Advancing the smart city objectives of electric demand management and new services to residents by home automation—learnings from a case

A smart sustainable city is an innovative city that uses information and communication technologies and other means to improve quality of life and efficiency (ITU, 2014). Smart city programmes guide the transformations of the respective cities in different policy areas, such as the use of renewable energy and reduction of greenhouse gas emissions (Ojo et al., 2015), and the significance of developing energy management and demand response solutions has been recognised (Miceli, 2013). As urban energy usage is largely concentrated in buildings, many smart city programmes explore reducing greenhouse gas emissions by increasing the intelligence of infrastructure and buildings (European Commission, 2021). In residential blocks, the majority of energy is consumed in individual homes. The claimed benefits of home automation systems include improved awareness of energy usage, reduction of energy consumption (KNX Association, 2019), and their potential to interconnect home energy management with smart grids (Ford et al., 2017; Hui et al., 2017). It can be logically concluded that in a smart city attempting to meet its energy efficiency and carbon neutrality targets, homes should also be smart.

In the autumn of 2013, the city of Helsinki launched the Smart Kalasatama project. It is an umbrella under which various kinds of smart city functions are being tested. These include energy solutions that aim to save energy and support the increase of wind and solar power in the grid. The city aims to make the power grid of Kalasatama as energy-efficient and flexible as possible by optimally utilising building automation in commercial buildings and home automation in apartments (Matschoss et al., 2016). However, making these objectives a reality requires that the city uses its regulatory power. Building automation has a centric role in optimising energy flows and reducing the unnecessary use of energy in commercial buildings, but automation is rarely seen inside residential apartments. No nationwide regulation exists that would require the installation of home automation during the construction phase, nor are there any norms or best practices of home automation functions and/or interfaces when home automation is present. Apartments are most commonly built by private companies, who determine apartments’ features based on market demand.

Under these conditions, the city of Helsinki decided to follow a home automation standard, drafted earlier by a business consortium for apartments in Kalasatama (Matschoss & Heiskanen, 2018). The standard describes a set of home automation devices, which was expected to provide readiness for energy-related services and electric demand management. Products or services for electric demand management in residential apartments were not existing during the compilation of the home automation standard, but the city expected that such would become available for future Kalasatama residents thanks to the readiness provided by the standard. The standard only covers readiness in the form of hardware, i.e. relays, meters, and communication capable controllers; the eventual utilisation of the home automation for energy-saving and demand management and the business models that the future services will employ have been left open. To ensure that the home automation systems in Kalasatama would create an attractive number of potential customers for companies offering energy-related applications and services, the city has made it mandatory to follow the standard in every apartment built in the district.

The focus of this paper is on the home automation standard of Kalasatama from an industrial perspective. Thus, the research question is as follows: How effectual are the home automation requirements (specified by the city) in leading to the emergence of new energy-related services for the city’s residents and integration of the apartments’ home automation systems into the city’s smart grid via demand management? In this study, we will outline the carbon neutrality targets of the city of Helsinki, explain the process that led to the designation of the new city district Kalasatama as a smart city project, describe the regulations the city has prescribed within this specific area to enable the emergence of energy management services for its residents, and describe the stakeholder interviews. Our findings suggest that the home automation in the area is in smaller use than was envisioned.

This paper is organised as follows: “Kalasatama—a smart city” describes the smart city of Kalasatama and its regulations related to home automation. “Home automation in demand management” evaluates the applicability of that home automation in demand management. In “The implementation of the study”, we empirically examine the market reaction to Kalasatama’s home automation through stakeholder interviews. “Results” presents our conclusions.

In this section, we evaluate the applicability of Kalasatama’s home automation in the electric demand management applications available in the region.

Electric demand management actions may be categorised depending on their timing and impact. The quicker changes are processed and completed, the more unwanted impact they potentially have. Spinning reserves, implemented by loads, represent the upper (quicker) end. Loads can correlate their power consumption to a signal or the grid frequency. In the simplest implementation of this approach, consumers use less power if the frequency drops, and they are then rewarded with an incentive (Palensky, 2011; US Department of Energy, 2006). In Finland, the transmission system operator and a state-regulated monopoly Fingrid is responsible for the incentive-based balancing of energy markets. Fingrid purchases frequency containment reserves for normal operation (FCR-Ns) and frequency containment reserves for disturbances (FCR-Ds) from the domestic yearly and hourly markets. The FCR-N is a symmetrical reserve product. The reserve capacity is completely activated as upward balancing (implying that the power of the connected loads is minimized) when the grid frequency is 49.9 Hz or less. Correspondingly, when the grid frequency is 50.1 Hz or more, the reserve capacity is completely activated as downward balancing (the power of the connected loads is maximized). In the grid frequency range of 49.9 to 50.1 Hz, the volume of the activated capacity is proportional to the magnitude of the frequency deviation. With the FCR-Ds, the activation of the reserve capacity begins when the grid frequency falls below 49.9 Hz, and the reserve capacity is completely activated when the grid frequency is 49.5 Hz or less. The minimum reserve capacities needed to participate in the FCR-N and FCR-D markets are 100 kW and 1 MW, respectively (Fingid 2019).

