mathit{k}}\right)}^2/{\mathit{N}}_{\mathit{p}}}}{\overline{\mathit{m}}}$$ (2)

where m_k and s_k are the respective measured and simulated data points for each model instance 'k'; Np is the number of data points at interval 'p' (i.e., N_monthly = 12) and $\overline{m}$ is the average of the measured data points. The acceptable error limit for MBE is ≤±5 %, while for CV (RMSE) is ≤+15 %. Based on calibration, the MBE and CV(RMSE) came to be +0.99 % and +3.45 %, respectively, hence within an acceptable range.

2.2. Monitored data analysis

Data from January 2018 to August 2020 was utilized for the study. In British Columbia, the main lockdown came into effect from the beginning of March 2020 and continued to the summer of 2020. The research team was could not access data beyond August 2020 as the project associated with data collection was officially closed at this time. Hence for this study, data related to the year 2018 and 2019 present pre-COVID periods, and data for the year 2020 is related to post-COVID occupancy. Due to this data limitation, only the data from of January to August was compared for the three years: 2018, 2019, and 2020. Another data limitation occurred due to a malfunctioning sensor collecting a load of the HVAC system, due to which majority of the data for May 2020 was lost.

Data collected from the sensors were analyzed in three main categories: HVAC loads, non-HVAC loads, and overall loads. For this study, HVAC loads comprise energy use from the natural gas furnace, air conditioner, and HRV. Non-HVAC loads include energy use from lighting systems, domestic hot water, refrigerator, washer, dryer, and plug-in loads. Overall loads are the sum of HVAC and non-HVAC loads. This categorization is necessary to separate the impact of external weather changes on home energy usage [7]. Previous studies have shown a high correlation between external weather with HVAC load [57]; hence the external environment impacts need to be considered to evaluate variation in HVAC loads before and after COVID-19 lockdowns. Compared to HVAC loads, non-HVAC loads are not dependent on environmental conditions [58]. For non-HVAC and overall loads the medians hourly loads were determined for each month and plotted against time of the day. The rate of change metric for non-HVAC loads was also determined to measure the increase and decrease in load for each hour. Though overall loads depend on external weather, they were plotted against the time of day to show variation in overall energy.

Two main metrics are used to assess variation in non-HVAC and overall loads. Daily energy consumption based on hourly median values in one month of a particular year was calculated using Eq. (3).

$${\mathit{E}}_{\mathit{d}\mathit{m}\mathit{y}}(kWh)=\sum _{i=1}^{24}{E}_i$$ (3)

where, E_dmy represents the sum of hourly median values in one month of a particular year; i is the hour of the day, and Ei is the median energy used in that hour. Variation between energy spent in months of 2018 and 2019 compared to those in COVID year 2020 was quantified using an indicator $I$ Eq. (4). This metric helps to consider both the increase and decrease of energy at different times by considering proportion of the hourly change in energy. A higher percentage of $I$ indicates a more significant divergence and vice versa [18]. This indicator also helps in comparing energy variation between 2018 and 2019.

$$\mathit{I}(\%)=\sqrt{\frac{1}{24}\sum _{i=1}^{24}{\left(\frac{{\mathit{Ei}}_{2020}}{\sum _{i=1}^{24}{\mathit{Ei}}_{2020}}-\frac{E_i}{\sum _{i=1}^{24}Ei}\right)}^2}$$ (4)

where, ${\mathit{Ei}}_{2020}$ represents hourly median energy values in one month of year 2020; Ei is the median energy used in that hour of the pre-COVID year.

Change in HVAC energy over time is determined by constructing regression models [7]. These models were evaluated based on the value of R-squared (coefficient of determination). Models were considered good quality if their coefficient of determination greater than 0.75 [59]. The models generated for HVAC loads can be conceptually represented as Eq. (5) .

$$\mathit{y}=\alpha +\sum _{i=1}^{24}{\beta }_i\left(Drybulbtemperature\right)+{\epsilon }_i$$ (5)

Where y is the daily energy consumed by HVAC at a given dry bulb temperature, α is the intercept between dry bulb temperature (°C) and Total Daily HVAC Loads (kWh), ${\beta }_i$ is the regression coefficient, $i$= 24 and ${\epsilon }_i$ is the error value that is not correlated with the dry bulb temperature.

2.3. Data modelling

The data modelling phase involves three steps: energy, carbon footprint, and cost modeling. The energy modeling process was performed using the HOT2000 (V11.10) energy modeling tool, while both carbon footprint and cost modeling were performed using an excel spreadsheet prepared explicitly for this research.

2.3.1. Energy modeling

HOT2000 simulations provided annual electricity and natural gas consumption under various energy upgrade combinations. The HOT2000 tool is specifically designed for residential buildings in Canada and is extensively used by building construction practitioners and researchers in Canada [51]. The tool uses the information on local weather, occupancy, home characteristics, floorplans, equipment, and energy sources to create a standard energy model. The tool operates on steady-state models and behaves like a gray-box program to determine dwellings’ energy demands [60]. HOT2000 uses the Alberta Air Infiltration Model (AIM2), Mitalus, and bin methods to evaluate monthly energy loads [41], [61].

Calibrated energy model representing occupancy of ONS occupancy was tested with all possible combinations of EEMs (Table 2 ). The same base case model was then updated with 100 % occupancy to represent NNS and analyse the impacts of EEMs. It should be noted that, unlike EnergyPlus and TRNYSYS programs with pre-defined occupancy simulation codes [62], data related to lighting energy usage, plug loads, temperature set points, and amount of time spent indoors needs to be added manually in HOT2000. Though the massive increase in lighting and plug loads under COVID-19 lockdowns are well documented [63], these variations were beyond the scope of this work and hence not updated for NNS. EEMs combinations were tested for the two occupancies using parametric analysis to evaluate annual energy usage. The parametric analysis uses a base case modeled by changing one variable (energy upgrade for this research) at a time, keeping other variables constant [64]. This method is suitable only for limited variables as the computation and time requirements increase with the number of variables. In this research, the parametric analysis was found feasible as only a limited number of energy upgrade options were considered. The seven energy upgrades were parametrically modeled so that, in total, 127 simulations, including a base case, were conducted for each occupancy. The annual energy consumption for these 127 combinations was used to evaluate environmental and economic parameters.

Table 2.

Weighting scenarios for the energy upgrades' decision criteria.

Scenario	Description	Energy resources	Environmental	Economic
		Energy consumption (GJ)	Carbon emissions (kgCO2eq.)	Payback period (Years)
S1	Pro-Energy	1	0	0
S2	Pro-Environmental	0	1	0
S3	Pro-Economical	0	0	1
S4	Neutral (Equally important)	0.33	0.33	0.33
S5	Intermediate	0.75	0.25	0
S6	Intermediate	0.75	0	0.25
S7	Intermediate	0.5	0.5	0
S8	Intermediate	0.5	0	0.5
S9	Intermediate	0.25	0.75	0
S10	Intermediate	0.25	0	0.75
S11	Intermediate	0	0.25	0.75
S12	Intermediate	0	0.5	0.5

2.3.2. Environmental modelling

The annual carbon footprint was evaluated using the energy simulation results for each set of energy upgrade combinations. In Okanagan, BC, residential buildings are operated on two main energy sources: electricity and natural gas. Compared to other regions of the world, the electricity provided to homes in the Okanagan Valley has a very low emission factor. This is because a large portion (about 95 %) of the energy mix is based on renewable energy [65]. The natural gas emission factor (49.87 kgCO₂eq./GJ) is almost 18 times higher than electricity emission factor (2.8 kgCO₂eq./GJ) [66]. The annual carbon footprint for each combination is evaluated by Eq. (6):

$${\mathit{G}\mathit{H}\mathit{G}}_{\mathit{a}\mathit{n}\mathit{n}\mathit{a}\mathit{u}\mathit{l}}={\sum }_{\mathit{j}=1}^{\mathit{n}}{\mathit{E}}_{\mathit{j}}\mathbf{\ast }{\mathit{E}\mathit{F}}_{\mathit{j}}$$ (6)

where GHG annual is the sum of carbon emissions by all energy resources over a year, j is the type of fuel (electricity or natural gas), E_i represents the yearly energy consumed using the fuel j, n is the number of energy sources, and EF_f is the associated emission factor (kgCO₂eq./GJ).

