A Method and Metrics to Assess the Energy Efficiency of Smart Working

Abstract

The paper discusses the energy efficiency of smart working (SW) as a solution to traditional work-approach issues, with a focus on evaluating benefits for both employers and employees. Remote working, while offering environmental advantages such as reduced commuting and office space use, poses challenges in assessing its true impact. The study presents results from a dynamic analysis on a real residential building, typical of an architectural style diffused in northern Italy, revealing a 15% average increase in energy consumption when all work tasks are performed from home. To address concerns about the environmental impact of SW, the research proposes a method and metrics for evaluation. Four novel indices (SWEET, SEE, SSEE, and 4E) are introduced, providing a structured approach to assess the energy efficiency of SW initiatives. The paper outlines a methodology for data gathering and metric application, aiming to acquire quantitative insights and mitigate disparities arising from a transfer of burdens to employees. This contribution not only signifies a ground-breaking methodology but also addresses an unresolved research question concerning the evaluation of the actual energy efficiency of smart working implementations for both employers and employees. The results underscore the importance of understanding the nuances of SW’s impact on household energy usage and its broader implications for sustainability goals.

1. Introduction

In the last two decades, flexible working practices have been developed and employed to address some traditional working approach issues. The revolutionary aim of such an approach, which deconstructs the traditional working mode strictly defined in place, time, and pyramidal structures, is to achieve a “win-win” condition for both employers and employees, providing benefits for both categories [1].

During the pandemic period, remote working, which includes, as said by Soga et al. [2], home working, teleworking, and smart working (SW), named also hybrid working according to Mc Phail et al. [3], has been the only mean to mitigate the lockdown impact on productivity. One of the most important legacies left by such a period seems to be a trend confirming hybrid working, already experienced to have good results before the pandemic, as the new configuration for several companies. This is because, not only do many employees prefer such a working method, as remembered by Vij et al. [4], but also because office spaces have been reduced and the working chain has been adapted to such an approach [3]. It is worth mentioning that smart working strongly differs from teleworking or home working, as emphasized by Loi [5]. In fact, it is based on flexibility and autonomy, thus it is not bound to a place or to the same working hours of the office [5]. Moreover, it allows employees to develop their duties remotely (at home, co-working spaces, other suitable places chosen by workers) while still having the option to access the offices when needed, maintaining contact with colleagues. Smart working can, thus, be a means to obtain the advantages of remote working while mitigating some of its drawbacks.

From the environmental point of view, the main positive voices related to remote working can be summarised as workers’ energy and emission savings due to reduced commuters’ travel, as highlighted by Roberto et al. [6], and energy and emission savings for companies due to different office use, as remembered by Osman et al. [7].

Besides the other acknowledged advantages of the flexible working practice, like those remembered by Angelici et al. [8], there are also some issues [2]. In particular, one of the open research questions concerns the real environmental advantage evaluation connected to work carried out from home or other locations, different from the traditional office, as underlined by Hook et al. [9]. The mentioned research underscores the need for an evaluation approach capable of considering the dynamic configurations of remote working. Specifically, it is required, firstly, to calculate the possible environmental impacts considering the peculiarities of each remote working configuration and secondly to take into account not only the employers’ advantages, in terms of energy saving, but also the possible energy consumption increments borne by employees. This second point is also remembered in [2], where such a research gap is reformulated as the lack of understanding for whether the companies’ energy saving results from a contextual burden shift to the employees or not. Crow et al. [10] state that the household increased consumption, due to hybrid work adoption, can be estimated between 7% and 23%. On the other hand, in the same research, it is said that commuting energy savings counterbalance and exceed, on average, such expenses. Also, Cassetti et al. [11] calculated a positive average result. Nevertheless, it is important to avoid specific situations, that are not so uncommon, like those described by Battisti et al. [12], where the fears expressed by Soga et al. come true because the net economic balance of workers is negative. Thus, a more thorough analysis is required to capture the peculiarities of different situations, particularly evaluating and taking into account the possible increased energy consumption of employees’ dwellings, or other working places, when used instead of the classic offices.

The current research aims to give a first answer to the cited open research questions by proposing a method and metrics that permit users to assess the net environmental advantages of remote working on a case-by-case basis. Such a method takes into account the peculiarities of the adopted working configuration, making it particularly suitable for the current evolving situation, where new forms of remote working are continuously developing [13].

The method is based on a comparison between an “ante” and a “post” configuration considering, in the balance, the main factors affecting energy use: the possible increased energy consumption related to the different employees’ dwellings, or other working spaces, occupation, the avoided commuting expenses, and the office energy savings.

The method provides an approach and a metric able to determine the net environmental impact, allowing users to understand to what extent the proposed working configuration could mean to shift the energy expenditure from the employers to employees.

Due to the said international trend of increasing hybrid work experiences, the current research particularly focuses on SW. The research also presents an overview of the Italian situation, which is particularly interesting because the concept of “agile” work was officially introduced only in the year 2017 by the law 81 [14] and, ultimately, it has been largely employed, primarily due to the pandemic crisis, even if some SW experimentations had been already carried out before such an event. Furthermore, a detailed analysis of the energy consumption variations in a real building, representative of the typical architectural style prevalent in the 1960s, the era during which the majority of the national building stock was constructed, was conducted. This analysis considered different occupancy profiles to comprehend the extent to which SW can influence household energy consumption.

The research work is, thus, developed in the following steps:

• A short methodology description;
• A short overview of the Italian situation and a brief description of the main examples of smart working practices;
• A dynamic analysis of the energy consumption of a real building, taking into account different occupancy profiles and days of SW;
• The proposal of an evaluation method and related metrics;
• Some considerations about possible future developments.