The electricity consumption of large industries, such as the forestry and the metal and chemical industries, has long been used as a balancing reserve. The balancing markets have lately also become available for smaller-scale customers. Meeting the minimum power requirements of the balancing reserve market may require aggregators, who combine small-scale consumption from different consumers into a larger entity (Fingrid, 2019). Recently, certain groups of buildings, such as a chain of 136 groceries (Siemens, 2020) and 17 public buildings belonging to the city of Lappeenranta (Energy Efficiency Agreements 2020), have been aggregated for electric demand management. A large shopping centre, exceeding the minimum power requirement from a demand management contract partner, has been connected to the balance reserve market without an aggregator (Janhunen et al., 2020).

The lower (slower) end of demand management actions is shifting the time of use of energy, according to its time-based price. Tariffs penalise certain periods of time with a higher price. Customers can reduce their energy costs by adjusting the timing of their consumption, by consuming more during lower-priced periods and less during higher-priced periods (Palensky, 2011; US Department of Energy, 2006). There are two types of contracts available to buyers of electricity. In fixed-price contracts, the price of energy is known in advance, and in market-price contracts, the pricing is tied to the hourly changing market price (Energy Authority, 2020). The latter allows consumers to save on their energy costs by shifting their time of use.

Hypothetically, a large group of residential consumers could provide a considerable demand management potential, but not all electrical loads in residential apartments are shiftable. Non-shiftable loads, such as lighting and cooking, must be available at times desired by the users. Examples of shiftable loads are washing machines and dishwashers (Haider et al., 2016). Shiftable loads may be further categorised into deferrable and thermostatically controlled loads (Elghitani et al. 2018). Deferrable loads can be controlled to defer their energy consumption to a future time, but within a certain deadline to avoid inconvenience to the user. Most of the electric power in homes that is suitable for demand management comes from deferrable loads (Favuzza et al., 2018), but it is not possible to implement time shifting of common appliances solely via power cut-off with home automation. The research indicating that demand management has the potential to reduce peak energy usage in smart homes is commonly based on a presupposition that individual home appliances are intelligent and communicate externally (Li et al., 2011; Mohsenian-Rad et al., 2010). Thus, making demand management processes fully automated would require one to replace all the appliances in a home, as they should be smart and connected (Barsanti et al., 2020). Another challenge comes from smart home administration issues. Users of smart home systems are ordinary people of various ages and backgrounds, and no technical expertise can be expected, which may constrain the level of functionality the system may realistically provide (Balta-Ozkan et al., 2013). The willingness of consumers to change their energy use practices due to demand management is also expected to be limited (Haider et al., 2016; Yan et al., 2018). Investing in smart home technology solely for energy management is less appealing to the public than policymakers presume (Balta-Ozkan et al., 2014; Wilson et al., 2017), and interest in such investments is also low among renters who may not pay the energy bills themselves (Solà et al., 2021).

Next, we will assess the viability of the load groups (described in the “The home automation specified by the land transfer conditions”) for electric demand management. Here, each load group defined in the land transfer conditions (City of Helsinki 2016) has been assessed, assuming that they can be controlled on and off by a home automation system’s relay, as specified in the land transfer conditions. With that relay, the home automation may connect or disconnect the load groups from the main power, but whether the load group consumes power also depends on the condition of its local control (the room switch of a light group, thermostat of a thermostatically controller load, etc.). The column ‘typically existing’ states whether the load group typically exists in apartments at Kalasatama. In the column ‘upward balancing’, we estimate whether that group is eligible to participate in the incentive-based remand response by switching its relay open (which would result in reduced power consumption) without sacrificing user comfort or safety. In the column ‘downward balancing’, we estimate whether it is possible to increase power consumption by closing the relay. In the column ‘shift in time’, we estimate whether it is feasible to move the consumption of the load group in time by opening the relay during high-cost periods and closing it during low-cost periods (Table 1). The major difference between the Table 1 and the other categorisations of domestic loads found in the literature (e.g. Croce et al., 2020) is that we have followed the load groupings and their control principles as outlined by the land transfer conditions of Kalasatama, and have assumed that the existing household appliances are generic, i.e. not having hypothetical smart features. The assessment indicates that it is difficult to find a coupling between the load groups of a typical apartment and demand management. The most obvious load groups for demand management would be heat reservoirs, such as electric space or water heaters, but such are not existing in apartments with district heating.