2.3.3. Economic modeling

The life cycle cost (LCC) method evaluated costs associated with energy upgrade alternatives. LCC analysis accounts for costs in a product or building [67], [68]. The relative LCC of the energy upgrades were evaluated using Eq. (7) based on US Federal Energy Management Program [69]. This model is adopted because it requires less information that can be used effectively in decision-making

$$\mathit{d}\mathit{L}\mathit{C}{\mathit{C}}_{\mathit{t}}=\mathit{d}{\mathit{C}}_{\mathit{I}}+\mathit{d}{\mathit{C}}_{\mathit{O}}+\mathit{d}{\mathit{C}}_{\mathit{E}\mathit{O}\mathit{L}}$$ (7)

where $\mathit{dLC}{C}_t$ is the difference in the total life cycle costs of base case and an energy upgrade combination, t is the period under study (30-year for this research), $d{C}_I$ is the difference in the initial investment costs, $d{C}_O$ is the difference in the operational costs (energy, maintenance, and replacement), and $d{C}_{\mathit{EOL}}$ is the difference at the end-of-life cycle costs. In addition to relative life cycle costs, economic criterion, payback period, was calculated as:

$$\mathit{P}\mathit{a}\mathit{y}\mathit{b}\mathit{a}\mathit{c}\mathit{k}\mathit{P}\mathit{e}\mathit{r}\mathit{i}\mathit{o}\mathit{d}=\mathit{d}\mathit{L}\mathit{C}{\mathit{C}}_{\mathit{t}}/\mathit{d}{\mathit{C}}_{\mathit{O}}$$ (8)

For this research, the life period of 30 years is adopted because 25–30 years is the usual mortgage time for residential buildings in Canada [40]. In the same vein, recent studies on residential buildings in Canada have used a 30-year lifetime [70]. In addition, the end-of-life cycle costs were considered negligible, and only costs involved in the investment and operation stage formed part of decision making. It was also assumed that over 30-year the annual energy consumption remains constant.

2.4. Decision making

2.4.1. Scenario development

The construction industry deals with many stakeholders (developers, contractors, governmental policymakers, and end-users). These stakeholders have diverse preferences regarding the importance of different decision criteria for a green building project [71]. In most cases, construction industry practitioners value economic sustainability more than the overall sustainability of different construction methods [72]. Though governments and building professionals have started promoting sustainable building practices [73], the higher initial investment costs are still one of the main criteria in decision-making [74].

Here three parameters: energy, environment, and economy (3E), are the main decision criteria. These have been extensively used in research related to sustainable building design due to their important role in achieving mitigation goals [68]. Existing literature has emphasized the need to consider stakeholders' priorities for different criteria since using equal weights for different criteria will provide misleading results [69]. In this research, twelve weighting scenarios were developed to address variation in stakeholders' preferences on the environmental and economic criteria. Table 2 lists the three main decision criteria (i.e., annual operational energy, annual carbon emissions, and payback period) and the twelve weighting scenarios defining stakeholders' preferences for the decision criteria involved in the decision-making process. Though the amount of energy use is directly proportional to the operational carbon emissions, for this research, they are considered two separate parameters. The main reason for keeping these separate is the high variation in the carbon factor of natural gas (49.87 kgCO₂eq./GJ) compared to electricity (2.8 kgCO₂eq./GJ) [66]. Hence, even if two retrofits give the same energy savings, their carbon footprint can be significantly different depending on the specific type of energy (natural gas or electricity) reduced. Likewise, the weighting given to energy and operational carbon footprint varies from one stakeholder to another. For instance, a local government may prefer GHG reduction, while utility providers may give more weight to energy to reduce grid loads. The TOPSIS MCDA aggregation method was employed to rank these upgrade combinations under the two occupancy profiles (ONS and NNS homes).

2.4.2. Topsis MCDA

Multi-Criteria Decision Analysis (MCDA) is a well-recognized method in sustainability studies [75], [76]. Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) is a powerful MCDA technique. This technique identifies and ranks alternatives that are closest to the ideal solution and furthest from the worst [75]. TOPSIS compares a set of alternatives selected from pre-specified criteria. Researchers have utilized this method to rank energy upgrade options in a residential building due to its ability to find the closest alternative to the positive ideal solution and farthest from the ideal negative alternative [77]. The steps followed in the TOPSIS method are presented in Appendix B.

3. Results and discussion

3.1. Monitored data analysis

This section aims to determine the changes in energy consumption during COVID-19 lockdowns based on the data collected from sensors. This section is organized into three main categories based on monitored data from 2018 to 2020 to show energy use variation for non-HVAC, overall, and HVAC loads. First, the variation in non-HVAC loads is evaluated for each month, followed by overall load variation shown as median hourly energy change per month of the year and HVAC loads normalized with respect to external dry bulb temperatures.

Fig. 4 Plot shows the median hourly non-HVAC loads for eight months (January to August) for 2018, 2019, and 2020. The median non-HVAC loads (kWh) per day of the month are shown across the y-axis, while x-axis represents the hours of the day. Starting from January and moving towards August, a clear trend of increase in the energy loads for the year 2020 is observed. The increase in energy use is more pronounced in the months of April, May, and June, marking the most stringent lockdown period. The average for these three months has increased from average of 15.8 kWh per day in 2019 to 23.6 kWh per day for the corresponding lockdown months of 2020. $I$ value calculated for each month of the year show that highest change is observed for May when data of 2018 and 2019 are compared with 2020. For 2018 this variation was up to 15 % more, and for 2019 the increase was 11 %. For the majority of the months almost a constant increase of energy (7–8 %) was observed for the two years (2018 and 2019) compared to 2020; for instance, 5 % increase in March, 8 % increase in June, and 8 % increase in July. For August of 2018 large variation is observed. Looking at Fig. 4 . It is observed that unusually low energy is used for this month. There can be numerous reasons for this form of energy use pattern. However, since the project was officially closed on September 2020, we could not consult occupants for this low usage. However, the variation of 2019 almost shows a similar variation of 7 %. In addition, the peak times of energy use have moved later in the day. This change in pattern is in line with previous studies [18], [78] that showed occupancy times and associated energy increased after COVID-19. This trend is present as an early morning commute is not required, and people working from home are spending more time indoors compared to pre-COVID lockdowns [7], [79]. Similarly, more use of household appliances and plug-in loads accompany this increased time spent at home.

Fig. 4.

Median hourly non-HVAC loads for months of year 2018, 2019 and 2020.

The highest increase is observed in the early hours of the day and later in the evening. This shows that during COVID lockdowns, the occupancy habits especially associated with sleeping times were impacted. It is also seen that August shows the highest increase from normal non-HVAC loads compared to other months. A possible reason for this much higher increase in energy during the COVID-19 period may be associated adoption of more indoor activities [7] or limited access to outdoor activities, as most provincial parks and camping sites were either partially or fully closed to public access during the lockdown periods [80].

In order to more holistically analyze the impact of COVID-19 lockdowns on the energy use of the case study home, the median hourly loads over the 24-h period were also analyzed. Fig. 5 shows the variation in overall loads (HVAC and non-HVAC loads) for all months (January to August) for the three tested years. The y-axis represents the median hourly overall loads, while the x-axis represents the time of the day (hour). The results show an average increase of 18 % and 7 % from 2018 and 2019, respectively, to 2020. The average loads for the eight months are 23.3 kWh, 25.6 kWh, and 27.4 kWh for 2018, 2019, and 2020 respectively. It is observed that similar to non-HVAC load curves, the curves for overall loads are, on average higher than in pre-COVID years. Interestingly addition of HVAC loads adds more pronounced peaks. Since energy use in the HVAC system is dependent upon local environmental conditions, energy in some months of the year 2019 surpass the peak lockdown period. Hence, considering the influence of external environmental conditions is essential. The increase in energy use is observed for the three lockdown months, March, April, and May. $I$ values for overall loads showed extremely high variations. For example, the highest $I$ value was found for May at 23 % and 11 % for 2018 and 2019 compared to 2020, respectively. However, the majority of HVAC data was lost for May 2020; hence this value may not be a true representative of data. An increase in overall load was also observed for the summer months of June, with a 13 % increase to 2018 and a 14 % increase to 2019. Since these loads are not normalized against weather conditions, the variation is likely lower than these values. In contrast to non-HVAC loads, which showed the highest variation for May for both 2018 and 2019, the highest $I$ was present in May for 2018 (of 23 %) and in June for 2019 (of 14 %). For remaining months 4–7 % change in $I$ was observed. Similar to non-HVAC data, the results for $I$ value of August 2018 are an outlier and should be excluded when comparing variation in overall energy consumption.

Fig. 5.

Median hourly overall loads for each month of year 2018, 2019 and 2020.