The present study contributes to the existing body of knowledge about the energy efficiency of SW, introducing an innovative approach and presenting a novel method and metrics aimed at providing a nuanced assessment of the net energy efficiency on a case-by-case basis. This method takes into account the specificities of each working configuration, tailoring the analysis to the continually evolving landscape of SW, where new forms and practices continuously emerge. Importantly, in addressing a current research gap, the metrics not only offer a global evaluation but also provide a means to quantify potential energy improvements or worsening, considering perspectives from both employers and employees, thus contributing to a comprehensive understanding of the energy implications of SW.

2. Methodology

In order to address the research gap described in the introduction, the research work focused on two primary aspects: understanding the main factors influencing energy use affected by smart working (SW), and conducting a comparison between “before” and ”after” SW application configurations. The key methodologies employed in the research can be summarised as follows:

• Bibliographic research. An analysis of the state of the art of the current research works was carried out, examining not only papers concerning the environmental impact of SW, but also the existing indices related to such a topic. Using platforms like Google Scholar and Science Direct of Elsevier, it was possible to identify gaps in the existing literature, particularly the lack of evaluation tools and methods and the lack of an employees’ perspective on the energy advantages or disadvantages related to the SW application.
• Overview of the Italian Situation. To contextualize the research, an overview of the Italian situation was provided. This part, based on bibliographic research, includes a brief description of SW practices in Italy, which is particularly interesting due to the massive increase in the SW application during and after the pandemic period.
• Real Building Dynamic Analysis. As an important part of the research, a dynamic analysis of the energy consumption of a real building was conducted, considering different occupancy profiles and SW days to understand how SW can influence household energy consumption in a practical setting. The dynamic simulation is a novelty in this field and can be a means to extract, from the Energy Performance Certificate, energy consumption data nearer to real ones.
• Metric Development. Specific metrics to quantify the net environmental impact of SW to provide a measure of the energy efficiency in different configurations was developed. Four new indices were built, using a method that avoids the main drawbacks of such evaluating tools (i.e., arbitrary weights and judgements). These four new indices address the lack of evaluation tools found during the bibliographic research, providing an innovative method to measure the improvement or worsening percentage in energy saving. Moreover, these metrics distinguish perspectives between employers and employees, answering the open research question described in the introduction.

3. Smart Working Practices in Italy

In Italy, Law 81/2017 [14] introduced the concept of “agile” work. The aim of such a working practice, as outlined in Article 18 of the law, is to enhance the competitiveness and facilitate a better balance between work and life. “Agile” work is defined as a flexible way to perform work duties, based on objectives, without particular constraints on the place and time, except for the limit of maximum daily and weekly working hours. Upon analysis, it becomes evident that “agile” work is very similar, if not identical, to smart work. Since 2017, various experiences of hybrid working have been carried out; however, the COVID-19 pandemic was a significant game-changer. In 2019, according to the Confesercenti Dossier 2022 [15], Italian smart workers accounted for only 5.7% of the total workforce, whereas the percentage was higher in other European countries (23% in France, 12.3% in Germany, and 8.4% in Spain). In 2020, during the pandemic crisis, due to related restrictions, the percentage of Italian smart workers rose to 40%, surpassing France, Germany, and Spain. In 2022, the dossier estimated that around 4.5 million employees, constituting 20% of the total workforce, worked remotely in the post-lockdown period. Regarding transportation, the dossier estimated that 20% of remote workers avoided using public transport, while 80% refrained from using private transportation means, such as cars or motorcycles, resulting in a significant reduction in air pollution levels. The same survey reports that the number of professions “potentially workable remotely” is approximately 36% of the total employed, namely 8.2 million. An insightful survey of ENEA on home-to-work mobility, based on 5500 employees belonging to the public administration, revealed that the average commuting time was 1 h and 30 min (one in four people takes more than two hours). The study also highlighted that the average covered distance is 49 km per day (one in four individuals travels more than 70 km daily). Notably, the car is the primary mode of transportation, with 5 out of 10 people using it exclusively.

The survey carried out by the Smart Working Observatory Politecnico di Milano [16] contains interesting data. Such a survey is continuously updated and it describes the experience of companies operating in Italy, currently adopting smart working. The Smart Working Observatory, estimated, also, that working remotely for 2 days a week avoids the emissions of 480 kg of CO₂ per person per year.

Table 1. SW Italian examples (Ref. stays for references).

