Impact of energy conservation measures in residential buildings in very remote communities in Australia

Abstract

This study investigated the impact of various energy conservation measures (ECMs) on total annual electricity consumption on a single level, two-bedroom residential building in an arid climate and in a very remote area of Australia. Base case scenario of annual energy consumption profiles of the building and its existing systems were modelled and simulated using DesignBuilder software, and the results were validated by on-site measured electricity data. Two categories of ECMs (major intervention and minor intervention) were investigated and analysed. The findings show that the total annual energy consumption of the case study building can be reduced by up to 44% (2 kWh/m²·yr) when compared against the base model and measured data if selected ECMs are implemented. Significant savings from implementing selected ECMs can help alleviate the cost-of-living pressure and stress currently experienced by many households on low incomes especially for residents of very remote communities in Australia.

KEYWORDS:

• Energy conservation measures
• thermal comfort
• building energy simulations
• very remote communities
• electricity consumption
• energy conservation
• arid climates

1. Introduction

Energy consumption in buildings is subject to several factors such as climatic conditions, building design and orientation, building envelope, occupant behaviour, building systems and controls, maintenance, etc. (Qian et al. 2019). Between 1994 and 2014, worldwide energy consumption increased by about 92% with current research indicating an increase in global demand for oil, coal and natural gas by 30% and 50%, respectively, between 2007 and 2035 (Berardi 2015). This implies an annual increase of carbon emissions by more than 2% (Qian et al. 2019). In Australia, the energy consumption rose by 0.6% to reach 6196 petajoules in 2019, representing an average annual growth of 0.7% over the past 10 years. Heavily dependent on fossil fuels (Coal, Oil and Gas), these accounted for 94% of Australia’s primary energy mix, where oil took the largest share at 39%, coal at 29% and natural gas at 29%, while renewable energy accounted for the remaining 6% (Department of the Environment and Energy 2019). In 2017, Australia’s energy supply was more carbon intensive (‘dirtier’) than China’s, and the rest of the world (WED 2020a; 2020b; 2020c).

The residential sector in Australia is a big consumer of energy and a significant carbon emitter recording 57% of emissions in the sector (Energy Consult Pty Ltd 2015). It remains the fourth largest consumer of energy after transport, electricity supply and manufacturing (Department of Industry 2020). Typical building energy consumption goes towards space heating, space cooling, lighting, water heating and ventilation (McGee 2017). Based on the above, it is evident that reducing energy consumption in the residential sector is an effective and practical way to reduce energy consumption and carbon emissions domestically and globally.

The Nationwide Housing Energy Rating Scheme (NatHERS) in Australia is a direct response to the need for an energy-efficient residential sector and specifies minimum energy requirements for residential buildings in Australia. While this initiative has greatly benefited the housing sector in Australia, the impact has been minimal in remote communities due to the peculiar characteristics associated with remote communities.

Spanning 90% of Australia’s land mass and between 6% and 23% of the entire population, the characteristics of remote Australia differ from other parts of Australia. These remote regions present with common systemic properties including a history of persistent social and economic disadvantage that also manifest in poor health statistics and weak infrastructure. These systemic remote region characteristics when interfaced with changing climates and different energy futures, results in impacts, opportunities and vulnerabilities that differ from those applicable to coastal and urban regions of Australia (Maru, Chewings, and Sparrow 2012). The challenges buildings and their occupants in remote regions of Australia face when compared with urban areas include generally harsher climates with greater temperature extremes and much higher electricity costs ($5/kWh versus $0.2/kWh). This is due to the higher cost of electricity infrastructure design, installation, operation and maintenance in remote areas compared to metropolitan areas, as such cost of electricity supply to the consumer is therefore much higher in remote areas; ‘for example, $2 per kilowatt hour (for the full life-cycle cost of electricity generation), compared to about 14 cents per kilowatt hour in town’ (CAT 2021).

Approximately 100,000 Aboriginal and Torres Strait Islander people live in remote communities (Buergelt et al. 2017) and it is projected to reach 145,000 by 2021 (ABS 2017). Extreme climatic conditions and overcrowding, combined with significantly lower incomes and significantly greater health problems means that their energy needs take on particular importance. Therefore, the supply of affordable energy to these communities is critical to their daily survival and their longer-term opportunities, including improving their life expectancy and overall wellbeing (Energy Consumers Australia 2019). It is estimated that an additional 5500 homes will be required to address overcrowding in remote communities by 2028 after accounting for population growth (Towart et al. 2017). This expected increase in the supply of housing also means increased energy demand, consumption and emissions. In 2012 alone, total power consumption in remote communities was 5739 kWh (Foran et al. 2015) and given the significant investment in housing targeted towards remote communities by the Australian government, opportunities continue to abound to reduce energy demand, consumption and carbon emissions via the identification and implementation of energy conservation measures. Building Energy Simulation (BES) is a popular approach towards achieving this outcome as it facilitates the exploration of climate-responsive design and the testing of alternatives for building energy efficiency and thermal performance without compromising occupant comfort. However, most of such studies in Australia and beyond have been representative of urban and coastal areas and do not capture the systemic properties associated with remote and very remote regions which are largely Indigenous. Consequently, baseline data from the findings may not necessarily produce accurate results when directly applied in the remote context/settings. While the number of projects looking into the issue of energy in Indigenous communities continues to grow, it remains a very new field of research and focus in Australia. This paper attempts to augment this growing field of study using an evidence-based approach.

This research, therefore, presents a pilot analysis using a case study approach to explore the effectiveness of several energy conservation measures aimed at improving the energy performance of residential buildings in very remote communities without compromising thermal performance to produce the following outcomes:

1. Generate data that will allow sustainability gaps in existing remote housing to be identified so that appropriate solutions can be proffered.

2. Generate data that will aid evaluation and prioritization when implementing strategies for achieving energy efficiency.

3. Generate results that will form a basis for management and decision making.

4. Provide workable and reliable data that will inform the planning, design, construction, operation and maintenance of sustainable housing in remote communities.

5. Establish key performance indicators for future use, potential benchmarking and measuring improvements.

6. Identify the utility of environmental and IEQ data as inputs into evidence-based design via building performance quantification techniques (monitoring and simulations) for generating new housing concept designs for different climatic zones in remote communities.

2. Literature review

2.1. Overview

Electricity is an essential service that building occupants rely on to maintain good health and comfort. According to the world bank overview on energy, about 789 million people live without electricity with hundreds of millions more living with insufficient or unreliable access to it (Bank 2021).

Since 2007, Australian residential retail electricity prices have increased faster than those of any other OECD country and in the last decade, Australia has moved from being a low-price electricity market to around the OECD average (Queensland Productivity Commission 2016). Increasing tariffs and growing consumption have seen the cost of household electricity in Australia continue to rise resulting in undue pressure on household budgets and further stresses for those on low incomes across the country (McKenzie 2013). In Queensland (QLD), a total of 29,692 households in the state were disconnected for not paying an electricity bill, and as of June 2015, it was reported that over 15,000 households were managing their bills under a hardship programme arrangement with their retailer (Bright Actions 2016). In Victoria (VIC), the proportion of households participating in retailer hardship programmes grew by 35% (Strempel, Nicholls, and Strengers 2015). Data across the different states and territories highlight similar trends with Western Australia (WA) recording the highest number of disconnection rates in comparison to states like New South Wales (NSW), Victoria (VIC) and South Australia (SA) (Economic Regulation Authority 2019).