Regression models for the HVAC loads compared to dry-bulb temperatures are presented in Fig. 6 . Second-order polynomial regression model was found to provide the best-fit for the models. Other researchers have also found that HVAC system performance can be represented through quadratic regression models [81], [82]. Here the daily HVAC loads for each day of the month are plotted against the average external temperature data collected from the thermal sensors. The regression models for three years and three summer months, June, July, and August were generated. Unfortunately, due to malfunctioning sensor data of the HVAC system, more than 90 % of the data for the month of May 2020 was lost, and a regression model for comparison with May 2018 and May 2019 could not be created. Based on the available data, the generated regression models and the related equations and coefficients of determination can be seen in Fig. 6. The normalization of the HVAC against external temperature at the time of use shows that for June and August, the loads for 2020 is more than either of the previous two years. In the case of July certain degree of overlap is observed for the COVID and pre-COVID periods. The higher use of HVAC is related to longer periods spent at home. On non-HVAC loads also. Compared to the pre-COVID period, when the lower thermostat temperatures were not required to be maintained all the time, the energy used for running HVAC systems was also low.

Fig. 6.

Regression models developed for total daily HVAC loads of three summer months of 2018, 2019 and 2020.

Overall, there is an increase in the energy use for non-HVAC, HVAC, and overall loads of the house for post-COVID period compared to the pre-COVID. However, there is variation in the amount of energy increased in different months. For instance, the overall average energy consumption per day increased from 2018 for 2020 year from 21.6 kWh/day to 27.9 kWh/day in March, 23.6 kWh to 28.9 kWh/day in April, and 19.0 kWh/day to 25.4 kWh/day in May In addition, depending upon the weather conditions, the highest changes in non-HVAC and HVAC loads can occur in different months of the lockdown period. For example, for February, average daily loads increased by 23 % for non-HVAC and decreased by 36 % for HVAC loads for 2018 compared to 2020. The weather history of this month showed that the average temperature was −4.9 °C for 2018 and −0.1 °C for 2020 [83]. Hence, more energy for HVAC was required in 2018 compared to 2020, which is evident from the decrease in HVAC load (−36 %) for the two years. These results accord with recent studies performed for energy use during COVID-19 lockdowns, such as work by Abdeen et al. (2021) [63] work on residential buildings in Ottawa, Canada.

3.2. Modelled data analysis

In this section, results of the data analysis related to energy, cost, and carbon footprint modelling are presented and discussed. The data analysis aims to identify energy upgrades that can be useful for the tested occupancies (ONS and NNS) and stakeholders' preferences. First, the results of energy simulations and annual carbon emission reduction potential due to different energy upgrade combinations are presented. Then, the investment analysis of the energy upgrade combinations is presented, followed by the results of MCDA analyses for ranking the energy upgrade combinations.

As stated earlier, seven energy upgrades have been selected with the potential for energy conservation and environmental impact reduction. A total of 127 combinations, denoted by C1, C2, …, and C127, were developed for each ONS and NNS home. Each upgrade combination was separately applied to the base home with the help of the HOT2000 tool. Subsequently, for each upgrade combination, the results of energy consumption (sum of natural gas consumption and electricity consumption) (in GJ/year), payback period (in years), and operational carbon emissions (in kg.CO₂eq/year) were extracted. Due to space limitations, the detailed results have been presented in Appendices C and D.

3.2.1. Energy analysis

The energy analysis results revealed that each energy upgrade combination could result in different amounts of reduction in energy resources, environmental impacts and, payback periods for two occupancies, ONS and NNS. As Fig. 7 (a) shows, the energy-saving due to different upgrade combinations can range between 36.5 GJ and 116.3 GJ in the ONS home and between 47.6 GJ and 119.5 GJ in the NNS home.

Fig. 7.

Chart of annual energy consumption and box-plots of savings potential of energy efficiency measures.

Two main groups are visible for tested combinations. These groups are formed due to geothermal heat pump, which has a much higher energy-saving potential than other EEMs. Hence, any combination that contains an HVAC upgrade falls into the lower energy tier. Likewise, it is also observed that the savings due to the HVAC upgrade are higher for ONS compared to NNS. The figure also contains a box-plot that shows a range of potential energy savings in percentage for two occupancies.

It can be observed that the average energy saving potential of ONS is higher (37 %) compared to NNS (32 %) for the two profiles. Although some EEMs (WL, WN, FN, and PV) provide a slightly higher percentage of energy savings for NNS than ONS. Marshall et al. [46] work on occupancy profiles in U.K. dwellings also showed that wall insulation had higher energy-saving potential for longer occupancy times. Overall, the saving potential is higher for ONS than NNS, mainly due to the more saving potential of HVAC in ONS than NNS. The main reason for less saving from HVAC systems is that space heating is still the largest source of energy consumption in Canadian dwellings [84]. When people spend more time at home, amount of internal heat gained from the occupants and building systems (such as lighting) also increases [85], [86]. This results in decreased heating load demand, resulting in less savings from the HVAC system.

3.2.2. Environmental analysis

Similar to energy consumption, EEMs can reduce operational carbon emissions. In contrast to energy consumption, the annual emissions are higher for NNS upgrades than for ONS, as illustrated in Fig. 8 (a). This unexpected result is related to the heating load requirements of the home. When people spend more time at home, the demand for heating load decreases due to more internal heat gains from the metabolism and activities of occupants. In addition, longer occupancy is also accompanied by the use of building energy systems and appliances that also produce some heat energy [85], [86]. Therefore, despite an overall lower annual energy consumption in ONS than in NNS, more energy is consumed for heating purposes in ONS. These changes are clearly seen in the case of energy measure combinations that do not contain an HVAC upgrade (which is operated on electricity). As natural gas has 18 times higher emissions than electricity, a slight variation in its use increases overall ONS emissions. Fig. 8(b) shows that the range of carbon emission reduction potential due to the energy upgrade combinations is almost the same for both ONS and NNS and can reach up to 95 % due to environmentally friendly HVAC upgrades.associated with DHW, CA$(-1040) for ONS, and CA$ (−632) for NNS. It can be seen that the only difference evident in the costs of two occupancy profiles, ONS and NNS, are the associated energy costs. Except for annual energy costs, all other costs are linearly added to determine the costs of energy efficiency combinations. Appendix C contains details of relative costs associated with all tested combinations.

Fig. 8.

Chart of annual carbon emissions and box-plots of reduction potential of energy efficiency measures.

3.2.3. TOPSIS MCDA results

Table 3, Table 4 show the top ten energy upgrade combinations concerning the three extreme scenarios (pro-energy, pro-environmental, pro-economic) and the neutral scenario for the ONS and NNS homes, respectively. These four scenarios are of particular importance as they represent the extreme positions; therefore, any variation in weighting would lie in between them. Due to space limitations, complete results from TOPSIS are provided in Appendix D.

Table 3.

Top 10 energy upgrade combinations under four extreme scenarios in the ONS home.

Rank	S1 (Pro-Energy)	S2(Pro-Environmental)	S3(Pro-Economical)	S4(Neutral (equally important))
1	WL RF HVAC WN PV DHW	WL RF HVAC WN PV DHW	DHW	RF HVAC WN PV
2	WL HVAC WN PV DHW	WL HVAC WN PV DHW	RF DHW	WL RF HVAC PV DHW
3	WL RF HVAC WN PV DHW FN	WL RF HVAC WN PV DHW FN	WL DHW	WL HVAC PV DHW
4	WL HVAC WN PV DHW FN	WL HVAC WN PV DHW FN	WL RF DHW	RF HVAC PV DHW
5	RF HVAC WN PV DHW	RF HVAC WN PV DHW	RF HVAC WN PV	HVAC PV DHW
6	HVAC WN PV DHW	HVAC WN PV DHW	PV DHW	WL HVAC WN PV DHW
7	RF HVAC WN PV DHW FN	RF HVAC WN PV DHW FN	WL PV DHW	WL RF HVAC WN PV DHW
8	HVAC WN PV DHW FN	HVAC WN PV DHW FN	RF PV DHW	HVAC WN PV DHW
9	WL RF HVAC PV DHW	WL RF HVAC PV DHW	WL RF PV DHW	RF HVAC WN PV DHW
10	WL HVAC PV DHW	WL HVAC PV DHW	HVAC DHW	WL RF HVAC PV

Note: The details of energy upgrade rankings for all twelve scenarios in the ONS home can be found in Appendix D.

Table 4.

Top 10 energy upgrade combinations under four extreme scenarios in the NNS home.