Company	Sector	Number of Employees	SW Applications	Estimated Impact	Ref.
A2A	Environment, Water, and Energy	13,000	From 20% to 60% of remote working time	Reduction of 2000 tons of CO2 per year	[16]
Kuehne + Nagel Italy	Logistics, Sea, Air, and Land Transport	3000	Up to 50% of remote working time	-
Sanofi	Pharmaceutical	2000	40–60% of remote working time every four weeks	Savings on transportation expenses for home-office commuting
Angelini Industries	Pharmaceutical, Perfumery, and Other Sectors	About 5800	2 days in office, 3 days remote	-
Mediaset	Television Broadcasting	3000	Averagely, 2 days per week remote; at least 1 day per week remote and 1 day in office	30% reduction in environmental impact due to home-to-office commuting
Autostrade per l’Italia	Highway Management and Maintenance	5300	At least 2 days per week remote	-
Nestlé	Food Sector	4600	In experimentation with SW since 2022
Baker Hughes	Industrial Technologies	5000	-	900 tons of CO2 reduction
Coop	Food Sector, Consumer Cooperatives	18,000	2–4 days per week remote	Reduction in commuting of about 830,000 km and associated travel time of about 17,500 h
Agos	Consumer Credit Sector	2000	20–50% of working time by remote. At least 1 day in office per week	-
DB Group	Shipping Sector	300	2 days per week remote	-
Kuwait Petroleum Italia (Q8)	Fuel Distribution Sector	600	2 days per week remote	-
Caffeina	Marketing/Communication Sector	200	Various remote work options (e.g., work from office, work from close by, work from anywhere, work from remote)
Chiesi	Pharmaceutical	1560	From 2 to 5 days a week remote	Reduction in CO2 emissions by 10%, including green commuting initiatives
Ducati Corse	Motorcycle	1300	80% of employees involved in SW
AON	Insurance	1600	Flexible hours and presence
Credit Agricole	Insurance	230	Alternation of three days in the office and two in remote working
Danone	Food	500	20% of the month in presence (1 day out of 5)
Pomellato	Jewelry Manufacturing	400	Partially remote tasks defined in three levels
Grenke	Services	220	Office presence at least once a week; Compliance with workplace safety and privacy rules
Siemens	Manufacturing	2700	SW applied without limitations	60% reduction in CO2 emissions following office redesign and management
Plenitude (Eni)	Energy	Not specified	12 days of remote working per month		[17]
Generali	Insurance	Not specified	At least two days in the office during the week, with the opportunity for remote work on Fridays, holidays, and in August.	Motivated by the extraordinary increase in energy prices.	[18]
Tim	Telecommunications	32,000	Up to three days of remote work per week; desk sharing in the office, space rationalization	Reduction of 50,100 tons of CO2 due to avoided commuting.	[19]
Vodafone	Telecommunications	6000	Three days of remote work and two in the office voluntarily each week. Possibility of remote work even during public transportation strikes and red alerts or holidays.	-	[20]

below summarises not only the findings of the Smart Working Observatory, but also some other important experiences of SW in Italy:

Analysing the above data, it is possible to identify an attention to CO₂ emission savings as well as energy cost savings, with a particular focus on commuting issues. It is important to note that the energy consumption related to transport accounts for the 25% of total energy consumption, as reported by Linstad et al. [21], and SW could be a means to reduce it and to significantly contribute to the emission mitigation targets, at least in the short period [11]. In their survey of 800 individuals across Italy, Fabiani et al. [22] calculated that beyond 10 km of distance, remote working gives a positive environmental impact, in terms of global warming potential. Also, Noussan et al. [23] underlined that such a working practice can be an effective action for energy consumption and emission reduction, in particular, allowing for a 16–17% of emissions reduction when long commuting distances are avoided. Notably, when remote working is applied, office space is often rationalized, leading to increased cost (and emissions) savings. Remote working appears to be encouraged before the weekend, to increase the energy savings of the companies. From the various experiences, reported above, hybrid working seems to be the preferred choice for the companies. In general, remote working is allowed for at least two days a week. During such days, employees can choose, in autonomy, their working locations, such as co-working spaces, libraries, green spaces, etc. Nevertheless, one of the most used places to perform the working tasks is one’s own dwelling. The increased energy consumption in “home offices”, already highlighted in the cited literature, could play a non-negligible role in the environmental balance of SW, making its estimation important. In [22], such a consumption was calculated on the basis of the construction year of houses, while other studies extracted it from surveys. In the current research, a dynamic simulation of a real building was carried out, considering different occupancy profiles to better estimate the potential impact of remote work on household thermal energy use. The analysis aims to discern when the increased consumption is closer to 7% or 23% [10]. In the following section, an example of such an analysis is presented.

4. Building Energy Consumption Evaluation in Different Occupancy Profiles

In order to carry out a reliable assessment of the potential impact of SW on the household energy consumption, it was decided to model a real building and conduct dynamic simulations, varying its occupation.

The chosen building is a condominium built in 1961, which can be considered a typical example of post-war architecture characterized by the extensive use of reinforced concrete, perforated bricks, and prefabricated panels, with little or no attention to energy losses. Situated near the centre of Genoa (north Italy), the condominium is part of a densely urbanized architectural context.

In the said city, reinforced concrete buildings with more than four floors make up 30% of the entire building stock, and the characteristics of the envelope materials and the geometry of the building are similar to other structures in the same area, making it potentially emblematic of the entire region (data elaboration from the 14th ISTAT Report of 2001) [24]. Thus, such a building can be a good candidate as an archetype to be used with the Urban Building Energy Models (UBEMs) [25] in the context of the research project Urban Reference Buildings for Energy Modelling (URBEM) [26]

Details about the actual state of the building, including a careful analysis of the layers and materials composing the external frames, the characteristics of the heating system and related actual energy consumption data, were extracted from [27].

The building layout is composed by a U-shaped part (main tower) and a rectangular part (secondary tower). The apartments, with an internal height of 3 m, have variable sizes based on their location. Two main modules are distinguished based on the floor plan dimensions: 90 m² for the two largest dwellings in the main tower and approximately 60 m² for all other units. The total usable area of the apartments is about 5602 m², with a ratio of exposed surface area-to-heated volume of 0.4.

In Figure 1, there is a plan of one of the first eight floors.

The external wall stratifications have been identified through on-site investigations. The perimeter walls are mainly composed of walls with a double layer of plastered brick, except for some load-bearing surfaces in reinforced concrete, characterized by two layers of perforated bricks with an air gap of 8 cm. The average transmittance of the external closures in bricks is equal to 1.105 W/(m²K), while the transmittance of the bearing parts in reinforced concrete is 3.48 W/(m²K).

