Virtual Disaggregation — a game changer for energy efficiency innovations
Non-intrusive load monitoring (NILM), also known as Virtual Disaggregation, is a data-driven approach to identify the operational state (on/off) and the accurate power consumption of individual electrical loads, considering as input only the aggregated consumption of these loads. In this report we take a look at why and where NILM can help energy decarbonization in buildings.
The developments on NILM goes back to 1992 with the initial efforts of Hart [1, 2] and then strongly advanced with the developments in Artificial Intelligence and Machine Learning (ML). As the terms virtual and non-intrusive imply, this method is implemented with minimum interruption in the building and its micro-grid. Measurements are obtained only from a single smart meter (Fig. 1), so there is no need for deployment of extra equipment that would increase the complexity and the cost of the installation.
Why NILM?
Our experience at Builtrix large scale pilots with entities managing over 1000 buildings — distributed around the country or in some cases around the continent, tells us that the major drawback of the smart buildings is the purchase, installation, and maintenance of many IoT devices and sensors.
Reports indicate that the capital cost for installation of each electricity smart meter is around 200 EUR [3]. This cost in conjunction with the occupants’ disruption during the installation do not allow property owners/managers to buy and install them. However, the cost and installation time are not the only issue and deployment of such huge sensor network, connectivity, data collection, and storage may have destructive environmental impacts in terms of increased electronic wastes and carbon emissions. As the wise man once said “less is more”, less hardware requirement offered by NILM supports the best practices for building energy efficiency and decarbonization projects.
Demands for NILM
The major demands for NILM are coming from three main sources: I- Utility companies, II- Commercial and industrial buildings, III- Residential buildings.
The number of utility providers in our cities are increasing. Electricity suppliers are looking for innovative approaches, mainly proposed by Small-Medium-Size Enterprises (SMEs) to improve customer engagement and differentiate in this competitive markets. In addition, NILM can even create new revenue streams for energy retailers based on consumer needs which can be derived from their consumption behavior. Clearly, NILM sounds very fit for this purpose, while the demand from retail suppliers keeps increasing. Meanwhile, utilities also consider NILM as a good solution for demand-side flexibility management. Thus, NILM helps utilities to gain real-time granular consumption data and to conduct load-shaping actions such as customizing Time of Use (ToU) rates, management of local energy communities, deployment of photovoltaic plants, batteries, and EV chargers. NILM helps in generating appliance level data and smarter contribution in demand side flexibility.
Commercial and industrial buildings need tools to help them not only optimize their energy assets to achieve energy savings but also detect potential equipment malfunction before further deterioration causes more significant issues. Furthermore, sustainability regulations and need for dynamic green certification is pushing service buildings towards demanding more granular energy consumption data.
Increasing adoption rate of smart appliances, electric vehicles (EVs), and building-integrated PV plants demonstrates a rising number of energy producers and customers (so called prosumers) who request granular and instant information about their energy consumption and generation in order to optimize their self-consumption plans.
Business models around NILM
In the current business models, either sensor-based or SaaS products are delivering real-time and granular consumption monitoring service and more new value-added services based on consumption data analysis. These business models include:
B2C Plug-and-play service with the key value-added around load-shaping recommendations for energy savings and home-automation for residential buildings.
B2B plug-and-play services (or paired with existing sub-meters) with the key value-added around load optimization for energy savings, predictive maintenance and dysfunction detection for appliances and equipment in commercial and industrial sectors.
B2B2B services mainly delivered by white-label products or API to help utilities improve consumer engagement. The Key added value services include market place for consumers to buy new energy-efficient appliances, energy infrastructure plans (ex: EV charging station), ToU tariff design and home-automation for load-shaping.
NILM as a decarbonization enabler
Commercial buildings worldwide are a major source of carbon emissions and changing energy use behavior in these environments has the capacity for large Greenhouse Gas (GHG) reduction. As stated in the Energy Performance Directive, the European Union (EU) is committed to developing a sustainable, competitive, secure, and decarbonized energy system by 2050. To meet this goal, special attention must be paid to the building sector as it accounts for 38.9% of all the energy consumed in the EU and is among the largest end-use consumer sectors. Various approaches have been proposed to reduce the energy consumed by buildings, including the adoption of building energy efficiency standards, promoting building renovation and implementing applied ICT solutions for building automation, among others.