For remote and very remote communities in Australia, these stresses take a far greater dimension due to the peculiar characteristics associated with the region. It is well documented that residents of remote communities currently experience disadvantage across a range of social and economic indicators including health, housing, income and education which makes such households vulnerable to relative poverty. Median household incomes are significantly less than their city counterparts thus, are at particular risk of the negative health and social effects of high-energy costs and poor reliability of supply caused by frequent disconnections (Queensland Council of Social Service 2014). While low household income does not directly cause an increase in high-energy consumption, it is a factor which directly impacts one’s ability to cope with energy cost. This is because most remote and very remote communities in Australia are not serviced by an electricity grid but employ the extensive use of diesel as a means of power generation through diesel generators. These communities are isolated by hundreds of kilometres of untarred (dirt road) which can be cut off by flooding especially during the rainy seasons, which makes reliable supply of diesel difficult around certain times of the year (Green, Jackson, and Morrison 2009). The majority of Indigenous people residing in remote and very remote areas are in receipt of Commonwealth income support payments hence, are considered to be low income since individuals in these communities live on an average median income between $209 and $237 per week (Statistics 2006). The median personal income in 2018–2019 was around $489 per week for Aboriginal and Torres Strait Islander adults and $767 per week for non-Indigenous adults (Provision 2020). This disparity in the median equivalized gross weekly household (EGWH) income between residents of remote communities and their city counterparts generally decreases as remoteness increases. For instance, in 2018–2019, the median EGWH income ranged from $719 in major cities to $400 in very remote areas (see Figure 1 for remote classification according to the Australian Statistical Geography Standard (ASGS)). The lack of employment opportunities also means that the individuals in receipt of Commonwealth income support payments typically tend to spend more time at home during the daytime thus, driving up relative energy costs (Urmee, Thoo, and Killick 2012). Going by the statistics above, it is not surprising that households on low and fixed incomes spend a greater proportion of their income on such an essential service when compared with households with higher income (Fagan and Australia 2008). Statistically, low-income households spend approximately 10% of their disposable income on household energy costs, compared to the national average of 5%, which means that even the slightest increase in utility bills can have a significant impact on already stretched household budgets (GHW 2016).

Figure 1. 2016 Australian Statistical Geography Standard: Remoteness Structure Remoteness Area Boundaries.

The growing concern about how rising costs and consumption of electricity impacts households in remote and very remote communities in Australia, especially those on low income, disadvantaged or otherwise marginalized, means it remains critical that more effort is channelled towards alleviating the pressures faced by such communities in accessing this service. Wanarn Community, WA, the case study for this paper rates as very remote which implies that residents of this community experience great difficulty in accessing the services they need and the products they wish to purchase (Trewin 2001). Hence, the need for a targeted approach to reduce energy need, demand and consumption through the design and delivery of energy-efficient housing that not only minimizes ongoing operational and maintenance cost but also provides greater opportunities for the occupants to adopt energy-efficient practices through behavioural changes.

2.2. Cost of remoteness

The disadvantage associated with remote communities when it comes to the cost-of-service delivery due to remoteness is visible in various sectors such as health (Welfare 2011), housing (Fien et al. 2007), education (Stokes, Holdsworth, and Stafford 1999), transportation (Spandonide 2014) and energy (AECOM 2014; Nous Group 2017). Remoteness significantly impacts the cost of building materials and construction due to high fuel and transportation costs (Szava et al. 2007). The effect of this is further pronounced in maintenance costs which are around 8.5 times higher in very remote communities than in remote communities, with emergency repair activities costing around 75% higher, and planned repairs costing about 50% higher than in remote communities due to travel cost per job (Nous Group 2017). For remote communities, vast distances often correlate with decrease in population (Figure 2). The more remote a community is, the more reduced the population will be as such, some communities could have a population of less than 50 people resulting in a higher cost per unit per capita towards service delivery (Australian Government Productivity Commission 2017). In terms of energy service provision, the cost of setting up the necessary infrastructure including maintaining poles and wires and the management of power distribution alone remains significant and ongoing. For instance, the price of diesel fuel to run power generators is one factor particularly in the NT and in northern WA where, in some Aboriginal communities, those costs can account for as much as $2000 per quarter, translating to $0.25–$2.00/kWh (Ding et al. 2015)

Figure 2. Discrete Indigenous communities by size and remoteness, 2011. Source: ABS (Census of Population and Housing, unpublished).

From a cultural point of view, connections to family and community, to the land, to culture and traditions, are all fundamental to how indigenous people in remote communities feel about themselves, and their sense of a good life (Australian Government Productivity Commission 2017). Certain cultural practices when overlooked, create significant spikes in energy costs for Aboriginal and Torres Strait Islander households for instance, it is not unusual for Aboriginal and Torres Strait Islander households to have large numbers of people living in a single house at any one time. The reason being that many households have large families, as well as cultural expectations that Aboriginal and Torres Strait Islander communities generally have an ‘open door’ to relatives and a cultural responsibility to take in relatives in times of need. This also includes distant relatives (Queensland Council of Social Service 2014) as such these socio-cultural characteristics contribute to high electricity consumption and costs.

To address this problem, the federal government as part of their climate change strategy introduced the Low-Income Energy Efficiency Program (LIEEP) in 2011(Yang et al. 2020b). Aimed at improving the energy efficiency of low-income households and facilitate better management of energy use by these households, the LIEEP programme had two main objectives:

1. To assist low-income and disadvantaged households overcome identified barriers to energy efficiency to better manage their energy use and,

2. To capture and analyse data and information to inform future energy efficiency policy and programmes.

The federal government awarded grants worth $55.3 million to 20 recipients where 20 energy efficiency projects were trialled out across Australia (Department of Industry 2016). About 20,000 households were engaged across the 20 programmes with rather diverse participant characteristics. The key approaches to data collection included participant engagement through online surveys, social media tools and telephone interviews (Australian Government – Department of Industry 2016; Skillset Environment 2016; Sustainable Business Australia 2016; Swinton et al. 2016), energy efficiency workshops (BEST Employment Ltd 2016; Button, Lodge, and Taylor 2016; Cooper et al. 2016; Romanach et al. 2016; Yang et al. 2020a) and personalized home visits and energy assessments (Alexander et al. 2016; Alliance 2016; Bedggood et al. 2016; Buergelt et al. 2016; Byrne et al. 2016; Lynch et al. 2016; Northern Grampians Shire Council 2016; O’Mullane and Hoch 2016; Skillset Environment 2016; Steinborner et al. 2016; Sullivan 2016; Uniting Communities 2016; Ward and Brent 2016). Table 1 provides a snapshot on the programmes highlighting key aspects of each programme such as type of programme, location, approach, type of interventions implemented, outcomes as well as the challenges faced in carryout carrying out the various projects.

A summary of the findings from the LIEEP programme is provided below:

Table 1. Overview of LIEEP programmes and key elements.