The building’s roof is flat, insulated, and accessible. The southeast and southwest sides have large openings (doors–windows) and balconies, while the northeast and northwest sides have fewer windows and no balconies. The windows are made of double glazing and wooden frames, with a total transparent surface of 600 m²; the average transmittance of the glazed window components is 2.98 W/(m²K).

The building, that is one of the two twin constructions depicted in Figure 2, consists of 17 floors, with the first 8 comprising both the main and secondary towers, while the remaining 9 complete the main tower alone, as it can be seen also in the prospect reported by Franco et al. [28].

The heating system is centralized, with column distribution and radiators as emission terminals. Two natural gas generators with a total nominal power of 200 kW, regulated by a thermostat, are located in a thermal plant. The electrical power of the circulation pumps is about 800 W. The supply water temperature is maintained at 70 °C and regulated by a single climate control unit with an external sensor. The chimney losses with burners on have been evaluated at about 8%, while those with burners off at about 1.5%. The internal temperature of the apartments is maintained on average at 20 °C.

Based on the structural and plant data provided earlier, a building model was created in TerMus-PLUS (version 5.0.3.30718) [29], a software that uses EnergyPlus (version 22.1.0) for dynamic energy simulations through a graphical interface.

The actual consumption data, available for the late autumn–winter–early spring season in 2003–2004 were used to validate the model. A simulation was conducted using external climatic conditions obtained from a monitoring station located in the immediate vicinity of the building during the same time interval. The error between the simulation results and real consumptions is within 5%, confirming the reliability of the model.

In order to evaluate the energy impact of SW, the first simulation was carried out under the hypothesis that the heating is provided for 12 h a day, maintaining the same internal temperature. Such a configuration is named “Case a”.

The second simulation scenario, labelled “Case b”, considers the same building occupation, but increases the heating time to 16 h a day, prioritizing thermal comfort at the expense of increased energy consumption.

The last simulation configuration has the following assumptions: the building’s apartments are, in part, occupied by “Workers” who spend approximately 10 h away from home for five working days per week. During such days, they use heating for only 6 h per day (early morning and evening). Workers use heating throughout the available activation period, 12 h, on weekends and holidays. The other building occupants are “Residents”, who predominantly stay at home and, thus, use heating continuously, 12 h a day, each day of the week. To manage the system’s usage during the daily heating period, the availability of temperature regulation systems for each apartment is assumed.

To estimate the number of inhabitants in these two categories, it was considered that Residents are mostly individuals aged 65 or older and children aged 3 years or younger. Based on national demographic statistics as of 2023 [30], the population aged 65 and over constitutes approximately 24% of the total, while children up to 3 years old make up 2.7% of the population. If considering pensioners, it should be noted that the retirement age of 67 has varied over time, with different options for early retirement. In 2021, the percentage of pensioners was 38.5% of the total residents.

In order to give a good approximation of the real occupation situation, “Case c” represents the scenario where 26% of the building inhabitants are typically at home, while the remaining 76% goes out to work. From this perspective, the two previous cases, a and b, consider the same occupation of the building such that all the inhabitants are Residents.

For all the above situations, two different internal temperature settings were considered: 22 °C and 20 °C. The latter is in compliance with collected data concerning actual energy consumption, while the former represents a condition of increased thermal comfort, using a tolerance of ±2 °C, for the internal temperature, allowed by the national law, Decree 412/1993 [31] and Decree 74/2013 [32].

In Figure 3 results, in terms of yearly natural gas consumption, are reported for all the scenarios and temperature settings. The percentages above the arrows represent the increment in the energy source consumption between two cases: the first one is individuated by the arrow start, while the second one is showed by the arrow end.

The comparison between Case c and Case a represents the difference between an occupation with all the Workers out for work 5 days a week, while all Residents remain at home, and a situation of full remote working (5 days a week), that means all at home. The increment in both temperature scenarios, 20 °C and 22 °C, fall approximately in the middle of the range found by [10], ranging from 7% to 23% as previously mentioned. Increasing the heating duration from 12 to 16 h results in an additional consumption increase of 13.7% or 13.6%. Comparing Case c and Case b means that not only do all the inhabitants stay at home, but that all of them, without distinction, decide to increase the maximum heating period by 4 h. Under such a hypothesis, the consumption increment exceeds the superior limit of the cited range because in both situations of temperature setting, the increment is about 30%. Nevertheless, this last configuration seems non-representative of the actual trend in the country, where, even in the coldest regions, the heating period is averagely maintained under the 12 h [33].

On the basis of the previous observations, Case c can be considered the baseline for further considerations about possible energy consumption increments due to work carried out remotely. Moreover, given the slight difference in increment for the two temperature settings, the configuration of 20 °C, aligning more with the current trend of energy savings recommendations, was used. Thus, taking into account Case c for the rate between Workers and Residents as well as their heating trend, other simulations were carried out. These simulations considered varying building occupation and heating hours in accordance with the increasing days of smart working carried out remotely, ranging from 1 to 5 days a week. Results are presented in Figure 4 as a graph.

Analysing the graph above, it can be seen that the consumption trend is linear. Such a trend expresses how the energy consumption, E_r, increases with the days of remote working, r. The general expression can be written as

$$E_r=k r+E_0$$

(1)

Equation (1) permits for the calculation of building consumption E_r, given the r days of remote working, knowing the building consumption before the SW application E₀, when r is equal to 0, and the coefficient k. It is worth observing that E can be expressed in different measure units, depending upon the employed energy source; in the case study example, it is expressed in Sm³/year. The k coefficient depends upon the considered building archetype and has, as measure units, the same measure unit of E₀ divided by day. In the examined case for this kind of building, k is equal to 1343 Sm³/day.