Rank	S1(Pro-Energy)	S2 (Pro-Environmental)	S3(Pro-Economical)	S4(Neutral (equally important))
1	WL RF HVAC WN PV DHW FN	WL RF HVAC WN PV DHW FN	DHW	WL RF HVAC PV DHW
2	WL HVAC WN PV DHW FN	WL HVAC WN PV DHW FN	RF DHW	WL HVAC PV DHW
3	WL RF HVAC WN PV DHW	WL RF HVAC WN PV DHW	WL DHW	RF HVAC PV DHW
4	WL HVAC WN PV DHW	WL HVAC WN PV DHW	WL RF DHW	HVAC PV DHW
5	RF HVAC WN PV DHW FN	RF HVAC WN PV DHW FN	PV DHW	WL HVAC WN PV DHW
6	HVAC WN PV DHW FN	HVAC WN PV DHW FN	RF PV DHW	WL RF HVAC WN PV DHW
7	RF HVAC WN PV DHW	RF HVAC WN PV DHW	WL PV DHW	HVAC WN PV DHW
8	HVAC WN PV DHW	HVAC WN PV DHW	WL RF PV DHW	RF HVAC WN PV DHW
9	WL RF HVAC PV DHW FN	WL RF HVAC PV DHW FN	HVAC DHW	WL HVAC PV DHW FN
10	WL HVAC PV DHW FN	WL HVAC PV DHW FN	WL HVAC DHW	WL RF HVAC PV DHW FN

Note: The details of energy upgrade rankings for all twelve scenarios in the NNS home can be found in Appendix D.

In the case of ONS, it is observed that similar EEMs are ranked high for pro-energy (S1) and pro-environmental (S4) scenarios. This is not surprising since higher-ranked combinations consist of multiple EEMs that operation either electrical or renewable energy. For instance, Table 3 shows that the preferable option for S1 and S2 is WL RF HVAC WN PV DHW. Here all the active energy systems (HVAC and DHW) are operated on electrical energy, while PV is a renewable energy generation source. All the top ten combinations in S1 and S2 have these three EEMs. However, as we move down in the ranking list, the EEMs are not the same for all cases since some EEMs focus on reducing energy (either natural or electrical) and others on reducing GHG. Since natural gas carbon footprint is almost 18 times larger than electricity, these variations become more visible for EEMs ranked lower.

Among the four scenarios displayed, EEMs ranked as pro-economic (S3) are in clear contrast to the remaining scenarios. Since cost becomes the decision criteria, energy and environmental impact fall into the backline. This ranking also indicates that cost should never be the main criterion for installing EEM. Other researchers have also shown that least-cost methods are not always environmentally efficient and can even prevent the implementation of more ambitious climate targets for these buildings in the future [87]. Since DHW has the lowest payback time, it was ranked highest for the pro-economic scenario. In reality, the stakeholders would not have such a rigid weighting but a preference that lies somewhere between the four extreme scenarios. Hence, a neutral (S4) scenario that gives equal weightage to the 3Es (energy, environment, economy) is also shown in Table 3. The ranking, however, gives an idea of the sensitivity of different upgrade combinations with changes in stakeholders' preferences.

Rankings provided by Table 4 for NNS display similar trends as ONS described above. Hence, the interrelation between energy use and GHG emissions for S1 and S2 show similar EEMs for top-ranked combinations. Cross comparison between the ONS and NNS cases also reveals that the top eight upgrade combinations are the same, with differences in ranking orders. For instance, WL RF HVAC WN PV DHW is ranked first for pro-energy and pro-environmental scenarios in the ONS home, but it is ranked third for the same scenarios in the NNS home. These results occur as top-ranking EEMS aim to reduce energy and use electricity. Similar to ONS, as we go beyond the top ten upgrade combinations, the variation in upgrade combinations increases for S1 and S2. Variation in ranking also occurs for the two occupancies. The primary source of this variation is the heating load energy.

requirements for ONS and NNS. Space heating accounts for 64 % of end use energy for residential buildings in Canada [84]. Since occupants of a building also act as internal heat sources, the heating load requirement for higher occupancy also decreases. In contrast to the S1 and S2, the rankings are the same for the S3 for both ONS and NNS occupancies. Hence, if a low payback period is stakeholders main concern, the same set of EEMs would be feasible for both occupancies (i.e., ONS and NNS). Most of the upgrade combinations for the S4 are different for both ONS and NNS cases. However, the top four combinations (WL RF HVAC PV DHW, WL HVAC PV DHW, RF HVAC PV DHW, and HVAC PV DHW) are ranked the same between the ONS and NNS homes when all criteria are given the same importance.

Overall, the ranking results of different energy upgrades indicate that the same energy upgrade combinations could not be feasible for pre-COVID and post-COVID occupancies (due to differences in their rank orders). Moreover, only a limited set of energy upgrades was explored in this research. For instance, the ranking difference for passive measure wall insulation is 10 units for ONS,increasing to 13 units for NNS. In our work, only one type of insulation was considered. However, numerous types of wall insulations are available in the market that vary in size, shape, material, thermal properties and, costs [88] and hence will result in variation of energy performance, GHG emissions and, the payback period. Other researchers, such as D'Agostino et al. (2019) [89], have considered a wider variety of EEMs for decision-making and verified that change in building technology would impact building performance. Therefore, a considerable degree of variation is anticipated when the number of energy efficiency options increase.

With the advent of the COVID-19 pandemic and the associated lockdowns, many institutions now provide flexibility to work and study at home. Therefore, both the cases of ONS and NNS will be prevalent in the future, and both cases must be considered as part of policy planning. The existing energy modeling guidelines by local and national governments recommend 50 % of occupants' time at home [45]. Therefore, these guidelines may need to be revisited at least to account for the degree of variability that can come into effect as a larger set of the population adopts NNS as the norm.

4. Conclusions, limitations and future work

This work investigated the impacts of the COVID-19 lockdown period on the energy consumption of a residential building in Okanagan Valley (British Columbia, Canada) using high-resolution energy data from 2018 to 2020. The study further investigated the impacts of occupancy times (related to pre-pandemic and post-COVID (new normal)) on EEMs and quantitatively found the energy, environmental and economic potentials. In addition, the impacts of stakeholders’ preferences were investigated for selectingEEMs suitable for each occupancy. The main findings of this research are:

• Non-HVAC loads in 2020 increased by 10 % for 2018 and 8 % for 2019. With the advent of COVID-19 lockdowns, the energy and peak times also increased. When normalized with the external dry bulb temperature, HVAC loads also increased compared to pre-COVID.

• Two extreme occupancy situations were investigated to determine the impact of pre-COVID and full occupancy (i.e., 100 % of the time spent at home). The results indicated that the electricity load increased by 16.4 %, and natural gas decreased by 7.6 %.

• It was observed that the suitability of EEMs could vary for the two occupancies, even for the same stakeholders. For example, in the case of single upgrade, wall insulation ranked 125 and 122 for pro-energy, 114 and 109 for pro-environmental, 63 and 25 for pro-economic, while 107 and 100 for neutral stakeholder choices in ONS and NNS occupancies, respectively. This implies that a different set of EEMs will be suitable for the same stakeholders if they adopt the teleworking practice for work.

The present research assumed the most extreme case for stay at home situation for NNS. People will stay 100 % of the time only for limited durations under the most stringent pandemic lockdowns, natural disasters, or a strong curfew from governments. Considering the two occupancies helped intuitionally visualize the extreme energy, economic, and environmental impacts associated with energy upgrades. Since any other occupancy would fall between these two occupancies, the effects of adopted upgrades have been effectively studied. Furthermore, some researchers have used more detailed working scenarios, where three main occupancies in an office building (0 %, 50 %, and 100 %) were investigated [90]. A similar approach can be adopted for residential buildings. More studies based on monitored data are still needed to assess variation in occupancy patterns for different households and archetypes. Since the residential sector is composed of various groups (e.g., Canadian families can be grouped into seven categories [91]), the availability of such studies will enable a cross-comparison and design of better building energy models. Future research can also compare the findings of this micro-level study with those carried out at regional levels, such as Abdeen et al. [63].

It should be kept in mind that the study is limited to energy consumption and performance of EEMs. In addition, to changes in energy use and occupancy patterns, the lockdowns also impacted thermal comfort, visual comfort, acoustical comfort, and indoor air quality (IAQ) of the home [92]. Evaluations of these parameters were beyond the scope of this work. Recent work on occupants’ comfort under the COVID-19 lockdown is still inconclusive, with some indicating decrease [92], [93] while others show an increase in comfort [94]. Therefore, more studies need to be conducted that ensure dwellings are comfortable under longer occupancy times. Evaluation of comfort levels will require collecting and analyzing data related to home characteristics and occupants. In the same vein, different restorative environmental design strategies [95] (such as spatial reconfiguration, natural and artificial lighting redesign, and ventilation systems) need to be evaluated to increase the comfort of occupants for longer occupancy times.

The findings of this study are helpful for local governments to direct resources efficiently under the changing occupancy profiles and to be ready for possible surges in residential sector energy demand under possible future lockdowns. The proposed methodology can be adopted for research other building archetypes and household sizes to help move toward more energy-efficient and sustainable residential sector. The work can also be useful for utility providers to assess energy demands and ensure the smooth working of the grid.