No.	Project	Location	Remote Classification	No. of Households	Approach	Electricity data source	Interventions	Outcome	Challenges
1	Power Save	NSW	Regional	900	Engagement and workshops, Home Energy Assessment	Energy Bills	Energy Saving devices & appliance upgrade	10% reduction in electricity consumption per household	Lack of available energy consumption data
2	GHW Program	Queensland	Major City – Brisbane	1000	Home Energy Assessment and Energy Workshops	Energy provider	Energy Saving devices & appliance upgrade	Reduction in electricity cost across all households by $35,184.52 per year	Lack of available energy consumption data, difficult to implement a one size fits all approach
3	Home Energy Efficiency Upgrade Program (HEEUP)	Victoria	Major City & Regional	793	Home visits	Metre readings	Hot water system upgrades	20% decrease daily in electricity consumption	Reluctance to participate due to the need to provide personal data
4	Switched on homes	Western Australia	Regional	240	Online surveys and telephone interviews	Data Loggers, energy bills & energy provider	Installation of Voltage optimization units, generic energy efficiency tips sent by SMS, personalized feedback sent by SMS	Voltage optimization – 8% reduction in electricity use, SMS tips – 5.8% reduction in electricity use. Personalized feedback – no impact	Lack of available historical data on building records
5	Reduce Your Juice (RYJ)	Queensland	Major City – Brisbane	1000	Using social media to engage participants to foster behavioural change	Self-reported billing data & data from energy provider	Education and energy efficient equipment upgrades	$54.82 average saving on quarterly electricity bills ($219.28 annual saving)	Low interest/low involvement topic, Information failure
6	Smart Cooling in the Tropics (SCIT)	Northern Territory	Major City – Darwin	480	Home energy assessments	Self-reported billing data & data from energy provider	Structural modifications, appliance upgrades and temperature setpoint adjustments	Structural modifications: 30% reduction in energy consumption & Appliance upgrades: 14% reduction in energy consumption	Lack of available historical data on building energy consumption records. Incomplete or inaccurate data
7	Future Powered Families (FPF)	Victoria	Major City & Regional	6133 participants	Home Energy Assessment and Energy Workshops	Energy provider	Energy Saving devices & appliance upgrade	10.78% reduction in energy consumption	Difficulties with recruiting participants, low rates of data collection, insufficient energy consumption data
8	Power down project	Victoria	Major City & Regional	1350	Home Energy Assessment and Energy Workshops	Metre readings & energy provider	Structural modifications, energy Saving devices & appliance upgrade	Reduction in annual energy consumption by 222.65 kWh and $62.35 annual savings at 17 cents per kWh per household	Missing/ incomplete data, data collection
9	Manymak Energy Efficiency Project	Northern Territory	Remote	663	Home visits and Energy monitoring using data loggers	Metre readings & data loggers	Structural modifications, energy Saving devices & appliance upgrade	8.52 kWh reduction in daily energy consumption	Distances, limited transport infrastructure, Harsh climatic conditions and cultural complexities
10	Koorie Energy Efficiency Project	Victoria	Major City & Regional	1124	Home visits and use of In-Home display monitors	Electricity bills from energy provider	Appliance replacement – Refrigerator	4% reduction in electricity consumption due to behavioural change with respect to energy management	Lack of adequate data, difficulty in obtaining electricity consumption data
11	Get Bill Smart	Tasmania	Major City	504	Home Energy Assessments	Data Loggers	Structural modifications, energy Saving devices & appliance upgrade	Average electricity saving of 2.8 kWh per day	Low interest and participation rate
12	Power Savers Program	NSW	Major City & Regional	1010	Home energy assessments	Energy bills	Structural modifications, energy Saving devices & appliance upgrade	An average reduction in electricity consumption by 0.5 kWh per household per day	Consent from landlords to implement energy efficient retrofits, Incomplete/unavailable data on electricity consumption from households
13	Northern Grampians Shire Council	Victoria	Regional	40	Home Energy assessments	Self-reported billing data	Structural modifications, energy Saving devices & appliance upgrade	3.213% and 8.712% reduction in peak and off-peak electricity usage respectively per household	Gaps in electricity consumption and/or billing data
14	Bright Actions Project	Queensland	Major City	1767	Personalized home visit and energy-efficiency workshops	Energy bills	Appliance upgrades – Refrigerator and washing machines	Undisclosed	Gaps in electricity consumption and/or billing data
15	Energy + Illawarra	NSW	Regional	830	Energy-efficiency workshops and indoor temperature monitoring	Energy bills	Structural modifications, energy Saving devices & appliance upgrade	Inconclusive	Gaps in electricity consumption and/or billing data
16	Powerplay	NSW	Major City & Regional	164	Surveys and energy efficiency awareness through gamification	Self-reported billing data	Nil	Inconclusive	Low participation rate
17	Energy Saver Study	Victoria	Regional	320	Personalized home visits	Self-reported billing data, Data Loggers & data from energy provider	Energy Saving devices	7% reduction in total energy use, improved average NatHERS rating from 2.7 to 3.4 stars	Nil
18	Our Green Home	Nationwide	Major City & Regional	600	Energy consumption awareness to inform behavioural change	Data Loggers & data from energy provider	Installation of energy monitors	Annual savings of 261 kWh per household	Low participation rate, Landlord objection to installation of the monitors
19	Beat the Heat	South Australia	Major City	200	Home energy assessment	Data Loggers	Appliance upgrade: high efficiency reverse-cycle air conditioner	12% reduction in annual energy consumption, an annual saving of $220	Low participation rate, tenant reluctance to participate, faulty equipment and failed installation
20	Glenelg Saves	Victoria	Regional	300	Home energy assessments	Data from energy provider	Nil	Rise in energy consumption by 5%	Reluctance to participate and lack of trust

Location: Following a review of available reports on each of the projects conducted in the LIEEP, the state of Victoria was the state with the highest number of projects around 35% (7) of the projects, NSW a distant second at 20% (4) (Figure 3(A)). From 20 projects executed, 95% (19) of the projects were carried out in major cities and regional areas, and 5% (1) in remote communities (and this was in the Norther Territory, specifically targeted towards indigenous Australians). No programmes were carried out in very remote communities which is the location of the case study project in this study (Figure 3(B)). This finding underscores the importance of the contribution of this study.

Figure 3. Proportion distribution of programmes. (A) Proportion distribution by state, (B) proportion distribution by remote structure/classification.

Approach/methodology: The approach to data collection has been classified into three categories: surveys, energy conservation workshops, and home visits and energy assessments. By proportion, surveys accounted for 20% (4), energy conservation workshops accounted for 25% (5) and home visits and energy assessments accounted for 55% (11) (Figure 4(A)). The energy consumption data was via self-reported energy bills by households, metre readings from energy providers and monitoring using data loggers, however, energy data was predominantly obtained from energy providers/retailers (Figure 4(B)).

Figure 4. Approach/methodology proportion by type. (A) Approach/methodology proportion, (B) proportion of energy consumption data collection by source.

Interventions: The LIEEP saw the implementation of energy conservation measures in most of the programmes (18 out of a total of 20 programmes had interventions). These interventions in the form of energy conservation measures consisted of the supply and installation of energy-saving devices such as standby power controllers (Alliance 2016), appliance upgrades such as air-conditioners (Byrne et al. 2016; O’Mullane and Hoch 2016; Sullivan 2016; Ward and Brent 2016), structural modifications such as ceiling and floor insulations (Steinborner et al. 2016), education and awareness via personalized feedback and energy conservation tips via SMS (Australian Government – Department of Industry 2016). The impacts of these interventions led to reductions in annual household electricity consumption between 3% and 30% depending on the type of intervention implemented where structural modifications had the greatest impact on energy consumption by 30%. It is noted, however, that one programme with no interventions saw an annual increase in energy by 5% (Lynch et al. 2016). Figure 5(A,B) provides some insights on the interventions implemented including impact on annual household energy consumption.

Figure 5. Interventions by type and impacts on annual electricity consumption. (A) Proportion of intervention by type, (B) percentage reduction in annual electricity consumption.

Challenges: Many challenges were also highlighted in the reports predominantly around the lack of available historical building and energy consumption data including difficulty obtaining such data. Even when these data were made available, cases of missing and inaccurate data were also reported particularly around the electricity consumption data. Challenges relating to remoteness were also reported with respect to logistics, climatic conditions, and cultural complexities. Further, cases on the reluctance of landlords to grant consent for households to implement energy conservation measures (ECMs) in the rented homes as well as low participation rates, faulty installations and failed equipment were also reported. Table 2 summarizes the magnitude of the challenges faced in terms of dominance.

Table 2. A snapshot on the challenges reported during the implementation of the LIEEP.

Challenge	Count	Percentage
Lack of available historical data	6	28
Low participation rate	7	33
Inaccurate/inadequate energy consumption data	4	19
Remoteness	1	5
Resistance from Landlords on implementation of ECM’s	2	10
Faulty equipment & failed installations	1	5

The findings above highlight the need to continue to grow the body of knowledge on energy management to support low-income households in Australia, particularly in very remote communities to enable them better manage their energy needs. The peculiarities and limitations in undertaking energy efficiency and conservation projects in remote communities means that access to reliable data at the very local level is needed to inform better decision making across regions and local levels. Reliable data that is easily accessible and understood can go a long way to empower communities in advocating for the local needs of their communities. This will also aid governments and other service providers in making better-informed decisions, improve efficiency and accountability and measure outcomes (SCRGSP 2020).

2.4. Residential energy efficiency and building energy simulation

The energy efficiency of residential buildings denotes the required amount of energy to facilitate the useful services of everyday practices in households. NatHERS remains the most popular pathway to meet the National Construction Code (NCC) energy efficiency requirements in Australia (Willand, Maller, and Ridley 2019). Dynamic energy simulation is one method for quantifying energy performance of buildings via the use of energy simulation tools thus, an accredited method for NatHERS and NCC energy efficiency compliance in Australia. Currently, there are over 400 tools applicable to the evaluation of energy efficiency, renewable energy and sustainability in buildings (IBPSA 2020). From this number, approximately 120 tools are just for whole building energy simulations. The reliability of such simulations, however, is dependent on several factors such as the quality and detail of inputs into the energy model, underlying uncertainties linked to available weather databases, the in-field measured parameters, and the data obtained from existing documentation (Ding et al. 2015; Erba, Causone, and Armani 2017; Littlewood et al. 2017). Further, the selected software tool and knowledge of the user also impact the outcome of simulation results. The use of computational software as a calculation method provides a standard and repeatable methodology and output values for selected criteria. This standardization is a benefit to designers, checkers and approvers as it provides familiarization, expediency, confidence and accuracy benefits (Australian Building Codes Board 2018). To demonstrate compliance, the NCC requires that the choice of simulation engines used for energy modelling meet the requirements of ANSI/ASHRAE Standard 140 (2007) and allow using a time-step of not more than one hour throughout an entire year from which to assess building performance, using representative local weather data as an input to calculations (Board 2019).