Analyses carried out on the building proceeded to set two important points. First of all, dynamic simulations confirmed international findings about the possible energy consumption increment resulting from a different dwelling occupation, and the value is not negligible. Secondly, simulation results highlighted a linear relationship between the number of remote working days and the observed increment in the case study, representative of a large percentage of Italian dwellings. Therefore, understanding the impact on household management for each SW configuration is crucial to accurately assess the environmental advantages of such a working practice.

5. Smart Working Environmental Impact Evaluation: Method and Metrics

The previous section pointed out the possibility of a meaningful increase in household thermal energy consumption, emphasizing the need for a punctual evaluation of potential environmental benefits of smart working. The current section introduces a method that allows users, with the help of numeric metrics, to determine the extent of actual energy savings and the possible burden shifted to employees, aiming to propose suitable compensation. While SW presented promising results for decarbonisation, its widespread adoption during the pandemic raised important questions about the actual net impact of the practice. To address the green deal, synergic actions are required. Effective laws to enhance good SW practices and correct evaluation of their application within companies are necessary. The goal is to achieve “win-win-win” scenarios benefiting companies, employees, and the environment. However, formulating effective laws and providing accurate assessments require an evaluation tool, leading to the development of a method to calculate the energy effectiveness of SW. The proposed metrics and method are aimed to serve as tools for policymakers, businesses, and stakeholders interested in promoting energy-efficient smart working practices. A metric suitable for such purposes should be characterized by clarity, comprehensibility, conciseness, ease of use, and effectiveness in providing a simple tool to support comparisons and evaluations. Authors oriented their research to find numeric indices, which were able to give, at a glance, a sort of energy effectiveness related to the SW policies. It is worth remembering that indicators and indices are a favourite means to translate complex phenomena into simple metrics, even if an index field is somewhat controversial, because it often implies arbitral choices, as underlined by Greco et al. [34]. To develop an indicator and/or an index is not a simple task, because it requires not only finding and quantifying the main parameters ruling the complex phenomena, but also to parametrize and combine them in order to obtain a synthetic, but meaningful, metric. Nevertheless, the European Community [35] strongly recommends such a practice, in particular, for sustainability and environmental themes. As far as the authors’ knowledge goes, the only existing index, peculiarly thought for SW, concerns the propensity to such a working format, developed by Astorquiza-Bustos et al. [36]. In [22], instead, the Global Warming Potential (GWP) index was used to assess the avoided emissions given by SW practice. Such an index is valuable in this context as it allows users to assess the potential emission impact of SW. However, it is not a metric for determining energy efficiency, nor a means to understand if the practice provides savings for the main parties involved in the change in working paradigm, namely the company and employees.

The novelty of the current research is the proposal of a set of four indices based on the energy efficiency concept given by the intended savings, avoiding, thus, arbitral weight coefficients or subjective judgments. In such a way, one of the main drawbacks related to the index building is overcome [34]. The proposed metrics not only enable users to determine the overall advantage of SW from an energy-saving perspective, but also assess whether such a practice benefits both employers and employees. Furthermore, the research introduces a method, detailed below, that describes how to collect and organize data for these indices.

The main energy factors taken into account in this first proposal are as follows:

• Energy savings related to the reduced use of HVAC office plants;
• Energy increments due to the increased house use;
• Energy savings related to the decreased workers’ commuting;
• Company electrical energy savings due to the avoided use of computers and other required equipment;
• Employees’ increased electrical energy use due to computers and other required equipment.

Evaluating other emission savings, such as avoiding plastic bottles, reducing waste production from office cleaning, or managing water consumption, requires further analysis. This aspect would be developed in a later stage of the research and can be seamlessly incorporated into the considered aspects using the same criteria.

The method is based on the fruitful collaboration between employers and employees in order to obtain clarity and transparency in results. In recognition of the importance of privacy and adherence to local privacy laws, participants should be explicitly informed about the nature of the data collected, the purposes for which it will be used, and the safeguards in place to protect their privacy. Ensuring transparency and providing individuals with the opportunity to grant explicit consent not only aligns with legal requirements but also contributes to create awareness about the sustainability aim of the work.

The following outlines five steps that encapsulate a synthesis of the evaluations needed to implement the proposed methodology. It starts with the calculation of potential companies’ energy savings (Step 1), followed by the gathering of employee data, particularly focusing on transportation means and daily covered distance (Step 2), and additional details such as the location of remote working activity, whether it be at home or in a co-working space (Step 3). In the case of remote work from home, the suggestion is to utilize the building Energy Performance Certificate (EPC) and potentially normalize consumption with reference to building archetypes.

Subsequently (Step4), the data facilitating the quantification of various aspects are distilled into indices:

1. Smart Working Energy Efficiency Tool (SWEET):

Definition: SWEET is an overall efficiency measure that considers various energy consumption factors before and after the implementation of smart working.

Calculation: it is calculated as the ratio between the estimated global energy savings due to SW and the global energy consumption before SW.

2. Smartworkers’ Energy Efficiency (SEE):

Definition: SEE offers an employee-centric perspective, calculating the energy advantage or disadvantage for all individuals engaged in smart working.

Calculation: it is determined by the ratio of the energy consumptions and savings before and after SW application, particularly focusing on the possible increased energy use of the buildings.

3. Specific Smartworker’s Energy Efficiency (SSEE):

Definition: SSEE provides a more granular evaluation at the individual employee level, offering insights into the energy impact for each person engaged in smart working.

Calculation: the tool has the same structure as SEE, but considers the specific energy advantage or disadvantage for each employee.