This qualitative case study was conducted as follows: first, publicly available information about the smart city of Kalasatama was collected. This included announcements and publications by the city of Helsinki, the land transfer conditions and associated documentation, publications by companies that had either participated in the start-up phase of the smart city programme or later during its implementation, and other publications concerning Kalasatama. We also collected additional information through stakeholder interviews. We wanted to compare the objectives to the implementation of home automation—enabling new business and services for residents and enabling electric demand management—to how companies actually utilise it in Kalasatama, so we planned to seek interviewees from three groups of stakeholders.

The first group (‘initiators’) consisted of key persons in organisations that participated in the initiation phase of the smart city programme and took part in compiling the land transfer conditions. We interviewed a sales manager from industry and a project director, a head of unit, and a R&D manager from organisations of the city of Helsinki. The second group (‘implementers’) consisted of companies that are required to follow the land transfer conditions in their activities: two consultants from planning offices and a project manager, a planning manager, and a vice president from constructor companies. The third group (‘utilisors’) consisted of companies that are delivering solutions and services that are based on the home automation systems described in and required by the land transfer conditions. Only one such company could be recognised, whose business unit director was interviewed.

Potential interviewees were found by contacting known companies and organisations connected to the research topic and by using snowball method during the interviews. Interviews were conducted in person between January and September 2019. The average duration of an interview was 40 min. The interviews followed a semi-structured approach, where the topics of the interview were first given by the interviewer and then they were openly discussed. The interview topics covered how the interviewees, representing their organisations, assessed the usefulness, feasibility, and consequences of the home automation section of the Kalasatama land transfer conditions. The interviews were recorded, when approved by the interviewee, and afterwards a transcript of the interview was sent to the interviewees for a review.

We made a conceptual content analysis by reading the transcribed interviews and listing the arguments in which the interviewees were interpreted to express a position about the research question (Table 2). For groups 1 and 2, it was found that the views of all interviewees within the groups were largely in line with one another. Arguments that supported the home automation section of the land transfer conditions were almost solely presented by the group 1 interviewees. All group 1 interviewees stated that a home automation standard would not arise by itself. Thus, if a district-sized standard for residential demand management piloting is to be created, some regulation has to be exercised. The importance of open standards to make competition among prospective service providers possible was often emphasised. However, seemingly no measurable objectives were discussed when the land transfer conditions were created. The new GDPR legislation was considered to hinder the development of services via home automation due to strict demands in the processing of personal data.

It wouldn't make sense to have sensors from three different companies in the living room because the homeowner wants to use three different services.

- An interviewee from the group 1

Arguments from the group 2 interviewees were mostly critical. The overall objective of making the district a testbed for smart grid piloting was agreed, but the section in the land transfer conditions covering home automation was not regarded as leading to an optimal use of resources. Instead, it was considered to be somewhat burdensome, disproportionately increasing constructor costs. Most interviewees also pointed out that the electric energy consumption of a typical apartment is small and even has decreased further during the validity of the land transfer conditions, due to the improved energy efficiency of household appliances and light fittings. It was thus believed that loads of a reasonable size in electric demand management applications are uncommon in residential apartments. Some interviewees were worried about whether technical progress would outstrip the home automation standard described in the land transfer conditions. Although the standard may have initially been considered modern, it may become outdated during the two decades it will take until the construction in Kalasatama is completed.

All interviewees from the construction companies expressed their concern that the requirement to equip apartments with home automation was not directly beneficial to their business operations. It increased their costs, but they could not pass on this premium cost in the apartments’ sales prices, nor did it have influence on the selling time of apartments. The developers also expressed that the land transfer conditions do not take into consideration demographic issues in the district and their effects on housing. Since solvency varies between customer groups, construction companies tend to make small rental apartments less equipped than large owner-occupied apartments. Tenants usually live for a relatively shorter period of time in an apartment and are thus less motivated to invest in products and services aiming to improve the energy efficiency of the apartment compared to those living in owner-occupied homes. The development and construction companies also did not want to make binding contracts for energy-saving services on behalf of the prospective customers during the construction phase of the apartments (i.e. before apartment ownership has been handed over), as such contracts could increase the cost of living of those customers who have no interest in using them.

We still build stylish, smart, fine, architecturally high-quality homes. Then there’s the issue of the money that we spend on planning, the money that we spend on contracting, the money that we spend on this learning (of home automation), and we can't transfer that cost to the sales price of the apartment.