2.5. Climate of Wanarn and selected weather file

Wanarn is situated on the Indigenous lands of Ngaanyatjarra in WA. The climate is classified as arid to semi-arid (Desert), with hot summers. Temperatures are often above 40°C with moderately cool winters and often very cold nights. The meteorological view of the Aboriginal and Torres Strait Islander people is one of great diversity that cannot be meaningfully simplified to a rigid European seasonal calendar for the entire continent. Because of all this, seasonal cycles as described by the various Aboriginal cultures differ substantially according to location. In WA, climate is described according to the Nyoongar calendar which consists of six seasons namely: Kambarang, Birak Bunuru (Dry and hot season – October to March) and Djeran, Makuru, Djilba (cool and wet season – April to September; MacFarlane 2016). This classification splits the weather into two main seasons of hot and cool, where the hot season spans October to March and the cool season spans April to September. Typical Meteorological Year (TMY) and the Actual Meteorological Year (AMY) are the two common types of weather files used in building energy simulation studies (Bianchi and Smith 2019; Burleyson et al. 2018; Chan 2011; Hong et al. 2013; Hong, Chang, and Lin 2013; Lin and Hong 2013; Pyrgou et al. 2017; Tsoka et al. 2018). TMY’s contain statistically derived weather data from the past 10–30 years of historic data depending on availability, presents in 8760-hour format for use in simulation engines and do not capture extreme periods. AMY’s on the other hand, represent actual hourly observations for a given year and capture any extremes observed within the given year or any given timeframe. The choice of weather data for the simulation for this study was influenced by the need to account for seasonal variations and temperature extremes as this is reflective of the micro-climate. As such, the AMY data set was deemed appropriate over the TMY data set for this purpose.

The AMY data set for the 2018 calendar year representing 8760 h of weather data in .epw format, based on data obtained from Giles weather station 94461 from the Australian Bureau of Meteorology (ABOM) was used to perform the simulations. Prepared by Exemplary Energy in partnership with ABOM and the Commonwealth Scientific and Industrial Research Organisation (CSIRO), Satellite-derived solar Global Horizontal (GHI) and Direct Normal Irradiance (DNI) estimates are based on images from the Geostationary Meteorological Satellite GMS-5, Geostationary Operational Environmental Satellite (GOES-9) and MTSAT-1R satellites, which are provided with permission of the Japan Meteorological Agency (JMA) and the United States National Oceanic & Atmospheric Administration (NOAA) (Energy 2019). Data quality is assured via routine application of a capping algorithm to reduce over or under estimation impacts in line with the time convention of the applicable climate data. For solar data, each hourly value was calculated based on the minute-by-minute interpolated value for 60 min before the hour stated (1–24 h) in keeping with the convention of the TMY2 and EPW formats. Cloud cover data was inferred from solar data with the same time convention by relating it to coincident Clear Sky values. This was based on the total satellite-derived global irradiance over ∼5 km square pixels nearest to the target site, expressed as a fraction of the equivalent clear sky irradiance, to give a measure of fractional cloud cover that correlated well with observations (Maleny and Island 2017). For temperature, atmospheric Pressure and Humidity, all values were instantaneous on the hour as measured by ABOM, while wind was on the basis of averages for the 10 min preceding the stated hour as measured by ABOM (Lee 2011; Partners 2014). On-site measurements hourly measurements for temperature and humidity for 8760 h were then imported to replace corresponding ABOM data for simulation in DB and EP. A brief analysis of the weather data including the corresponding heating degree days (HDD) and cooling degree days (CDD) is presented in Table 3 and Figure 6. The base values for HDD and CDD calculations were 15 and 21°C heating and cooling, respectively. These fall within the ABOM recommended basepoints of 12 and 18°C for heating and 18 and 24°C for cooling (Australian Bureau of Meteorology 2022).

Figure 6. (A) Monthly radiation. (B) Monthly heating and cooling degree days.

Table 3. Brief analysis of selected weather file.

Month	Temperature °C			RH%	Dew point °C	Global solar Wh/m^2	Windspeed m/s	Annual heating degree days (base temperature = 15°C) = 185 days	Annual cooling degree days (base temperature = 21°C) = 2213 days
	Min	Max	Mean
Jan	17.2	47.4	31.5	47.8	11.0	312.6	5.2	0	329
Feb	19.1	47.9	30.4	45.3	14.2	289.0	4.9	0	263
Mar	20.3	46.6	30.6	28.3	4.4	252.0	4.9	0	297
Apr	14.5	45.3	27.9	29.8	2.3	201.0	4.5	1	216
May	9.1	40.6	19.6	23.5	−1.9	180.8	4.0	23	64
Jun	4.4	32.9	16.3	21.4	−4.0	162.4	4.3	59	26
Jul	5.1	34.4	17.8	20.8	−4.5	179.1	4.0	44	48
Aug	3.1	37.5	18.2	18.9	−6.8	212.0	4.3	47	58
Sep	9.0	40.1	23.0	17.5	−5.2	252.0	4.4	10	121
Oct	14.9	45.2	27.6	32.3	4.3	277.2	4.8	0	215
Nov	14.1	44.2	28.7	33.4	6.0	309.8	5.1	1	239
Dec	17.3	48.6	31.8	33.2	−1.8	257.9	5.2	0	337

2.6. Energy conservation

Energy conservation can be defined as the practice of reducing the amount of energy required for a building to operate in a manner that maintains the needed functionality for the occupants with respect to occupant comfort and wellbeing (Rahman, Rasul, and Khan 2010). Energy consumption in buildings is impacted by several factors such as climatic zones, building envelope, building materials, building occupants’ characteristics and behaviours, ventilation systems, artificial lighting and appliances and building design (Ochedi and Taki 2019). Energy efficiency interventions designed for implementation in households vary widely and include activities such as engineering retrofits to the home, undertaking home energy audits, the use of in-home displays, energy workshops and several behaviour-change activities (Geller et al. 2006). The key question for decision makers then becomes which of these intervention approaches should be adopted with respect to their effectiveness and impact while considering household comfort and wellbeing as well as cost implications. The link between accessibility and affordability of electricity to housing, education, health and social welfare outcomes discussed in the previous sections of this paper further suggests that benefits of an energy-efficient home go beyond reducing a household’s energy bill. In addition, there are also valuable societal benefits such as reduced hospitalizations due to improved health outcomes (Howden-Chapman and Preval 2014), a decrease in temperature-related mortality (Taylor et al. 2018), higher level of energy security and stability (Ayres, Turton, and Casten 2007; Selvakkumaran and Limmeechokchai 2013), as well as improved economic productivity (Anderson et al. 2014), which is highly beneficial to very remote communities. Energy conservation measures (ECMs) can be classified into two groups: Major intervention ECMs and minor intervention ECMs. The former are typically those interventions that relate to structural modifications or engineering retrofits on the building envelope such as wall and floor insulation upgrades, HVAC upgrades, window/glazing upgrades and external shading devices. These types of interventions typically attract a higher cost implication to implement. The latter, however, constitute those interventions that have minimal or no structural impact to the buildings as such lower cost implications to implement hence considered as minor investment ECMs (Rahman, Rasul, and Khan 2010). Examples include interventions like lighting upgrades, draft proofing using door and window seals, blinds for internal shading, energy-saving devices such as remote-control sockets, stove timers, and so on. Other measures/approaches to energy conservation related to behavioural change such as adjustments to indoor temperature setpoints are generally classified as no intervention/no investment ECM’s and may be considered however, studies have shown that energy conservation via the behavioural change approach has typically resulted in minimal impact towards reducing electricity consumption (Bird and Legault 2018; Delmas, Fischlein, and Asensio 2013; James and Ambrose 2017). Building energy simulations provide an effective means for the evaluation of the energy-saving potential of ECMs enabling informed decisions towards the selection of the most appropriate measure for implementation to be made especially in desert climates. For instance, Al-Tamimi (2022) used simulations to investigate the impact of different building envelope modifications on energy consumption of residential buildings in hot and arid climates in Qatar and found that thermal insulation can reduce energy consumption by 23% while an effective combination of shading, glazing and green roofs can reduce total energy consumption by 35%. Similarly, findings from studies on improving thermal behaviour of glazed windows in hot tropical regions show potential reduction in annual cooling loads by 7.2% and reduction in peak cooling loads by around 36% (Aboulnaga and Moustafa 2016; Al-Tamimi and Fadzil 2012; Sbar et al. 2012; Yu et al. 2020). Using thermal calculations and simulations, Goussous, Siam, and Alzoubi (2015) investigated the impact of green roof technology on HVAC energy consumption on a residential building in Jordan and the results showed that up to 17% of HVAC energy consumption could be saved with the simplest application of the green roof system. In terms of whole building energy conservation, Rahman (2010) investigated the various ECMs by optimizing HVAC systems, lighting, glazing, heating and cooling setpoints in an institutional building in Queensland, Australia and the results show a 41% energy saving without compromise to occupant thermal comfort. Other similar studies conclude that optimization of the building envelopes using energy simulations is an effective approach to energy conservation and can reduce annual energy consumption by up to 39% (Al-Sanea et al. 2016; Alaidroos and Krarti 2015; Alayed, O'hegarty, and Kinnane 2021; Çomaklı and Yüksel 2003; Dababt and Awad 2017; Hasan 2018). This study employs the method of building energy simulations to quantify the impact of energy conservation measures across the three mentioned categories to aid the decision process towards implementation in buildings in very remote communities. In this study, the interventions are categorized into major and minor intervention ECMs.