4. Employer’s Energy Efficiency Evaluation (4E):

Definition: 4E calculates energy efficiency from the company’s perspective in the smart working configuration.

Calculation: it measures the energy efficiency by considering office and equipment-related energy consumption before and after smart working.

Finally (Step 5), the benefits for the company and the employee are assessed, and any compensation measures for the employee are contemplated if they are found to be penalized.

5.1. Step 1

The first step consists of the assessment of potential company energy savings. This operation should be based on an energy audit or on data coming from energy management. Using real energy consumption data, related to office and equipment use, and considering possible savings from their optimization including space reduction, it is possible to determine the factors mentioned in points 1, 4, and 5. As a first approximation, the energy required for working equipment, such as computers, web cam, etc., is assumed to be the same for both office and remote activities.

Real consumption data should be collected, at least, on a yearly basis and, for each j-th type or source, the “before” SW configuration (b) should be provided:

$${{Eo}_{jb}=\left({\sum }_{i=1}^n{eo}_{ij}\right)}_b$$

(2)

The term eo_ij refers to the i-th quantity of energy, of the j-th energy type or source, consumed in the offices for HVAC, such as electrical energy for heat pumps, air conditioning, or natural gas for office heating and domestic water heating, and so on. Clearly, Eo_jb represents all the n consumptions of the same j-th energy type or source. All of the Eo_j are expressed into their typical measurement units. In order to obtain homogenous terms, it is necessary to convert them to Primary Energy (PE). PE is a diffused way to transform the different types and sources of energy into a common term, whose measurement unit can be expressed in “kWh”, as remembered by Hitchin et al. [37]. In Europe, it is mandatory to use PE for the energy performance of buildings, as reported in the directive 2010/31/EU [38]. In order to transform all the energy terms into primary energy, it is possible to use conversion coefficients f_j, such as those specified in the standard EN ISO 5200-1 standard [39].

Applying those coefficients for all the m energy sources or types, it is possible to express the entire HVAC office-related energy consumption, PEo_b, in the “before” SW condition:

$${PEo}_b={\left({\sum }_{j=1}^m{Eo}_j·f_j\right)}_b$$

(3)

Similarly, on the basis of the expected energy savings, it is possible to calculate the PEo_a, which is the entire office-related energy consumption in the “after” SW condition (a):

$${PEo}_a={\left({\sum }_{j=1}^m{Eo}_j·f_j\right)}_a$$

(4)

At this stage of the method, it is also important to determine the energy consumption related to the individual use of office equipment, including computers, web cams, etc. Taking into account the energy used by each employee’s workstation, x, the PE related to the said office equipment consumption, PEeq_xb, in the “before” configuration, can be expressed as follows:

$${PEeq}_{xb}={\left({\sum }_{j=1}^m{Eeq}_j·f_j\right)}_b$$

(5)

To obtain the total energy consumption related to office equipment, PEeq_b, simply multiply the above equation by the number, q, of the employees using

$${PEeq}_b={\sum }_{x=1}^q{PEeq}_{xb}$$

(6)

The after configuration can be calculated, for each employee, PEeq_xa, dividing PEeq_xb by the “before” number of working days, w, carried out in presence and then multiplied by the days in presence, pw, in the SW configuration:

$${PEeq}_{xa}={pw}_x ·\frac{{PEeq}_{xb}}{w_x}$$

(7)

The total energy consumption related to office equipment, PEeq_a, in the after configuration, can be calculated as

$${PEeq}_a={\sum }_{x=1}^q{pw}_x \frac{{PEeq}_{xb}}{w_x}$$

(8)

5.2. Step 2

The second operation requires employees’ data, in particular the transport mean and the daily covered distance. The latter could be seen as a redundant information, considering that employers already have the address of their employees. However, the “house–work” journey could differ from the shortest distance due to traffic conditions. Sometimes, the longest route is the fastest and, therefore, the chosen one. On the basis of the transport mode and of the declared journey, it is possible to determine the primary energy consumption related to commuting for each employee, PEc_xb, as follows:

$${{PEc}_{xb}=\left({\sum }_{i=1}^t{d}_i·{Ec}_i·f_i\right)}_b$$

(9)

The distance, d_i [km] covered by each of the t used transport means, is multiplied for the related specific energy consumption, Ec_i [40].

Taking into account the q employees, it is possible to determine the total energy consumption:

$${PEc}_b={\sum }_{x=1}^q{PEc}_{xb}$$

(10)

Similarly to Equation (8), to quantify the energy consumption for employee in the “after” configuration, it is possible to apply the following:

$${PEc}_{xa}={pw}_x ·\frac{{PEc}_{xb}}{w_x}$$

(11)

Thus, the total energy consumed for commuting in SW configuration is

$${PEc}_a={\sum }_{x=1}^q{PEc}_{xa}$$

(12)

5.3. Step 3

The third step requires other data from employees. Firstly, it should be ascertained where the remote working activity is taking place. If a public space is used, such as a library or an internet café, in this first approximation, no increased energy consumption in charge to the employees will be considered. If from home, or from a co-working space, when the diagnosis or a collection of real data consumption is available, it is possible to proceed in the same way of step 1 and determine, for each j-th type or source of energy, the “before” consumption for each employee:

$${{Er}_{xjb}=\left({\sum }_{i=1}^n{er}_{xji}\right)}_b$$

(13)

The term er_xji refers to the i-th quantity of energy, of the j-th type or source, consumed, in the chosen remote working location, for HVAC and equipment, by the x-th employee, in the “before” configuration. If a co-working space is considered, the above result should be divided by the number of workers, g, sharing the same office.