All these investments and deployments of new generation of IoT devices powered by 5G and other revolutionary technologies for fast and secure data transfer are shaping the common energy data space. A rich pool of data is the first step towards promoting data-driven energy innovations sophisticated on big data analysis and ML. These innovations are the back bones and intelligent brains of automation technologies, focused on running and actuation of energy intensive components of buildings such as Heating, Ventilation, and Air Conditioning (HVAC), Lighting, Refrigeration, Electrical Vehicle (EV) chargers, Photovoltaic plants, and Storage systems. The main responsibility of this brain is to enable the installed renewable energy, storage, metering and automation hardware. By enabling the most expensive parts of the the smart building projects (the hardware part), the return on investment (ROI) shrinks down and property owners may become more comfortable with decarbonization plans. Thus, solutions such as NILM that not only reduce the need for hardware installation, but also provide data for decision making of machines and humans are key the key smart building tools developed by data and AI.
Main driver of the NILM market
The main drivers of NILM market are the EU smart meter rollout, promotions, and intensives in the area of energy digitalization. Back in 2014, the EU proposed a document outlining the need for smart meters to be rolled out across EU member states. It proposed that each country should install electricity smart meters across all residential and non-residential buildings and achieve 80% coverage by 2020. Now in 2022, several EU countries have surpassed their requirements and reached to a second phase of upgrades. However, some countries are behind their milestones and some others abandoned the commitment. According to the European Commission’s review in 2017, on average, only 37% of EU consumers were equipped with smart electricity meters.
Germany, Czech Republic, Greece, Croatia, and Cyprus were the few countries who decided to reject the rollout mainly due to unprepared legal measures, unknown data sharing and privacy policies, or financial barriers towards supporting the costs. Some of these issues have been solved in regional or country level and these countries have started to meet the rollout as part of their green deal agreements.
Spain, Sweden, Finland, Estonia, and Denmark with +80% deployment of smart meters in 2020 are the leaders of this rollout. Actually, Spain had a smart meter price advantage (40% cheaper) and the government mandated the installations. So, they were the first to reach 100% installation for households in 2018 and 40% for non-residential in 2019. Finland achieved 100% rollout of its target and is now preparing for a second rollout with higher resolution data (15-min) to be centralized in a data hub. It should be indicated that many other countries such as France, Ireland, Portugal, and Lithuania are trying to meet the rollout (80–100% deployment) by 2024.
Therefore, the future of EU energy (maybe by 2030) looks like every building has one central meter providing high resolution but aggregated energy data. This is what NILM model developers should consider for their market development among other technical criteria such as data model and interoperability concerns.
A business use case for NILM
I decided to write this section about Energy Gamification business use case to share my experience about 3 years of developing software and hardware product — and business — in this area (Optishower). Me and my co-founder exited the business in 2020 with extensive learning on scalability. In 2017, we had to install 3-6 smart meters per hotel room to have a full overview on the electricity consumption. Besides the cost of smart meters, the time for installation and solving the connectivity challenges were insane! Although, the results of the pilots where successful (see the short feedback from Marriott hotel — Amsterdam), the technology was not mature enough to support the fast integration and scalability of the product. This is why I admire the recent efforts inML developments in Virtual Disaggregation field and I believe, if NILM was available by that time, it could help us to achieve bigger markets.
Please be aware that all the technological advances in building systems directly contribute to just 42% of energy efficiency, which suggests that an impact on energy savings is highly dependent on behavioral plasticity. The EU Environment Agency claims that behavioral change measures can deliver sustained savings between 5% and 20% on top of the technological measures.
According to the Oxford dictionary gamification means: the application of typical elements of game playing (e.g. point scoring, competition with others, rules of play) to other areas of activity, typically as an online marketing technique to encourage engagement with a product or service.
“gamification is exciting because it promises to make the hard stuff in life fun”.
Most of existing campaigns for fostering energy conservation behaviors are typically designed as information-intensive and they seem not to be enough motivating. However, a growing body of literature supports the use of gamification as a method of producing attitude and behavior change [4–7]. Within this context, serious games are defined as virtual simulations of real-world activities that can educate users and prompt behavioral change.