Declaration of Competing Interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Mohammad Kamali reports financial support was provided by Mitacs Canada. Anber Rana reports financial support was provided by Natural Sciences and Engineering Research Council of Canada. M. Mohammed Riyadh reports financial support was provided by Mitacs Canada. S. Rubaiya Sultana reports financial support was provided by Mitacs Canada. M. Rubayat Kamal reports financial support was provided by Mitacs Canada.

Appendix A.

Fig. A1

Fig. A1.

Measured vs Simulated energy results for ONS.

See Table A1, Table A2, Table A3, Table A4, Table A5

Table A1.

Characteristics of the sensors.

Parameter	Units	Sensor	Uncertainty	Reference
Temperature	[°C]	HYW-Outdoor Temperature Sensor W-C7023P2004/U	±0.2 °C	[1]
		HYW-Universal Sensor W-C7023N2001/U	±0.2 °C	[1]
Relative humidity	[%]	HYW-Room Air & Humidity Sensor W-H7635A2012/U	±3%	[2]
Electrical power	[W]	HYW-Current Sensor 50A/120 V W-CTP20050VFD001	±1.0 %	[3]

Table A2.

Information used to calibrate energy models.

No.	Source	Description
1.	Logged measured data	Lighting and general equipment electrical load data
2.	Spot measured data	Measured electrical load data
3.	Infiltration rates data	Air change rate from blower door test tests
4.	As-built documentation	Architectural and mechanical as-built drawings Wall materials and constructions taken from as-built drawings
5.	Standards and guidelines	BC Building Code 2012 HOT2000 input/output reference guidelines
6.	Stakeholder consultation	Validation of drawings and assumptions
7.	Energy bills	Monthly energy bills collected over one year time

Table A3.

Characteristics of the case study home.

Characteristics
Construction year		2016
Number of floors		2
Number of bedrooms		3
Temperature setpoints
Heating setpoint		21 °C [69.8 °F]
Cooling setpoint		22.64 °C [72.7 °F]
Hot water temperature		55.56 °C [132.0 °F]
Occupancy
	Number of occupants	2 working adults and 2 teenagers
	Occupancy time	Adults (50 %); children (50 %)
Area
	Total area of the household (m2)	291.25 [3,135 ft2]
	Floor area (m2)	249.91 [2,690 ft2]
	Wall area (m2)	364.57 [3,924 ft2]
	Opening (Door + Window) area (m2)	64.1 (17 %)
Envelope
	Foundation	5.08 cm × 15.24 cm [2″x6″] wood studs @ 60.96 cm [24″] OC, 0.95 cm [3/8″] OSB sheeting, R-20 insulation, 1.27 mm [1/2″] drywall
	Basement slab	10.16 cm [4″] concrete
	Exterior wall	5.08 cm × 15.24 cm [2″x6″] wood studs @ 60.96 cm [24″], 0.95 cm [3/8″] OSB sheeting, (USI 0.28 [RSI 3.71])
	Interior wall	5.08 cm × 10.16 cm [2″x4″] wood studs, 1.27 mm [1/2″] drywall
	Roof	Engineered trusses (wood), 1.27 mm [1/2″] OSB sheeting, Asphalt (USI 0.26 [RSI 3.85])
	Windows	Vinyl double glazed windows c/w 180 low-E
Energy systems
	Space heating and cooling	Energy star rated dual fuelPayne (PG92SCS) 17.58 kWh AFUE 92.1 % Natural Gas & Payne (PA14NC) 9.96 kWh 14 SEER Split A/C
	Domestic hot water	Standard DHW system, 227 L [50 Gallons], EF 1.901
	Appliances	Standard

Source:[30], [31].

Table A4.

Detail of Natural Gas Rates.

	Parameter	Value
Delivery charges	Basic daily charge	30 days at 0.4261 CA$/day [0.33 €/day]
	Delivery	5.024 CA$/GJ [3.9 €/day]
Commodity charges	Storage and transport	1.397 CA$/GJ [1.08 €/day]
	Cost of gas	3.844 CA$/GJ [2.98 €/day]
Other charges and taxes	Municipal operating fee	3.09 % of a amounts
	Carbon tax	2.3053 CA$/GJ [1.79 €/day]
	Clean energy levy	0.40 % of a amounts
	GST	5 % of b amounts

Source: [6].

^aDelivery charges + Commodity charges.

^bMunicipal operating fee + Carbon tax.

Table A5.

Detail of Electricity Rates.

Parameter	Value
Basic daily charge	$37.20/ 30 days [€26.53 / 30 days ]
Electricity used Block 1: 1,600 kWh	11.6 CA¢/ kWh [0.09€/ kWh]
Electricity used Block 2: greater than 1,600 kWh	14.118 CA CA¢/ kWh [0.11€/ kWh]
GST	5 % of total amounts

Source: [6].

Appendix B.

Procedure TOPSIS MCDA

The following steps are involved in performing the TOPSIS method for this particular framework.

1) Construction of the decision matrix and weighting matrices. The decision matrix considered 3 criteria: energy consumption (GJ), Operational carbon footprint (kgCO₂-e), and payback period (Years). Weighting matrix is generated by defining 12 weighting scenarios representative of various stakeholder’s preferences.

2) Normalization of the decision matrix through the conversion of different criteria into non-dimensional values using Eq. (1).

${\overline{\mathrm{z}}}_{\mathrm{i}\mathrm{j}}={\mathrm{z}}_{\mathrm{i}\mathrm{j}}/\sqrt{{\sum }_{\mathrm{i}=1}^{\mathrm{j}}{{\mathrm{z}}^2}_{\mathrm{i}\mathrm{j}}};\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{i}=\overline{1,\mathrm{n}}\mathrm{t}\mathrm{o}\mathrm{j}=\overline{1,\mathrm{m}}$(1)where, “${\mathrm{z}}_{\mathrm{i}\mathrm{j}}"$ is an entry in the decision matrix and m represents the total number of rows and n represents the total number of columns.

3) Generation of the weighted matrix through multiplication of weights of each criterion with entries in normalized matrix.

4) Determination of the best and worst alternatives for each criterion.

The best solutions (${\mathrm{V}}_{\mathrm{b}}^{+})$ for the weighted normalized matrix are selected out of all alternatives: ${\mathrm{V}}_{\mathrm{b}}^{+}=\left[{\mathrm{M}\mathrm{i}\mathrm{n}}_{\mathrm{E}},{\mathrm{M}\mathrm{i}\mathrm{n}}_{\mathrm{G}\mathrm{H}\mathrm{G}},{\mathrm{M}\mathrm{i}\mathrm{n}}_{\mathrm{P}\mathrm{B}\mathrm{T}},\right]$

The worst solutions (${\mathrm{V}}_{\mathrm{b}}^{-})$ for the weighted normalized matrix are selected out of all alternatives: ${\mathrm{V}}_{\mathrm{b}}^{-}=\left[{\mathrm{M}\mathrm{a}\mathrm{x}}_{\mathrm{E}},{\mathrm{M}\mathrm{a}\mathrm{x}}_{\mathrm{G}\mathrm{H}\mathrm{G}},{\mathrm{M}\mathrm{a}\mathrm{x}}_{\mathrm{P}\mathrm{B}\mathrm{P}}\right]$

Where: E represents the annual energy consumption; GHG is annual operation carbon footprint; and PBP is the payback period.

5) Calculation of distance from ideal solution using Euclidean distance.

The distance of alternative from the positive ideal was found by using Eq. (2-a) and for negative ideal using Eq. (2-b).

$${\mathrm{d}}_{\mathrm{i}}^{+}=\sqrt{\sum _{\mathrm{j}=1}^{\mathrm{m}}({\mathrm{v}}_{\mathrm{i}\mathrm{j}}-{\mathrm{V}}_{\mathrm{b}}^{+}})$$ (2-a)

$${\mathrm{d}}_{\mathrm{i}}^{-}=\sqrt{\sum _{\mathrm{j}=1}^{\mathrm{m}}({\mathrm{v}}_{\mathrm{i}\mathrm{j}}-{\mathrm{V}}_{\mathrm{b}}^{-}})$$ (2-b)

The relative closeness to the ideal solution was found by Eq. (2-c).

$$\mathrm{C}{\mathrm{L}}_{\mathrm{i}}^{\ast }={\mathrm{d}}_{\mathrm{i}}^{-}/{\mathrm{d}}_{\mathrm{i}}^{-}+{\mathrm{d}}_{\mathrm{i}}^{+}$$ (2-c)

6) Ranking the alternatives based on relative closeness to the ideal solution.

Appendix C.