2.7. Overheating criteria and thermal comfort

Fanger’s predicted mean vote (PMV) and predicted percent dissatisfied (PPD) model is the widely accepted thermal comfort model (Fanger 1970). The PMV/PPD method reduces thermal comfort to a steady-state heat balance equation. Overheating criteria such as the Chartered Institution of Building Services Engineers (CIBSE) Guide A: Environmental design 2006 (Guide 2006) sets static design maximum thresholds at 28°C for living areas and 26°C for bedrooms is based on the PMV/PPD thermal comfort model, and requires that PPD should not be more than 10%. This means a maximum of not more than 876 h out of 8760 h in a year of discomfort hours or unmet heating and cooling loads is acceptable. This study applies the same approach when considering the balance between energy conservation measures and thermal comfort in the selection of optimal solutions from the simulation exercise.

3. Methodology

The methodology for this study adopts a case study approach using a residential building and stems from a practical applications point of view, consistent with existing building simulation and energy conservation studies. The objective of quantifying the impact of various ECMs for implementation in a remote Australian context was undertaken via optimization analysis in Design Builder (DB) and EnergyPlus (EP), based on genetic algorithm that investigates the interventions that offer the best performance solutions regarding energy consumption and thermal comfort.

3.1. Building and system description

The analysis focuses on a single-storey two-bedroom case study located in the Indigenous community of Wanarn, Western Australia which is classified as very remote. Ngaanyatjarra Lands where Wanarn is located is one of the most remote parts of Australia and arguably the world’s most remote community (McCafferty 2016). The selection of this case study stems from its suitability to generate unavailable performance data that is reflective of very remote communities that complement existing studies, which to date are primarily in remote or regional areas such as Alice springs and Laverton (Duell et al. 2006). The location of the case study building contains within it the key elements of the social, cultural, economic, environmental and geographic attributes responsible for the challenges associated high cost of electricity resulting in poorer outcomes for very remote communities. The selected case study is relevant to the practice of a new paradigm in sustainable design based on evidence from field studies on energy performance of households in very remote communities towards the design, construction, maintenance and operation of remote clinics in Australia. The degree of remoteness means that climatic conditions are harsher, infrastructure provision and maintenance costs are higher and the ease of transport and assembly makes the lightweight construction the dominant building material typology, which differs from the other remote area classifications. Further, limitations with respect to financial, time and logistical constraints made it impossible to monitor more than one building type/design at a time. Access with respect to obtaining necessary permits and the willingness of households to participate also limited the selection of buildings to be investigated at the time of the research. It is however expected that findings from this pilot study will provide useful insights to trigger further work on a broader scale specifically targeting very remote communities to grow the energy performance database. The reference model occupies a net floor area which is 67.1 m² with one occupant and the net area of unoccupied roof space is 71.8 m². The building is characterized by lightweight steel-framed metal cladded walls on the exterior with plasterboard lining on the internal walls. Roof and walls are insulated, windows are single glazed, heating and cooling are achieved by mechanical means through unitary split AC systems (Figure 7).

3.2. Data collection

Actual annual electricity consumption was measured for a period of 12 months from January 2018 to December 2018 using the WattNode ModBus Wattmeter which recorded an annual electricity consumption of 275.5 kWh (4.5 kWh/m²·yr). The input values for the simulation exercise with respect to the thermo-physical characteristics of construction elements, environmental control parameters and schedules including lighting and occupancy were obtained from as-built documentation, site visit/building survey and casual conversations with the occupant. Other inputs such as lighting and equipment power density have been calculated based on site information with guidance from the provisions of the Building Code of Australia (BCA) for class 1 buildings (Table 4).

Figure 7. Floor plan, 3D model and Section of case study building using DB.

Table 4. Building simulation input data for base case energy simulation.

Characteristic/component	Description of base model
Location/orientation	Western Australia, Australia – East facing – 270° North
Building type/shape	Residential/L-shape
Building height	3 m
Floor area (occupied)	67.1 m^2 (61.8 m^2)
External wall	100 mm lightweight steel construction with metal cladding. R = 2.8 m^2K/W, U = 0.35 W/m^2/K
Internal wall/partitions	2 × 25 mm gypsum plasterboard with 100 mm cavity. R = 0.61 m^2K/W, U = 1.64 W/m^2/K
Internal floor	100 mm Concrete Slab with Tiles. R = 0.37 m^2K/W, U = 2.65 W/m^2/K
Roof	Light coloured colourbond roof sheet, R = 4.5 m^2K/W
Glazing	Single glazed (SHGC = 0.8)
Wall to window ratio	North = 0%, South = 1.5%, West = 13.5%, East = 13.3%
Number of occupants	1
Lighting	Internal – LED 8W (2700 K) – 3.62 W/m^2, External – 5 W
HVAC	Unitary split system. Heating COP – 5.81, Cooling COP – 5.95 Operating hours: Winter – 2 hrs per day, Summer – 4 h per day
Ventilation rate	10.0 L/s/person
Infiltration rate	0.3 ACH
Temperature setpoint	24°C in hot season (October–March), 24°C in cool season (April–September)
Occupancy profile	Single Adult – Fulltime worker
	Weekdays: 18 hrs
	Weekends: 24 hrs
Occupancy	Metabolic rate = 1, Clothing insulation = 0.5 (hot season), 1.0 (cool season)
Equipment	3.0 W/m^2

3.3. Base model development, simulation, calibration and validation

The base model was developed using DB and EP software. DB happens to be one of the most comprehensive user interfaces for EP and maintains the standards of the European Parliament Board of Directive (EPBD). The EP solution is based on the heat and mass balance technique referred to as the Predictor–Corrector Method and assumes that the room air is well stirred, providing a uniform temperature throughout. It also assumes that room surfaces such as walls, windows, ceilings and floors have uniform surface temperatures, uniform long and short wave irradiation and diffuse radiating and reflecting surfaces (Crawley et al. 2000). In the modelling, building blocks were considered as composed of building elements such as walls, floor slab and roof, with internal partitions forming four thermal zones to reflect the accommodation schedule. These zones, i.e. bedroom 1, bathroom, bedroom 2 and living/dining/kitchen were partitioned and modelled according to the HVAC drawing. The building is occupied by a single fulltime health care worker hence occupation is estimated at 13 and 24 h on weekdays and weekends, respectively. The air conditioning units operate for approximately 2 h per day in winter and 4 h per day in summer. Table 4 provides details of the base model inputs used to simulate the base model results. The subsequent base model energy simulation resulted in total annual electricity consumption of 339 kWh (5.5 kWh/m²·yr), an overestimation by 63 kWh (1 kWh/m²·yr), hence the model was calibrated for validation against actual data to ensure its suitability for optimization. To calibrate the model, dynamic parameters based on assumptions on operational schedules and user behaviour of the occupant profile were adjusted including lighting power density. Occupancy hours on weekdays and weekends were reduced to 16 and 18 h per day respectively, HVAC operation was reduced from 6 h per day to 4.5 h per day, while lighting power density was reduced from 3.62 W/m² to 2.95 W/m² as shown in Table 5. This is to account for the possibility that not all spaces are occupied at the same time for a sole occupant health care worker, working fulltime.