Applying the conversion coefficients f_j, it is possible to express the dwelling-related energy consumption, PEr_xb, for each employee in the “before” SW condition:

$${PEr}_{xb}={\left({\sum }_{j=1}^m{Er}_{xj}·f_j\right)}_b$$

(14)

Taking into account all the q smart workers, the entire dwelling-related energy consumption, PEr_b, is

$${PEr}_b={\sum }_{x=1}^q{PEeq}_{xb}$$

(15)

Referring to the case study, whose results are in general agreement with the other international findings, as already underlined, it is possible to give a first energy increment coefficient, k_x, which provides the increment of energy consumption due to remote working days, proportional to that defined in (1). It is worth mentioning that the k value is not affected by the conversion coefficients f_j, because it is the angular coefficient of the consumption trend, which does not depend upon the energy source. Thus, the energy consumption in the “after” SW configuration, for the x-th employee will be

$${PEr}_{xa}=r_x·k_x+{PEr}_{xb}$$

(16)

and

$$\left(W_x-{pW}_x\right)=r_x$$

(17)

The coefficient k_x, of the x-th employee’s dwelling, is proportional to k (1) by the ratio between PEr_xb and E₀ of the entire building, while W and pW represent respectively the weekly working days and the weekly working days in presence. Taking into account the q employees, it is possible to determine the entire energy consumption in the “after” configuration:

$${PEr}_a={\sum }_{x=1}^q{PEr}_{xa}$$

(18)

At this point, it is important to add an observation. It is difficult for private houses to undergo energy diagnosis, and obtaining reliable consumption data is not straightforward. Moreover, there is a potential bias connected to self-declared consumptions related to the willingness to benefit from SW. To address these issues, a sort of standard should be employed. The most commonly used document declaring the building energy consumption in such cases is the Energy Performance Certificate (EPC) [41]. This document allows users to extrapolate HVAC energy consumption in terms of primary energy, offering a potential solution to the data collection challenge. However, it is crucial to note that there is often a significant gap between real energy consumption and the calculated values provided by the EPC. Delghust et al. [42] found an average of 25% difference, between the real gas consumption and the calculated one for low-energy houses, on a conspicuous sample of 537 dwellings, underscoring the tendency of the EPC to overestimate the data. Van Hove et al. [43] highlighted the impact of higher overestimations on lower-class households. In compliance to this last observation, the identified gap for the dwelling analysed in the previous section (which is a poor insulated building) is approximately 35%. However, the EPC certificate might be the only official and available data on the building energy behaviour. Thus, it would be advisable to find a way to employ it in order to estimate, at least, the consumption magnitude order. For example, starting from the EPC results, including primary energy (PE) required for HVAC, and using data on building characteristics, the correspondence with a building archetype can be identified. Archetype databases such as those reported in [25], or the one in construction in the cited project URBEM can be employed for this purpose. These archetypes can then be used in dynamic simulations, akin to those conducted in Section 4, to determine more realistic energy consumptions and adapt the k coefficient, if necessary.

5.4. Step 4

Once the above pieces of information have been collected, the authors propose a set of indices to assess the energy efficiency of SW practices. The first metric, called Smart Working Energy Efficiency Tool (SWEET), is an overall efficiency measure that considers all the mentioned energy consumption factors in both the SW configurations, before and after. Its expression is

$$\mathrm{S}\mathrm{W}\mathrm{E}\mathrm{E}\mathrm{T}=\frac{{PEo}_b+{PEeq}_b+{PEc}_b+{PEr}_b-{PEo}_a-{PEeq}_a-{PEc}_a-{PEr}_a}{{PEo}_b+{PEeq}_b+{PEc}_b+{PEr}_b}$$

(19)

The SWEET index is the ratio between the estimated energy savings, assessed by comparing the consumptions before and after the introduction of SW, and the energy consumption before SW. The metric enables users to assess the environmental advantages (or disadvantages) of this practice.

Such an index can take on positive, negative, or equal to zero values. A positive value indicates a favourable environmental impact of SW, signifying a net energy saving. In this case, SWEET can reach a maximum of one, representing a scenario where all primary energy consumption, once SW is implemented, equals zero. A result of zero implies that SW has no discernible effects on energy savings. Conversely, a negative value indicates that SW leads to an increase in overall energy consumption.

The second proposed index, called Smartworkers’ Energy Efficiency (SEE), offers an employee-centric perspective. It calculates the energy advantage, or disadvantage, for all the persons engaged in SW:

$$\mathrm{S}\mathrm{E}\mathrm{E}=\frac{{PEc}_b+{PEr}_b-{PEeq}_a-{PEc}_a-{PEr}_a}{{PEc}_b+{PEr}_b}$$

(20)

The SEE index can be positive, negative, or equal to zero too, with the same meanings as the SWEET results.

A specific version could be very useful for detailing the results at the scale of individual employees. Thus, the following equation proposes the Specific Smartworker’s Energy Efficiency (SSEE) index:

$$\mathrm{S}\mathrm{S}\mathrm{E}\mathrm{E}=\frac{{PEc}_{xb}+{PEr}_{xb}-{PEeq}_{xa}-{PEc}_{xa}-{PEr}_{xa}}{{PEc}_{xb}+{PEr}_{xb}}$$

(21)

The final proposed index, called the Employer’s Energy Efficiency Evaluation (4E), represents the calculation of energy efficiency from the company’s perspective in SW configuration:

$$4\mathrm{E}=\frac{{PEo}_b+{PEeq}_b-{PEo}_a-{PEeq}_a}{{PEo}_b+{PEeq}_b}$$

(22)

In this scenario, the index is generally expected to take on values between zero and one. While a negative result is theoretically possible, it implies that the SW configuration leads to higher energy consumption compared to working in a physical presence. Such a case can be considered highly unusual and might be the consequence of some errors in data collection, or in the definition of the number r of remote working days.