The results of energy consumption, environmental impacts, and economic performance of different energy upgrade combinations in the old-normal-situation (ONS) and new-normal-situation (NNS) base homes are presented in Tables C1 and C2 , respectively.

Table C1.

Results of using different energy upgrade combinations in the ONS home.

Combination Code	Energy Upgrades	Energy Consumption			Environmental Impacts	Economic Performance
		Annual Electricity (kWh/Year)	Annual Natural Gas (MCF/Year)	Annual Energy Consumption (GJ/Year)	Annual Carbon Emissions (kgCO2./Year)	Relative Life Cycle Cost (dLCC30) (CA$)	Payback Period (Years)
Basecase		10381.7	75.3	116.8	4066.4	0.0	0.0
C1	WL	10508.4	71.9	113.7	3888.8	1778.7	38.5
C2	RF	10377.6	74.8	116.3	4040.0	515.6	41.7
C3	WL RF	10520.3	71.4	113.2	3862.6	2223.1	42.2
C4	HVAC	14704.7	0.9	53.9	195.6	37698.8	26.6
C5	WL HVAC	14449.8	0.9	53.0	193.0	39832.4	26.6
C6	RF HVAC	14662.9	0.9	53.7	195.2	38184.7	27.0
C7	WL RF HVAC	14410.9	0.9	52.8	192.6	40305.4	27.0
C8	WN	9928.8	69.6	109.2	3761.9	23161.2	85.3
C9	WL WN	9914.7	66.1	105.4	3577.6	25605.5	69.0
C10	RF WN	9924.2	69.1	108.6	3735.6	23679.0	83.2
C11	WL RF WN	9910.3	65.6	104.9	3551.3	26122.4	68.0
C12	HVAC WN	14071.7	0.9	51.6	189.2	59410.9	48.9
C13	WL HVAC WN	13816.1	0.9	50.7	186.6	61547.5	47.1
C14	RF HVAC WN	14036.5	0.9	51.5	188.8	59867.4	49.0
C15	WL RF HVAC WN	13774.6	0.9	50.5	186.2	62032.1	47.1
C16	PV	7884.8	75.0	107.5	4025.4	27002.1	23.0
C17	WL PV	7873.1	71.5	103.8	3841.2	29435.7	22.8
C18	RF PV	7880.4	74.4	106.9	3993.8	27558.5	23.3
C19	WL RF PV	7868.8	71.0	103.2	3814.8	29952.2	23.1
C20	HVAC PV	12023.6	0.9	44.2	168.5	65401.6	23.9
C21	WL HVAC PV	11771.1	0.9	43.3	166.0	67524.5	24.1
C22	RF HVAC PV	11982.3	0.9	44.1	168.1	65885.3	24.1
C23	WL RF HVAC PV	11731.3	0.9	43.2	165.6	68001.5	24.3
C24	WN PV	7276.6	69.3	99.3	3719.4	50853.9	38.4
C25	WL WN PV	7262.8	65.7	95.5	3529.9	53336.3	36.9
C26	RF WN PV	7272.1	68.7	98.7	3687.8	51410.7	38.4
C27	WL RF WN PV	7258.4	65.2	94.9	3503.5	53853.3	37.0
C28	HVAC WN PV	11392.0	0.9	42.0	162.2	87107.4	35.9
C29	WL HVAC WN PV	11136.7	0.9	41.0	159.6	89242.8	35.4
C30	RF HVAC WN PV	11137.1	0.9	41.0	158.8	133012.7	12.8
C31	WL RF HVAC WN PV	11095.3	0.9	40.9	159.2	89726.9	35.5
C32	DHW	9481.7	73.5	111.7	3962.6	8387.0	7.6
C33	WL DHW	9453.9	70.1	108.0	3783.5	10852.7	10.7
C34	RF DHW	9477.3	73.0	111.1	3936.3	8903.9	9.1
C35	WL RF DHW	9449.6	69.5	107.3	3751.8	11408.7	11.7
C36	HVAC DHW	13488.4	0.9	49.5	183.3	46781.9	19.7
C37	WL HVAC DHW	13232.3	0.9	48.6	180.7	48920.8	20.1
C38	RF HVAC DHW	13450.9	0.9	49.4	182.9	47248.7	20.0
C39	WL RF HVAC DHW	13192.8	0.9	48.4	180.3	49396.5	20.5
C40	WN DHW	8862.4	67.7	103.3	3651.2	32327.6	41.0
C41	WL WN DHW	8833.5	64.2	99.5	3466.8	34837.8	38.1
C42	RF WN DHW	8857.9	67.2	102.8	3624.9	32845.0	41.0
C43	WL RD WN DHW	8829.0	63.7	99.0	3440.4	35355.1	38.2
C44	HVAC WN DHW	12853.0	0.9	47.2	176.9	68504.6	36.4
C45	WL HVAC WN DHW	12596.4	0.9	46.3	174.3	70645.8	35.9
C46	RF HVAC WN DHW	12813.8	0.9	47.1	176.5	68978.9	36.6
C47	WL RF HVAC WN DHW	12553.0	0.9	46.1	173.9	71138.8	36.0
C48	PV DHW	6830.9	73.1	101.7	3914.9	36113.0	17.6
C49	WL PV DHW	6803.0	69.7	98.0	3735.7	38579.2	17.9
C50	RF PV DHW	6826.6	72.6	101.2	3888.5	36629.4	18.0
C51	WL RF PV DHW	6798.7	69.2	97.5	3709.3	39095.7	18.2
C52	HVAC PV DHW	10810.9	0.9	39.9	156.3	74468.7	20.5
C53	WL HVAC PV DHW	10554.3	0.9	38.9	153.7	76609.8	20.8
C54	RF HVAC PV DHW	10767.8	0.9	39.7	155.9	74960.4	20.7
C55	WL RF HVAC PV DHW	10514.8	0.9	38.8	153.3	77085.5	21.0
C56	WN PV DHW	6211.1	67.4	93.5	3608.7	60016.3	30.4
C57	WL WN PV DHW	6182.2	63.9	89.7	3424.3	62526.5	29.7
C58	RF WN PV DHW	6206.6	66.8	92.8	3577.1	60573.2	30.4
C59	WL RF WN PV DHW	6177.8	63.3	89.0	3392.7	63082.9	29.7
C60	HVAC WN PV DHW	10173.9	0.9	37.6	149.9	96198.5	30.6
C61	WL HVAC WN PV DHW	9915.7	0.9	36.6	147.3	98346.8	30.5
C62	RF HVAC WN PV DHW	10132.5	0.9	37.4	149.5	96682.6	30.7
C63	WL RF HVAC WN PV DHW	9875.1	0.9	36.5	146.9	98827.4	30.6
C64	FN	10168.2	73.5	114.1	3969.5	11660.0	180.7
C65	WL FN	10170.1	69.6	110.0	3764.4	14191.1	103.4
C66	RF FN	10164.2	72.5	113.1	3916.9	12372.5	149.1
C67	WL RF FN	10166.0	69.1	109.5	3738.0	14706.7	99.5
C68	HVAC FN	14507.5	0.9	53.2	193.6	48575.8	51.7
C69	WL HVAC FN	14251.7	0.9	52.3	191.0	50713.4	49.3
C70	RF HVAC FN	14469.8	0.9	53.0	193.2	49043.4	51.7
C71	WL RF HVAC FN	14213.5	0.9	52.1	190.6	51183.2	49.4
C72	WN FN	9589.9	67.4	105.6	3642.8	35536.7	100.0
C73	WL WN FN	9673.5	63.9	102.2	3459.5	37546.6	90.7
C74	RF WN FN	9585.5	66.9	105.1	3616.4	36053.7	98.1
C75	WL RF WN FN	9668.9	63.3	101.6	3427.9	38103.9	88.9
C76	HVAC WN FN	13878.3	0.9	50.9	187.2	70270.9	67.6
C77	WL HVAC WN FN	13615.6	0.9	50.0	184.6	72439.2	64.3
C78	RF HVAC WN FN	13836.3	0.9	50.8	186.8	70757.7	67.3
C79	WL RF HVAC WN FN	13577.6	0.9	49.8	184.2	72908.2	64.2
C80	PV FN	7513.5	72.7	103.7	3900.7	39561.2	40.5
C81	WL PV FN	7515.5	69.3	100.2	3721.8	41894.5	38.9
C82	RF PV FN	7509.4	72.2	103.2	3874.3	40076.8	40.5
C83	WL RF PV FN	7510.7	68.7	99.5	3690.2	42452.6	38.8
C84	HVAC PV FN	11824.3	0.9	43.5	166.5	76287.9	36.1
C85	WL HVAC PV FN	11567.5	0.9	42.6	164.0	78430.0	35.6
C86	RF HVAC PV FN	11782.8	0.9	43.4	166.1	76772.5	36.2
C87	WL RF HVAC PV FN	11529.3	0.9	42.5	163.6	78899.8	35.7
C88	WN PV FN	7048.3	67.0	96.1	3596.1	62777.1	50.9
C89	WL WN PV FN	7018.5	63.5	92.3	3411.7	65291.2	48.5
C90	RF WN PV FN	7043.7	66.5	95.5	3569.8	63294.9	50.8
C91	WL RF WN PV FN	7014.0	63.0	91.7	3385.3	65808.6	48.4
C92	HVAC WN PV FN	11189.1	0.9	41.2	160.1	98009.8	46.2
C93	WL HVAC WN PV FN	10929.4	0.9	40.3	157.5	100164.7	45.4
C94	RF HVAC WN PV FN	11153.3	0.9	41.1	159.8	98469.0	46.3
C95	WL RF HVAC WN PV FN	10888.9	0.9	40.2	157.1	100644.8	45.4
C96	DHW FN	9072.2	74.9	111.7	4032.2	19655.5	56.2
C97	WL DHW FN	9073.6	71.5	108.1	3853.3	21991.4	50.0
C98	RF DHW FN	9067.9	74.4	111.1	4005.8	20172.0	55.7
C99	WL RF DHW FN	9069.5	70.9	107.5	3821.7	22546.4	49.5
C100	HVAC DHW FN	13543.5	0.9	49.7	183.9	56536.8	39.6
C101	WL HVAC DHW FN	13291.0	0.9	48.8	181.3	58659.8	38.7
C102	RF HVAC DHW FN	13506.6	0.9	49.6	183.5	57000.9	39.7
C103	WL RF HVAC DHW FN	13253.6	0.9	48.7	180.9	59126.1	38.9
C104	WN DHW FN	8494.5	69.3	103.7	3731.7	43332.2	65.0
C105	WL WN DHW FN	8466.2	65.8	99.9	3547.3	45839.6	59.5
C106	RF WN DHW FN	8490.1	68.7	103.0	3700.1	43888.6	64.3
C107	WL RF WN DHW FN	8461.8	65.2	99.2	3515.7	46396.0	59.0
C108	HVAC WN DHW FN	12925.3	0.9	47.5	177.6	78183.1	53.7
C109	WL HVAC WN DHW FN	12663.6	0.9	46.5	175.0	80346.9	52.0
C110	RF HVAC WN DHW FN	12885.3	0.9	47.3	177.2	78661.0	53.7
C111	WL RF HVAC WN DHW FN	12620.7	0.9	46.4	174.6	80837.7	52.0
C112	PV DHW FN	6416.6	74.5	101.7	3984.3	47402.8	33.1
C113	WL PV DHW FN	6418.2	71.1	98.1	3805.5	49737.8	32.4
C114	RF PV DHW FN	6412.3	74.0	101.2	3958.0	47919.3	33.2
C115	WL RF PV DHW FN	6414.0	70.5	97.5	3773.9	50293.3	32.4
C116	HVAC PV DHW FN	10859.6	0.9	40.0	156.8	84252.1	31.7
C117	WL HVAC PV DHW FN	10605.1	0.9	39.1	154.3	86383.9	31.5
C118	RF HVAC PV DHW FN	10819.8	0.9	39.9	156.4	84729.1	31.8
C119	WL RF HVAC PV DHW FN	10566.3	0.9	39.0	153.9	86856.5	31.6
C120	WN PV DHW FN	5838.8	68.9	93.7	3683.9	71079.9	42.1
C121	WL WN PV DHW FN	5810.4	65.4	89.9	3499.4	73587.8	40.8
C122	RF WN PV DHW FN	5834.4	68.3	93.1	3652.3	71636.3	42.1
C123	WL RF WN PV DHW FN	5918.0	64.8	89.7	3469.0	73646.1	41.5
C124	HVAC WN PV DHW FN	10234.1	0.9	37.8	150.5	105930.8	40.9
C125	WL HVAC WN PV DHW FN	9974.1	0.9	36.9	147.9	108087.1	40.3
C126	RF HVAC WN PV DHW FN	10192.4	0.9	37.6	150.1	106416.3	40.9
C127	WL RF HVAC WN PV DHW FN	9931.5	0.9	36.7	147.5	108576.5	40.4