Table 5. Adjusted parameters for calibrating the base model.

Space	Occupancy ratio (%)	Occupancy (hrs/day)	HVAC operation (hrs/day)	Lighting power density (W/m^2)
Living/kitchen	21	Weekdays = 16	Summer = 3
Bedroom 1	42			2.95
Bedroom 2	0	Weekends = 18	Winter = 1.5
WC/bath	4
Unoccupied (at work/out)	33

Calibration results and validation with actual data is shown in Figure 8. Monthly electricity consumption data was used for validation in line with the provisions of International Performance Measurement and Verification Protocol (IPMVP) (IPMVP 2002). Three parameters considered for base model validation were annual (January to December), hot season (October to March) and cool season (April to September) energy consumption in the computational simulation process and results compared against acceptable tolerances. For instance, ASHRAE-209 Guideline 14 defines acceptable limits for variances between measured and simulated energy consumption results at 15% Coefficient Variance of Root Mean Square Error (CV-RMSE) and ± 5% Mean Bias Error (MBE), respectively, for energy simulation results (ANSI/ASHRAE 2018). Other studies recommend that the annual simulated energy use should be within 10% of metered data, while a difference of less than 25% is acceptable on a seasonal basis (Kaplan and Caner 1992; Rahman 2010). The calibration results show that annual and seasonal electricity consumption fall within acceptable tolerances, thus, validates the model as shown in Table 6.

Figure 8. Simulation results showing fuel breakdown by end use.

Table 6. Annual and seasonal energy consumption showing CV-RMSE & MBE metric comparison between measured and calibrated simulated data.

	Measured data (kWh/m^2·yr)	Simulated data (kWh/m^2.yr)	MBE	CV-RMSE
Annual	4.5	4.5	0.03	0
Hot season	2.6	2.5	0.94	0.24
Cool season	1.9	2.0	−0.91	0.34

3.4. Evaluation of ECMs

The retrofit options were investigated and discussed according to the classifications of major intervention ECMs and minor intervention ECMs. The major intervention ECMs were analysed based on changes to the properties of external walls, glazing, shading devices and orientation while the minor intervention ECMs analysed the impacts of lighting upgrades and controls, air tightness and adjustments to indoor temperature setpoints. The interventions are analysed and evaluated based on the annual energy use pattern of the building without compromising occupant thermal comfort. Figure 9 provides a graphical representation of the analysis methodological framework.

Figure 9. Graphical representation of analysis methodological framework.

3.5. Optimization

Prior to commencing the optimization process, a sensitivity analysis was conducted based on the multiple regression method to understand the input variables with the potential to have the greatest impact on electricity consumption and overheating metric (unmet cooling hours) (Figure 10). This is deemed useful to save time and effort by eliminating variables with minimal impact during the optimization process. Figure 10(A,B) shows the results from the sensitivity analysis for electricity and overheating metric, respectively. The variables are ranked in order of importance where green, yellow and red indicate high, medium and low importance of the variables, respectively.

Figure 10. Sensitivity analysis. (a) Electricity consumption (kWh/m² yr). (b) Overheating metric (hrs).

The multi-objective optimization process was employed to assess the impact of the various interventions as such, optimization objectives, constraints and variables were used to guide the simulation process in identifying the best measures/combinations of measures that meet the objectives. The preferred optimal solution is highlighted by the red box in each table. The goal of this phase was to investigate the option(s) that reduce total electricity consumption with minimal unmet cooling hours. Based on this approach, optimal solutions were those deemed to best balance the trade-off between energy consumption and this overheating metric. The constraint set was for the annual energy consumption to be less than 4.5 kWh/m²·yr.

4. Results

4.1. Major intervention ECMs

Orientation: The optimization run for orientation involved 100 generations that resulted in a total number of 2428 iterations when the orientation is adjusted between 0° and 360°. The results show four iterations highlighted in red in Figure 11. They offer the best performance since they are closer to the overall Pareto front, that is, the contour of solutions that are not surpassed by other solutions regarding minimum total annual electricity consumption and overheating metric. The results show that appropriate orientation of the building can reduce annual electricity consumption by 14% (0.62 kWh/m²·yr) per year and reduce overheating metric by 55% (21.5 h) per year when compared with the base case, depending on the optimal solution considered. Iteration 54 is selected as the optimal design as highlighted in Table 7.

Figure 11. Pareto front of orientation showing selected optimal solutions.

Table 7. Simulation results for orientation optimization parameter.

Iteration	Model	Electricity consumption (kWh/m^2·yr)	Overheating metric (hrs)	Orientation (°)
0	Base case	4.5	39.2	270
5	Optimal Designs	3.83	16.2	180
9	Optimal Designs	3.88	14.8	170
32	Optimal Designs	3.85	15.2	175
54	Optimal Designs	3.80	17.1	185

Wall parameters: For the wall parameters, two wall assembly options were simulated and targeted external walls only; state-of-the-art lightweight and medium/heavy weight wall construction. Figure 12 and Table 8 provide representation of wall assembly types and properties in comparison to the base model including the results.

Figure 12. External wall assembly options considered in the optimization process.

Table 8. Simulation results for wall construction optimization parameters.

Wall type	Description	Model	Iteration	Electricity consumption (kWh/m^2·yr)	Overheating metric (hrs)	R-value (m^2K/W)	U-value (W/m^2/K)
a	100 mm lightweight steel construction with metal cladding	Base	0	4.5	39.2	2.8	0.35
B	140 mm lightweight steel construction with metal cladding	Optimal design	1	4.2	34	3.8	0.26
C	330 mm medium/heavy weight brick and concrete wall assembly	Optimal design	2	3.5	25	3.9	0.25

For the lightweight optimized wall construction (Figure 12(B)), the U and R values were slightly improved which increased the overall thickness of the wall by 40 mm when compared to the base model. The medium/heavyweight wall construction had a significantly improved wall construction type to include brickwork to replace the metal cladding on the outer layer and an additional layer of concrete block (Figure 12(C)). This increased the overall wall thickness by 235 mm including corresponding U and R values for the wall assembly when compared against the base model. A total of 2232 iterations were completed during the optimization process and the results from Table 8 show that using medium/heavyweight wall construction can reduce the annual electricity consumption by 21% (0.96 kWh/m²·yr), also minimizing overheating metric by 35% (13.6 hrs). Similarly, the optimized lightweight wall construction can reduce the annual electricity consumption by 5% (0.2 kWh/m²·yr) while minimizing overheating metric by 11% (4.6 hrs) annually. This finding underscores the importance of a high thermal mass towards energy conservation within this climate for a residential building.

Glazing: This energy conservation measure evaluates the impact of upgrading the existing external glazing with more efficient options. Three types of glazing were compared as described in Table 9. Low-emission clear single glazing, clear double glazing and low-emission double glazing options were simulated. After simulating a combined total of 2227 iterations for the three glazing types, the results show a reduction in annual total electricity consumption by 7% (0.31 kWh/m²·yr), 10% (0.47 kWh/m²·yr) and 17% (0.79 kWh/m²·yr) for low-emission clear single glazing, clear double glazing and clear low-emission double glazing options, respectively. Similarly, overheating metric was reduced by 23% (9.4 hrs), 34% (13.5 hrs), and 57% (22.7 hrs), respectively. The low-emission double glazing option was applied to the eastern and western-facing facades in isolation and results showed that applying this option had a greater positive impact on the eastern façade than on the western façade indicating that more attention is needed in the façade treatment for the west facing walls with respect to reducing energy consumption. The preferred optimized solution is highlighted in Table 9.

Table 9. Simulation results for glazing optimization parameters.