The most recent gamification initiative was EU-funded Horizon 2020 project EnerGAware — Energy Game for Awareness of energy efficiency in social housing communities. In the EnerGAware project, an innovative serious game was developed and implemented to promote a reduction in energy consumption and CO2 emissions in social housing. The project outcome has been tested in 100 homes, as a serious game, that linked to the actual energy consumption (smart meter data) of the game user’s home. The EnerGAware solution provided an innovative IT ecosystem in which users can play to learn about the potential energy savings from installing energy-efficiency measures and changing user behavior. The EnerGAware project yielded positive influence. It provided an average electricity saving of 3.46% and gas saving of 7.48%. Optimal time course of the project was around three months. Positive behavior change were seen for relatively short term and diminished by the final stage. Moreover, it was also important to note that the game complexity and lack of support were critical issues, subjects did not perceive that game was linked to their behavior which explains lack of real-life energy saving as result of playing the game.
In Addition to EnerGAware, project ChArGED (CleAnweb Gamified Energy Disaggregation) ran parallel in the timeline. CharGED addressed energy consumption in public building with a framework that facilitates achieving greater energy efficiency and reductions of wasted energy. It leverages IoT enabled devices to improve energy disaggregation mechanisms that provide energy use and wastage events at the device, area, and end user level.
Another EU project which is ongoing is towards a new generation of EU peer-to-peer Energy Communities facilitated by a gamified platform and empowered by user centered energy training and business models is called NRG2peers (started 1st September 2020 and will last till 31st August 2023). This project mainly aims for global energy and CO2 emission savings at the community level and encourage investments in sustainable energy in the EU.
Questions that may rise here are:
- Is consumption behavior of tenants a real pain of property managers or its just nice and fancy to have a gamification tool in their building?
- Does it worse to invest on data-driven behavior change technologies?
- What is the annual saving and RoI period of behavior change actions?
While there is no strong answer for these questions, gamified energy saving solutions should always move toward shorter RoI period by reducing the need for hardware, setup, and their maintenance costs. If you run such a business, you should aim for fast integration and short RoI, now.
Conclusions
The common challenge of many data-driven energy efficiency and decarbonization projects is their major requirements to the high resolution consumption data. The more detailed data is involved, i.e., in the appliance level, the better insight, awareness, and action can be provided. Solutions such as NILM can play an important role in such innovation by non-intrusive monitoring of building appliances.
NILM can be the AI model deployed on the server as an API endpoint to generate accurate appliance level time series and deliver it to the game or decision support engine for the gamification/insight process. Although, NILM looks fancy, but it has its own challenges for scalability. I am happy that in my new business (Builtrix) we are trying to solve these challenges with the supports provided by outstanding research centers, data providers, and Digital Innovation Hubs (DIHs). If you want to be part of this exploration and problem solving, please feel free to contact us (info@builtrix.tech).
I will continue to update this post in future with the links and references to our new product and business achievements.
References
[1] G.W. Hart, Nonintrusive appliance load monitoring, Proc. IEEE 80 (1992) 1870–1891.
[2] Pujić, D., Jelić, M., Tomašević, N., Batić, M. (2020). Chapter 10 Case Study from the Energy Domain. In: Janev, V., Graux, D., Jabeen, H., Sallinger, E. (eds) Knowledge Graphs and Big Data Processing. Lecture Notes in Computer Science, vol 12072. Springer, Cham. https://doi.org/10.1007/978-3-030-53199-7_10
[3] Study on cost benefit analysis of Smart metering System in EU Member States. Available online at [link], accessed in June 2022.
[4] Gaglia, A. G., Dialynas, E. N., Argiriou, A. A., Kostopoulou, E., Tsiamitros, D., Stimoniaris, D., & Laskos, K. M. (2019). Energy performance of European residential buildings: Energy use, technical and environmental characteristics of the Greek residential sector–energy conservation and CO₂ reduction. Energy and Buildings, 183, 86–104.
[5] Casals, M., Gangolells, M., Macarulla, M., Forcada, N., Fuertes, A., & Jones, R. V. (2020). Assessing the effectiveness of gamification in reducing domestic energy consumption: Lessons learned from the EnerGAware project. Energy and Buildings, 210, 109753.
[6] Soares, F., Madureira, A., Pagès, A., Barbosa, A., Coelho, A., Cassola, F., … & Sørensen, T. (2021). FEEdBACk: An ICT-Based Platform to Increase Energy Efficiency through Buildings’ Consumer Engagement. Energies, 14(6), 1524.
[7] Staddon, S. C., Cycil, C., Goulden, M., Leygue, C., & Spence, A. (2016). Intervening to change behaviour and save energy in the workplace: A systematic review of available evidence. Energy Research & Social Science, 17, 30–51.