5.5. Step 5

Once the values of the aforesaid indices are obtained, it is possible to determine whether SW is advantageous, from an energy point of view, and who is advantaged or disadvantaged. In the flux diagram proposed below in Figure 5, there is a possible way to use the indices results.

Analysing the above flow chart, it is clear that once SWEET is determined to be positive, indicating a global energy efficiency improvement due to the SW measure, it is essential to examine the other indices.

If SEE is positive or at least equal to zero, it signifies that, in general, the worker community does not experience energy drawbacks due to the altered working conditions. Conversely, if the result is negative, it implies that the overall energy efficiency is partially achieved at the expense of the employees.

The comparison between SWEET and 4E allows an understanding of how much the energy gain is due to the burden shifting. In particular, when the difference between SWEET and 4E is positive or null, the energy savings in the offices are not counterbalanced by higher energy expenses of employees. Conversely, if it is negative, the result can be seen as an indication of the energy burden shifted to the workers. At this point, potential actions can be considered to mitigate the imbalance. For example, determine an acceptable percentage of the company gain to be converted to an economic bonus to be distributed to employees. Such a percentage should be equal, at least, to the calculated gap between SWEET and 4E.

Once such an operation is carried out, a more detailed analysis, based on SSEE, can be helpful. The said index, calculated for each employee, makes it possible to determine who is paying the highest energy price and to act in a focused way to obtain equality, for example, by providing part of the economic bonus found in the previous step. It is important to emphasize that SSEE results are significant even when SWEET and SEE are positive. This is because there may be cases where some employees are disadvantaged by SW while others are not. In such cases, it is important to discuss with affected employees the option of undergoing SW or, alternatively, to identify compensatory measures within the company’s overall gain margin resulting from the application of SW. It is possible that some employees may prefer SW despite its negative energy impact for them due to other advantages. They may choose to accept this working arrangement without expecting any compensatory bonus. In such instances, if SWEET is positive, the overall energy efficiency improvement associated with the working configuration remains positive.

6. Future Developments

The proposed method and metrics are a means to understand the true impact of SW on the company’s energy-saving strategies, aiming to quantify its contribution to energy sustainability. The next step in the current research involves applying these indices and methods to a real case study—a company characterized by an extensive use of SW practice aimed to reduce its environmental impact—to validate results and define potential equity strategies. This analysis will include an economic assessment.

Another potential step of future development concerns the expression of SWEET. As explained in the previous section, the index enables users to calculate a global energy efficiency value related to SW. In the proposed initial expression, SWEET takes into account some of the energy-cost voices, internationally recognised as the main factors involved by the application of SW. However, it is possible to broaden the equation to encompass additional potential savings, such as the energy required for office cleaning and the reduction in waste materials associated with in-person work (paper, toner, etc.).

Another aspect of the research will focus on building analysis. In addition to defining archetypes for climatic zones, the objective is to identify corresponding k values through dynamic simulations. These simulations, similar to the approach proposed in the current work, will be conducted on different archetypes. Simultaneously, a study on the percentage gap between the models and EPC results will be developed. The aim is to derive a set of coefficients capable of converting the procedural results into acceptable energy consumption data, ready for use in the determination of the proposed indices. A last interesting possible development concerns the Energy Communities. The European Commission defines them as energy initiatives led by citizens that contribute to the transition to clean energy, promoting enhanced energy efficiency within local communities [44]. In particular, in Italy, there are plans to support such communities through incentives on produced and self-consumed renewable energy. In this context, a diffused issue concerns the electricity generated by Photovoltaic Panels, whose production peak occurs during the central hours of the day, corresponding to a significant portion of typical office time. In a SW configuration, such an energy production could be immediately employed, optimizing the use of renewable resources and economically benefiting the community.

7. Conclusions

Smart working has the potential to be a game-changer in the new green deal, especially for achieving short-term goals. However, its application requires careful evaluation and monitoring to understand its actual benefits. It is of the utmost importance not to underestimate the risk of shifting the burden from companies to employees in terms of energy consumption.

In this study, we conducted a dynamic analysis based on a real residential building, representative of an architecture solution diffused in the northern part of Italy, in various occupation configurations. The results confirmed that remote working can impact employees’ energy consumption, leading to an average 15% increase when all work activities are carried out from home. Importantly, the study emphasizes the need to analyse each case individually to precisely determine the advantages or disadvantages of SW, also taking into account the role of the company.

To address these concerns, the research proposes a new method and introduces novel evaluation metrics specifically designed to assess the environmental advantages of smart working. Four indices, based on the concept of energy efficiency evaluated on savings and expenses, have been developed:

• SWEET: offers a global evaluation of SW application, considering both company and employee energy consumption data.
• SEE: provides a metric to assess the energy savings of employees as a whole.
• SSEE: calculates the energy situation for individual employees.
• 4E: provides the energy-saving perspective for the company.

These indices take into account the internationally recognized factors influencing energy consumption affected by SW, namely office plants and equipment, home or co-working plants and equipment, and commuting.

Additionally, a method for data collection and utilization of these indices is presented to enable users to obtain quantitative answers and initial indications to counteract potential inequities resulting from a burden shift to employees. The method and the indices can be generally applied wherever there is an interest in understanding whether SW is efficient from an energy perspective and whether employees are affected by a net additional expenditure on energy.

This contribution not only represents an innovative approach in the field but also addresses a previously unanswered research question regarding a means to assess the real energy efficiency of SW applications for both companies and employees.