Iteration	Model	Electricity consumption (kWh/m^2·yr)	Overheating metric (hrs)	Glazing type	SHGC	Solar transmittance	U-value (W/m^2/K)
0	Base case	4.5	39.2	6 mm clear single glazing	0.82	0.8	5.8
1	Optimal Designs	4.1	29.8	6 mm clear Low-emissions single glazing	0.72	0.68	3.8
2	Optimal Designs	3.6	16.5	6 mm clear Low-emission double glazing	0.57	0.47	2.4
3	Optimal Designs	3.9	25.7	6 mm clear double glazing	0.7	0.6	3.1
4	Eastern façade	3.7	19.2	6 mm clear Low-emission double glazing	0.57	0.47	2.4
5	Western façade	3.8	19.8	6 mm clear Low-emission double glazing	0.57	0.47	2.4

Shading devices: Given that the glazing on the building is concentrated on the eastern and western facades, shading devices were introduced to all external glazing in these areas to manage heat gains (Figure 13). The eastern and western facades of a building receive large amounts of sunlight in the early morning and late afternoon, respectively. In this instance, four shading types were explored on these facades as follows: 0.5 m overhang (Figure 13(A)), 0.5 m projection louvres (Figure 13(B)), 0.5 m projection overhang + side fins (Figure 13(C)), and a combination of 0.5 m projection louvre + overhang + side fins (Figure 13(D)).

Figure 13. (A) 0.5 m overhang, (B) 0.5 m projection louvres, (C) 0.5 m projection overhang + side fins, (D) 0.5 m projection louvre + overhang + side fins.

From the results as presented in Table 10, optimal design 4 has the lowest energy consumption and overheating metric. The results show that combining the three options can reduce electricity consumption by up to 6.9% (0.29 kWh/m²·yr) in comparison to a single option. Further, this option reduces electricity consumption by 12% (0.54 kWh/m²·yr), while minimizing overheating metric by 44% (17.4 hrs) when compared with base case.

Table 10. Simulation results for shading options on eastern and western façade glazing.

Option	Iteration	Model	Electricity consumption (kWh/m^2·yr)	Overheating metric (hrs)	Shading type
N/A	0	Base case	4.5	39.2	No shading
A	1	Optimal Designs	4.2	32.1	0.5 m overhang
B	2	Optimal Designs	4.0	28.2	0.5 m projection louvres
C	3	Optimal Designs	4.1	29.2	0.5 m projection overhang + side fins
D	4	Optimal Designs	3.9	21.8	0.5 m projection louvre + overhang + side fins

An optimization simulation process combining all four major intervention ECMs was also performed to understand overall impact on annual energy consumption and overheating metric when combining all four variables. A total of 4638 iterations for multiple combinations were performed to determine optimal design solutions. From this exercise, two optimal designs emerged as shown in Table 10. The results showed the variables with the best balance between minimizing annual electricity consumption and overheating metric. For the shading variable, options C and D (Figure 13) featured in all optimal solutions while orientation angles ranged from 185° to 200° for best outcomes. For the external wall construction variable, it is observed that 140 mm lightweight construction yields high performance when in combination with other variables as it occurred in 6 options out of 17 while the medium/heavy weight wall construction type occurred in 11. For glazing type variable, only double clear glazing and double low-emission clear glazing featured in the optimal solution when in combination with the other variables (Table 11). Iteration 17 provides the best outcome as it reduces electricity consumption by 33.5% (1.49 kWh/m²·yr) with overheating metric reduced by 100% (0 hrs).

Table 11. Simulation results for combination of all major intervention ECMs.

Iteration	Category	Electricity (kWh/m^2·yr)	Overheating metric (hrs)	Shading type	Orientation (°)	External wall construction	Glazing type
0	Base case	4.5	39.2	No shading	270	100 mm lightweight	Sgl Clr 6 mm
1	Optimal Designs	2.99	0.2	0.5 m Overhang + side fins + Louvres	195	330 mm Heavyweight	Dbl Clr 6 mm/13 mm Air
2	Optimal Designs	2.99	0.0	0.5 m Overhang + side fins + Louvres	200	330 mm Heavyweight	Dbl LoE (e2 = .1) Clr 6 mm/13 mm Air
3	Optimal Designs	2.96	1.8	0.5 m Overhang + side fins	195	330 mm Heavyweight	Dbl Clr 6 mm/13 mm Air
4	Optimal Designs	2.96	1.7	0.5 m Overhang + side fins	185	330 mm Heavyweight	Dbl Clr 6 mm/13 mm Air
5	Optimal Designs	3.27	1.9	0.5 m Overhang + side fins	185	140 mm Lightweight	Dbl LoE (e2 = .1) Clr 6 mm/13 mm Air
6	Optimal Designs	2.97	0.0	0.5 m Overhang + side fins	195	330 mm Heavyweight	Dbl LoE (e2 = .1) Clr 6 mm/13 mm Air
7	Optimal Designs	2.99	0.0	0.5 m Overhang + side fins + Louvres	190	330 mm Heavyweight	Dbl LoE (e2 = .1) Clr 6 mm/13 mm Air
8	Optimal Designs	3.29	5.7	0.5 m Overhang + side fins + Louvres	200	100 mm Lightweight	Dbl LoE (e2 = .1) Clr 6 mm/13 mm Air
9	Optimal Designs	2.97	0.0	0.5 m Overhang + side fins	190	330 mm Heavyweight	Dbl LoE (e2 = .1) Clr 6 mm/13 mm Air
10	Optimal Designs	2.96	1.7	0.5 m Overhang + side fins	190	330 mm Heavyweight	Dbl Clr 6 mm/13 mm Air
11	Optimal Designs	3.29	5.9	0.5 m Overhang + side fins + Louvres	185	100 mm Lightweight	Dbl LoE (e2 = .1) Clr 6 mm/13 mm Air
12	Optimal Designs	2.99	0.0	0.5 m Overhang + side fins + Louvres	185	330 mm Heavyweight	Dbl LoE (e2 = .1) Clr 6 mm/13 mm Air
13	Optimal Designs	3.36	8.2	0.5 m Overhang + side fins + Louvres	195	100 mm Lightweight	Dbl Clr 6 mm/13 mm Air
14	Optimal Designs	3.29	5.8	0.5 m Overhang + side fins + Louvres	195	100 mm Lightweight	Dbl LoE (e2 = .1) Clr 6 mm/13 mm Air
15	Optimal Designs	2.99	0.0	0.5 m Overhang + side fins + Louvres	195	330 mm Heavyweight	Dbl LoE (e2 = .1) Clr 6 mm/13 mm Air
16	Optimal Designs	3.36	8.5	0.5 m Overhang + side fins + Louvres	185	100 mm Lightweight	Dbl Clr 6 mm/13 mm Air
17	Optimal Designs	2.97	0.0	0.5 m Overhang + side fins	185	330 mm Heavyweight	Dbl LoE (e2 = .1) Clr 6 mm/13 mm Air

4.2. Minor intervention ECMs

Lighting upgrade and controls: lighting controls were introduced to internal lighting only given the building already had efficient LED lighting with a power density of 2.95 W/m². For external lighting, the optimization included upgrading the external lights to lighting with an absolute power of between 1 W and 5 W. The results show a decrease in annual electricity consumption by 6.7% (0.3 kWh/m²·yr), 1.7% (0.08 kWh/m²·yr) and 8.1% (0.36 kWh/m²·yr) when external lighting absolute power density is at 1 W, when lighting controls are introduced to internal lighting and a combination of both options are implemented respectively (Table 12). No significant difference in discomfort hours was observed between the options.

Table 12. Simulation results for lighting controls and upgrade options.

Category	Electricity consumption (kWh/m^2·yr)	Overheating metric (hrs)
Base model	4.5	39.2
External lighting	4.16	39.2
Lighting controls	4.37	40
External lighting + Lighting controls	4.09	40

Air tightness: In Australia, there is also no requirement for whole building air tightness testing as such there have only been a handful of relatively small buildings tested in Australia to date (GBCA 2015). The energy efficiency provisions of the National Construction Code (NCC) require the air permeability rate of the building envelope to be about 10 m³/hr·m² at 50 Pascals. NatHERS considers a rate of 15 air changes per hour (ACH) when the house is pressurized to 50 Pascals a rough acceptable average (Ambrose and Syme 2017). Given no testing was undertaken in this research, the optimization parameters looked at options for ACH values between 0 and 20.

After 4912 iterations, the results show no change between the baseline ACH of 0.3 and adjustments between 0.3 and 20 ACH in minimum energy or discomfort metric (Table 13).

Table 13. Simulation results for air tightness options.

Category	Electricity (kWh/m^2·yr)	Overheating metric (hrs)	Airtightness
Base	4.5	39.2	0.3
Optimized design	4.5	39.2	0–20

Adjusting heating and cooling setpoint temperatures: Heating and cooling temperature setpoints were adjusted by ±1° each and results on the impact on energy consumption and overheating metric compared with the base model. Table 14 shows a reduction in electricity consumption by 17.8% (0.8 kWh/m²·yr) and minimized overheating metric by 21% (8.6 hrs) when heating and cooling setpoint temperatures change from 24°C to 23°C and 25°C, respectively.