-An interviewee from the group 2

We could only identify a single company that is offering energy management services based on the defined home automation in Kalasatama (the interviewee group 3). The commodity of that company is a mobile phone app that customers can use to monitor their home energy usage and make minor changes to the functionality of their home automation systems. Demand management is not possible with the app. A similar app is also available to the customers of the company outside Kalasatama. Since the electrical systems in the Kalasatama smart city district are unique, the form of delivery of the solution this company brings to their customers outside Kalasatama is different and, instead of a fixed KNX installation, is based on wireless retrofit energy meters, actuators, and a gateway. One major real estate company also offers its lessees a mobile phone app with some smart home features, but this service is not different in Kalasatama than the other parts of the country and does not involve the home automation features of Kalasatama.

The original thesis is that consumption (from appliances in an apartment) has to be grouped together and that it must be considered which of these consumption groups can be connected to (electric) demand management; so maybe that's the original driver and motive, but finding significant loads is difficult.

-An interviewee from the Group 3

The home automation regulation of Kalasatama is a unique proposition among smart city programmes. In this article, we have documented the proposition for the audience following the scientific literature and have analysed its outcome using current knowledge of residential electric demand management and stakeholder feedback. Our views concern to the implementation of the home automation specification by the Kalasatama land transfer conditions and are not to be generalized to other smart city projects or to the potency of home automation in demand management or energy efficiency in general.

In the apartment panel boards of Kalasatama, loads must be arranged into load groups according to their function, and each load group must be able to be switched on or off using home automation. The objective of this is to improve energy efficiency by increasing the smartness of homes and creating the readiness for electric demand management; however, due to solution attributes—only switching of load groups was specified, instead of power adjustment of individual loads—the solution did not seem to adequately address the challenges presented in the literature, and a reasonable demand management potential provided by the load groups was not recognised.

The home automation standard in the land transfer conditions was based on the reasoning that although a single dwelling is not a considerable consumer of electric power, a large number of dwellings could be aggregated to provide a reasonable combined demand management potential. The economic competence of this idea was seemingly not assessed when the standard was drafted. As the minimum power to participate in the reserve market is high and the demand management potential of a single dwelling is low, it becomes a challenge for prospective service providers to aggregate a sufficient number of loads together. The load groups were not found to be particularly applicable for time-shifting applications either.

Assessing the size of a potential market is a key topic for any company considering investing in a new product or service. According to the smart city vision, the district of thousands of homes with standardised home automation would attract entrepreneurial companies to offer new kinds of energy management–related services, and generate a marketplace which would not emerge without the existence of the installation base of home automation systems. However, such services hardly seem to have appeared during the validity of the land transfer conditions. Either the business opportunity has not been recognised, or it has been disbelieved during the evaluation phase. If a company wanted to start offering services based on the home automation solution of Kalasatama, it would first have to invest in the development of the solution. The number of apartments in the smart city district multiplied by the estimated average revenue per apartment is seemingly insufficient. Smart home companies do not seem to consider Kalasatama’s local home automation standard attractive. Their response might be different if the standard covered a larger area (e.g. the entire country).

When policymakers set norms for construction, they consider them to be interpreted as minimum requirements. Construction companies are free to surpass the norms, but under cost pressures, the minimum requirement often becomes the de facto standard, and this seems to be happening with Kalasatama’s home automation as well. For example, the constructors could add luxury features to the home automation that presently only covers energy management, but seemingly no such demand has been identified in the housing markets of Kalasatama. However, the home automation requirement obviously does not have such a significant negative impact on profitability to lead to withdrawals from construction projects in the area. The existing land transfer conditions also have a degree of ambiguity, which has led to some confusion in planning and contracting. This can lead to inconsistencies between home automation systems built in different projects, which is the opposite of the smart city objective and may further increase the financial risks of potential service providers.

The original working group that drafted the home automation standard consisted of large companies that understandably had their internal business motives involved, but they were not in the business of providing home energy management services to private customers. Apparently, the reasoning of these companies was that they would supply technology and someone else would take advantage of it. However, the home automation standard seems to be not well in line with the earnings logic of the existing companies. Installation of home automation belongs to electrical contractors, but after the installation on a site is complete and people have moved into their apartments, their work is over. Energy service and building management companies operate with HVAC systems of public and commercial buildings and with professional customers, not with residential electrical installations and private customers. This may have contributed to the low utilisation of home automation since apartments in the district have begun to complete. Based on these findings, we propose that similar smart city projects should be subject to a wider market dialogue between policymakers and prospective market participants before new standards are introduced at this scale. Such a debate could generate valuable information to evolve smart city programmes easier for market participants to exploit. However, construction in Kalasatama is still expected to continue for approximately 20 years, and the number of apartments following the home automation standard is going to continue to grow. Future research following the development of the Kalasatama smart city is therefore recommended.