Table C2.

Results of using different energy upgrade combinations in the NNS home.

Combination Code	Energy Upgrades	Energy Consumption			Environmental Impacts	Economic Performance
		Annual Electricity (kWh/Year)	Annual Natural Gas (MCF/Year)	Annual Energy Consumption (GJ/Year)	Annual Carbon Emissions (kgCO2./Year)	Relative Life Cycle Cost (dLCC30) (CA$)	Payback Period (Years)
Basecase		14427.9	64.6	120.1	3544.2	0	0.0
C1	WL	14458.7	61.2	116.6	3365.650	2205.2	24.9
C2	RF	14,424	64.1	119.6	3517.878	514.7	41.9
C3	WL RF	14,455	60.7	116.1	3339.306	2719	27.5
C4	HVAC	17,671	0.9	64.6	225.475	38277.5	25.2
C5	WL HVAC	17,433	0.9	63.7	223.076	40,336	25.5
C6	RF HVAC	17636.5	0.9	64.4	225.127	38,731	25.7
C7	WL RF HVAC	17395.6	0.9	63.6	222.699	40802.3	25.9
C8	WN	13798.4	58.7	111.6	3227.462	24025.5	70.9
C9	WL WN	13791.5	55.2	107.9	3043.248	26437.8	60.2
C10	RF WN	13813.5	58.1	111.0	3196.047	24495.2	70.4
C11	WL RF WN	13787.5	54.7	107.3	3016.901	26952.9	59.6
C12	HVAC WN	16991.1	0.9	62.1	218.622	60198.2	46.1
C13	WL HVAC WN	16737.4	0.9	61.2	216.065	62326.4	44.6
C14	RF HVAC WN	16955.2	0.9	62.0	218.260	60657.8	46.2
C15	WL RF HVAC WN	16704.1	0.9	61.1	215.729	62774.5	44.8
C16	PV	11795.5	64.3	110.3	3501.905	27604.7	21.8
C17	WL PV	11808.2	60.9	106.8	3323.149	29890.3	22.0
C18	RF PV	11791.5	63.8	109.8	3475.558	28119.8	22.2
C19	WL RF PV	11804.5	60.4	106.2	3296.806	30404.1	22.3
C20	HVAC PV	15001.1	0.9	55.0	198.563	65930.5	23.4
C21	WL HVAC PV	14758.3	0.9	54.1	196.115	68010.3	23.6
C22	RF HVAC PV	14958.7	0.9	54.8	198.135	66419.1	23.6
C23	WL RF HVAC PV	14,722	0.9	53.9	195.749	68471.7	23.8
C24	WN PV	11166.8	58.3	101.7	3179.890	51666.1	36.6
C25	WL WN PV	11140.7	54.9	98.0	3000.744	54124.3	35.3
C26	RF WN PV	11162.7	57.8	101.2	3153.543	52181.7	36.7
C27	WL RF WN PV	11136.7	54.4	97.5	2974.397	54639.4	35.4
C28	HVAC WN PV	14316.9	0.9	52.5	191.666	87870.3	34.8
C29	WL HVAC WN PV	14063.7	0.9	51.6	189.114	89996.3	34.4
C30	RF HVAC WN PV	14276.2	0.9	52.3	191.256	88351.3	34.9
C31	WL RF HVAC WN PV	14032.6	0.9	51.5	188.800	90434.6	34.6
C32	DHW	13324.2	66.2	117.8	3617.279	7950.7	8.4
C33	WL DHW	13,334	62.9	114.4	3443.755	10209.8	11.9
C34	RF DHW	13320.3	65.7	117.3	3590.933	8465.5	10.0
C35	WL RF DHW	13330.1	62.3	113.7	3412.148	10,764	13.0
C36	HVAC DHW	16688.1	0.9	61.0	215.568	46322.7	20.3
C37	WL HVAC DHW	16449.3	0.9	60.2	213.161	48384.6	20.8
C38	RF HVAC DHW	16686.5	0.9	61.0	215.551	46629.8	20.9
C39	WL RF HVAC DHW	16448.9	0.9	60.2	213.156	48686.4	21.4
C40	WN DHW	13798.4	58.6	111.5	3222.201	27,739	77.4
C41	WL WN DHW	13791.5	55.2	107.9	3043.248	30111.8	65.7
C42	RF WN DHW	13813.5	58.1	111.0	3196.047	28169.2	77.2
C43	WL RD WN DHW	13787.5	54.7	107.3	3016.901	30626.9	65.0
C44	HVAC WN DHW	16991.1	0.9	62.1	218.622	63872.2	48.7
C45	WL HVAC WN DHW	16737.4	0.9	61.2	216.065	66000.4	47.0
C46	RF HVAC WN DHW	16955.2	0.9	62.0	218.260	64331.8	48.8
C47	WL RF HVAC WN DHW	16704.1	0.9	61.1	215.729	66448.5	47.2
C48	PV DHW	10673.5	65.9	108.0	3574.776	35636.8	18.2
C49	WL PV DHW	10683.2	62.5	104.4	3395.990	37935.8	18.5
C50	RF PV DHW	10669.5	65.4	107.4	3548.429	36,152	18.5
C51	WL RF PV DHW	10679.3	62	103.9	3369.644	38450.5	18.8
C52	HVAC PV DHW	14013.1	0.9	51.4	188.604	73998.3	20.9
C53	WL HVAC PV DHW	13773.4	0.9	50.5	186.187	76064.3	21.2
C54	RF HVAC PV DHW	13976.6	0.9	51.3	188.236	74460.7	21.1
C55	WL RF HVAC PV DHW	13735.9	0.9	50.4	185.809	76531.1	21.4
C56	WN PV DHW	10038.3	60	99.4	3257.957	59687.6	30.8
C57	WL WN PV DHW	10036.2	56.6	95.8	3079.052	62039.1	30.3
C58	RF WN PV DHW	10034.1	59.5	98.9	3231.608	60203.7	30.9
C59	WL RF WN PV DHW	10032.1	56	95.2	3047.443	62594.2	30.4
C60	HVAC WN PV DHW	13335.8	0.9	49.0	181.776	95907.4	30.9
C61	WL HVAC WN PV DHW	13081.7	0.9	48.0	179.215	98037.4	30.8
C62	RF HVAC WN PV DHW	13297.8	0.9	48.8	181.393	96376.4	31.1
C63	WL RF HVAC WN PV DHW	13045.9	0.9	47.9	178.854	98496.6	30.9
C64	FN	14,205	62.7	117.3	3442.012	11741.3	172.3
C65	WL FN	14178.6	59.4	113.7	3268.124	14161.3	104.4