Table 14. Optimization results for heating and cooling setpoint temperature.

Category	Electricity (kWh/m^2·yr)	Overheating metric (hrs)	Cooling setpoint temperature (°C)	Heating setpoint temperature (°C)
Base model	4.5	39.2	24	24
Optimized design	3.66	30.6	25	23

Similar to the major intervention ECMs, optimization was also done when all minor intervention ECMs are considered together to determine the impact on annual electricity consumption and overheating metric. The resultant optimal design minimized electricity consumption by 25% (1.16 kWh/m²·yr) and overheating metric by 21% (8.6 hrs) as shown in Table 15.

Table 15. Simulation results for combination of all minor intervention ECMs.

Iteration	Category	Electricity consumption (kWh/m^2·yr)	Overheating metric (hrs)
0	Base	4.5	39.2
1	Optimal Design	3.34	30.6

Ranking ECMs based on performance indicators: Following the optimization process for the various ECMs, the results facilitated an understanding of the key performance indicators based on the impact of individual ECMs and combining all major and minor intervention ECMs on energy consumption and overheating metric. These options were ranked in the order of performance from highest impact (1) to lowest impact (9) on the building’s energy consumption using total annual energy consumption as the primary basis for comparison, secondary to the corresponding overheating metric in hours (Table 16 and Figure 14). It is noted that the HVAC system was not considered as an ECM measure because the existing HVAC system is a highly efficient system with heating and cooling coefficient of performance of 5.81 and 5.95, respectively. Rather, the impact of ECMs is based on percentage variance against the reference model based on known inputs from site building and site survey investigations and compared with actual monitored electricity data.

Figure 14. Graphical representation of various ECMs and impact on energy consumption and occupant thermal comfort.

Table 16. Rankings of ECMs based on minimum annual energy consumption.

Ranking	ECM	Intervention type	Performance indicator	Energy consumption (kWh/m^2·yr)	Overheating metric (hrs)
N/A	Base model	N/A	Base model	4.5	39.2
6	Orientation	Major	185° Orientation	3.84	17.1
3	Wall parameters	Major	330 mm Medium/heavy weight wall	3.50	25
5	Glazing	Major	6 mm clear Low-emission double glazing	3.66	16.5
7	Shading device	Major	0.5 m projection Louvre	3.92	21.8
8	Lighting upgrade & Controls	Minor	External lighting (1W) + Internal lighting controls	4.09	40
9	Air tightness	Minor	0–20 ACH	4.46	39.2
4	Heating & Cooling setpoints	Minor	Cooling (25°), heating (23°)	3.66	30.6
1	Combined – Major	Major	Iteration 1	2.97	0
2	Combined – Minor	Minor	Iteration 1	3.34	30.6

A summary of the results from Table 16 and Figure 14 shows that overall, combining all minor intervention ECMs provides the best performance by reducing energy consumption by 33.5% (1.49 kWh/m²·yr) when compared against the base model and measured data. For individual ECMs, increased thermal mass by upgrading external walls to 330 mm with insulation performs best by reducing energy consumption by 21% (0.96 kWh/m²·yr) when compared against the base model and measured data. Adjusting heating and cooling setpoint temperatures ranked 4th, decreases electricity consumption by 17.8% (0.8 kWh/m²·yr.) which suggests that minimal investment measures can also significantly reduce energy need and subsequent energy cost for households in remote communities.

The findings provide insights on retrofit options aimed at reducing energy consumption while maintaining thermally acceptable indoor environments for existing buildings however, it is also useful towards the design of new buildings when setting performance targets. To this effect, cumulative energy saving percentages are presented in Table 17 and Figure 15. The results show the cumulative impact on energy reduction when intervention measures are added one after the other based on their initial performance ranking.

Figure 15. Graphical representation showing cumulative energy saving percentages from retrofit interventions.

Table 17. Cumulative energy saving percentages from retrofit interventions.

ECM	Energy consumption (kWh/m^2·yr)	Savings (kWh/m^2·yr)	Cumulative energy savings (%)
Air tightness	4.5	0.00	0
Air tightness + Lighting upgrade & Controls	4.1	0.36	8.1
Air tightness + Lighting upgrade & Controls + Shading device	3.9	0.61	13.6
Air tightness + Lighting upgrade & Controls + Shading device + Orientation	3.4	1.03	23
Air tightness + Lighting upgrade & Controls + Shading device + Orientation + Glazing	3.0	1.43	32
Air tightness + Lighting upgrade & Controls + Shading device + Orientation + Glazing + Heating & Cooling setpoints	2.8	1.71	38
Air tightness + Lighting upgrade & Controls + Shading device + Orientation + Glazing + Heating & Cooling setpoints + Wall parameter	2.5	2.00	44

The results from Table 15 show that implementing ECMs can reduce electricity consumption by up to 44%. This percentage can further be reduced when renewable energy systems such as PV solar power are considered (Industry and Science 2015). For very remote communities there is great enthusiasm towards this form of energy generation (Byrnes et al. 2016). However, the geographical remoteness, dispersed consumers, off-grid electricity supply reliance, low consumption, high supply and maintenance costs and limited ability to pay remain barriers to implementation (Byrnes et al. 2016; Reiche, Covarrubias, and Martinot 2000; Yadoo and Cruickshank 2010). The high upfront and subsequent maintenance cost does not make it economically viable to implement as such, has been excluded this study at this time.

5. Conclusion

Electricity costs remain a significant cost-of-living pressure and stress for many households on low incomes, with greater impact on residents of remote and very remote communities in Australia. Energy conservation measures in buildings have been known to be an effective measure in reducing energy need in households which in turn reduce the cost of electricity bills. This paper adopts an evidence-based approach on a pilot study to explore the impact of various ECMs on an existing residential building within a remote context with the objective of identifying suitable options for interventions that can minimize energy requirements and subsequent costs in very remote households if implemented. It also sought to provide a base for future similar studies in the region based on the findings.

Building performance quantification methods of physical measurements and simulations were used to monitor actual annual electricity consumption in the case study building for 12 months, and the resultant data was used for comparison and validation purposes with simulated results respectively. Further, a systematic evaluation was done for various ECMs on the basis of the relationship between their energy savings and occupant thermal comfort and ranked in order of performance following a sensitivity analysis and optimization process. The ECMs were categorized into major intervention and minor intervention where the minor intervention ECMs requiring minimal financial investment could directly apply to existing buildings while the major intervention ECMs such as orientation and wall parameters could typically apply to the design of new buildings.

The results show that implementing all the major ECMs provides the greatest impact in reducing energy consumption by up to 44% therefore has the highest ranking as shown in Table 16. Other performance indicators such as adjusting the heating and cooling setpoint temperatures of the HVAC system by ±1°C, lighting upgrades and controls, and the application of shading devices to all external windows can save up to 17.8%, 8.1% and 12% of annual electricity consumption, respectively. However, as a major outcome, it was observed that lightweight construction performs poorly (5% reduction) in this climate compared to medium/heavy weight (21% reduction) wall type when considered as a standalone intervention measure. This result is of particular significance given that the dominant wall construction type in very remote communities is lightweight steel construction due to the ease of assembly, transportation and delivery in response to logistical and economic challenges faced in this region. Therefore, careful consideration should be given to this element of the building envelope in terms of trade-offs during decision making. The study revealed that ECMs can be successfully applied to buildings located in desert like climates which is hot and arid where high-energy use of air conditioning systems is required to maintain thermal comfort. In very remote communities with extreme weather conditions such as in Wanarn, the implementation of minor intervention energy conservation measures as demonstrated in the results of this study is strongly recommended when considering any renovation or refurbishment works due to the economic viability. This study, while limited to a single building typology and occupancy profile does not lend itself to make generalizations, but rather speaks to the unique characteristics of very remote communities where the one size fits all approach is seldom successful in the region. Rather, the findings are deemed useful towards the gradual application across the differing typologies and occupancy profiles that characterize these communities, further expanding this body of work. This in turn will go a long way in alleviating the cost pressures on everyday living and hardship faced by households due to high-energy costs.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available from the corresponding author, [S.U.], upon reasonable request.