C66	RF FN	14200.2	62.2	116.7	3415.657	12,260	157.7
C67	WL RF FN	14174.8	58.8	113.1	3236.518	14715.1	99.3
C68	HVAC FN	17458.6	0.9	63.8	223.334	49222.1	48.9
C69	WL HVAC FN	17243.3	0.9	63.0	221.164	51179.6	47.5
C70	RF HVAC FN	17207.1	0.9	62.9	220.799	50640.6	45.3
C71	WL RF HVAC FN	17207.1	0.9	62.9	220.799	51640.6	47.6
C72	WN FN	13531.9	56.8	108.6	3124.811	35960.7	94.0
C73	WL WN FN	13512.6	53.8	105.4	2966.778	38230.7	83.3
C74	RF WN FN	13532.5	56.3	108.1	3098.511	36455.4	92.7
C75	WL RF WN FN	13508.4	53.3	104.9	2940.430	38746.8	82.2
C76	HVAC WN FN	16910.3	0.9	61.8	217.807	70557.5	66.2
C77	WL HVAC WN FN	16,662	0.9	60.9	215.305	72661.7	63.4
C78	RF HVAC WN FN	16872.7	0.9	61.7	217.428	71024.7	66.1
C79	WL RF HVAC WN FN	16624.3	0.9	60.8	214.925	73129.4	63.3
C80	PV FN	11,551	62.4	107.4	3399.476	39,442	40.8
C81	WL PV FN	11525.4	59	103.7	3220.334	41,898	38.9
C82	RF PV FN	11,547	61.7	106.7	3362.607	40036.1	40.6
C83	WL RF PV FN	11,521	58.5	103.2	3193.983	42414.9	38.9
C84	HVAC PV FN	14808.6	0.9	54.3	196.622	76786.6	35.3
C85	WL HVAC PV FN	14566.6	0.9	53.4	194.183	78862.8	35.0
C86	RF HVAC PV FN	14768.6	0.9	54.1	196.219	77264.5	35.4
C87	WL RF HVAC PV FN	14528.8	0.9	53.3	193.802	79330.9	35.1
C88	WN PV FN	10878.3	56.5	98.8	3082.279	63659.6	48.6
C89	WL WN PV FN	10855.8	53	95.0	2897.908	66141.3	46.5
C90	RF WN PV FN	10878.9	55.9	98.1	3050.718	64193.8	48.4
C91	WL RF WN PV FN	10851.6	52.5	94.5	2871.559	66657.3	46.5
C92	HVAC WN PV FN	14233.6	0.9	52.2	190.826	98240.7	45.8
C93	WL HVAC WN PV FN	13983.7	0.9	51.3	188.307	100352.1	45.1
C94	RF HVAC WN PV FN	14193.5	0.9	52.0	190.422	98719.1	45.9
C95	WL RF HVAC WN PV FN	13943.9	0.9	51.1	187.906	100829.1	45.1
C96	DHW FN	13092.2	64.4	115.1	3520.237	19,693	55.8
C97	WL DHW FN	13,079	61	111.4	3341.220	22093.9	49.3
C98	RF DHW FN	13087.5	63.9	114.5	3493.883	20211.3	55.3
C99	WL RF DHW FN	13075.1	60.5	110.9	3314.874	22608.6	49.1
C100	HVAC DHW FN	16510.1	0.9	60.4	213.773	57114.2	38.1
C101	WL HVAC DHW FN	16271.2	0.9	59.5	211.365	59176.7	37.5
C102	RF HVAC DHW FN	16476.3	0.9	60.3	213.433	57564.6	38.4
C103	WL RF HVAC DHW FN	16231.5	0.9	59.4	210.965	59653.2	37.7
C104	WN DHW FN	12433.2	58.5	106.5	3203.178	43849.7	62.0
C105	WL WN DHW FN	12413.5	55.5	103.2	3045.141	46121.5	58.1
C106	RF WN DHW FN	12433.7	58	106.0	3176.877	44344.9	61.8
C107	WL RF WN DHW FN	12409.2	55	102.7	3018.791	46,638	57.9
C108	HVAC WN DHW FN	15942.6	0.9	58.3	208.053	78,535	52.7
C109	WL HVAC WN DHW FN	15695.2	0.9	57.5	205.559	80635.2	51.2
C110	RF HVAC WN DHW FN	15903.2	0.9	58.2	207.656	79010.2	52.7
C111	WL RF HVAC WN DHW FN	15653.8	0.9	57.3	205.142	81119.4	51.2
C112	PV DHW FN	10451.3	64	105.2	3472.572	47,375	33.1
C113	WL PV DHW FN	10424.7	60.7	101.6	3298.681	49795.9	32.3
C114	RF PV DHW FN	10446.5	63.5	104.6	3446.217	47893.7	33.2
C115	WL RF PV DHW FN	10420.8	60.2	101.0	3272.335	50310.6	32.4
C116	HVAC PV DHW FN	13829.2	0.9	50.7	186.750	84816.2	31.1
C117	WL HVAC PV DHW FN	13589.4	0.9	49.9	184.333	86882.6	31.0
C118	RF HVAC PV DHW FN	13795.1	0.9	50.6	186.406	85267.8	31.2
C119	WL RF HVAC PV DHW FN	13551.3	0.9	49.7	183.949	87,352	31.1
C120	WN PV DHW FN	9778.7	58.2	96.6	3160.637	71552.6	41.3
C121	WL WN PV DHW FN	9755.6	54.8	92.9	2981.521	73997.5	40.1
C122	RF WN PV DHW FN	9779.2	57.6	96.0	3129.074	72087.3	41.3
C123	WL RF WN PV DHW FN	9751.3	54.2	92.3	2949.910	74553.5	40.0
C124	HVAC WN PV DHW FN	13259.1	0.9	48.7	181.003	106248.5	40.5
C125	WL HVAC WN PV DHW FN	13012.8	0.9	47.8	178.521	108343.8	40.0
C126	RF HVAC WN PV DHW FN	13221.4	0.9	48.5	180.623	106716.2	40.5
C127	WL RF HVAC WN PV DHW FN	12,973	0.9	47.7	178.119	108820.8	40.1

Appendix D. TOPSIS results

The TOPSIS MCDA ranking results of energy upgrade combinations in the old-normal-situation (ONS) and new-normal-situation (NNS) base homes uning are presented in Table D1 .

Table D1.

Results of TOPSIS ranking for energy upgrades combinations ONS (Left) and NNS (Right).

Data availability

Data will be made available on request.