SECAI
Sustainable heating through Edge-Cloud-based AI systems
| Short description |
| The SECAI project implements AI-controlled heating in private homes using an edge-cloud system, which not only reduces energy consumption but also can be retrofitted into existing buildings. |
SECAI is using an edge-cloud system to develop sustainable “smart living” solutions for more efficient home heating. These solutions can be retrofitted cost-effectively into existing buildings, are quick to implement, and provide a much-needed method, in addition to physical renovation measures, for meeting the climate targets set by the European Union and the German government. The project addresses three central aspects of intelligent heating: smart, AI-assisted adjustment of heating control in the home; optimisation of system control focusing on cost, climate, and user preferences; and encouraging the willingness of tenants to save energy. To achieve these objectives, SECAI integrates edge AI into the “Smart Living” data ecosystem and into existing concepts for secure and open data communication such as ForeSight and Gaia-X. Establishing acceptance factors for the use of the developed applications is also a top priority.
Market perspectives and product claims
In Germany, the heating and cooling of buildings and the provision of hot water account for approximately 18% of all CO2 emissions. More than 75% of homes in Germany use fossil fuels such as natural gas and oil for heating. The energy transition for buildings cannot be achieved solely through traditional renovation measures such as insulation. SECAI is developing an intelligent building control system that makes it possible to achieve climate targets in the building sector within a reasonable time frame. Thanks to the simple and cost-effective retrofitting of existing buildings (with edge sensor systems and cloud-edge infrastructure) and the provision of AI services, heating activities and demand can be analysed and optimised, leading to a significant long-term reduction in CO2 emissions and heating costs.
The project results have the potential to revolutionise the existing solutions offered by companies in the sectors of smart living, heating and cooling technology, and energy. The use of edge devices makes it possible to implement a service system according to the privacy-by-design principle. Consequently, the potential of AI can be put to use precisely where sensitive data is produced and should remain: in homes and buildings in compliance with data protection regulations. Communication with the cloud is carried out according to the principles of interoperability and data security by integrating European and German initiatives such as ForeSight and Gaia-X. This approach offers many opportunities to use AI in innovative business models while maintaining a high level of user acceptance. The solutions developed are scalable and can therefore be applied to multiple buildings, in particular to the building management of entire housing units.
Challenge and innovation
The project faces significant challenges, including that of combining the advantages of edge and cloud technologies while preserving residents' data sovereignty. To address this, the concept of federated learning is employed. This machine learning strategy involves training a model on various decentralised devices or servers, ensuring their respective data is retained.
The task of the AI models is to increase the efficiency of the heating system in order to reduce energy consumption and running costs. To this end, the entire living situation is taken into consideration – from the sensors in individual rooms (nano level) through the home (micro level) to entire buildings (meso level) and residential complexes or building stocks (macro level). The system determines the actual heating demand, for example using patterns of room usage over time, and at the same time synergies between the levels are identified and used to improve the efficiency of the heating.
In this way, both individual targets at the home level and the overall reduction of energy consumption in the whole system can be taken into account. Furthermore, SECAI allows external data such as weather data and forecasts to be included in the AI models in order to obtain the most accurate predictions possible. To this end, SECAI is working with UBIMET, an independent international weather service providing high-precision, customised and system-integrated weather data, in response to the “Edge data economy” call for funding applications issued by the Austrian and German ministries responsible for climate action.
Use cases
Over the course of the project, the system is made available to 20 tenants for them to use in their homes on a voluntary basis. The functionality of the edge-cloud system, its added value, and the AI services based on it are therefore all tested. To evaluate the efficiency and saving effects of the concept, data is monitored over three heating periods: Aggregated data from the last heating period is used to develop the system. Once the cloud-edge infrastructure has been installed, the data from the coming heating period is applied for use under real conditions. The complete system and the AI services are validated in the third heating period. In the process, the acceptance and effectiveness of the developed solutions are evaluated and optimised where necessary. To visualise the functionality and added value of the system, a demonstrator is also being developed.
Consortium
Strategion GmbH (consortium leader), Connectivity Solutions GmbH, Deutsches Forschungszentrum für Künstliche Intelligenz GmbH, Goethe University Frankfurt, GSW Gesellschaft für Siedlungs- und Wohnungsbau Baden-Württemberg mbH, Research Association for Electrical Engineering (FE) at the German Electrical and Electronic Manufacturers’ Association (ZVEI).
Benefits
| Current situation | Future vision |
| Digitalisation of the housing industry is stagnating at a low level; the potential of the edge data economy is not being fully exploited. | SECAI offers cost-effective AI and edge-cloud infrastructure solutions for retrofitting existing buildings in the Smart Living data ecosystem, thus advancing the digital transformation of the housing industry and other operators in the smart living domain. The project thus also supports the establishment of a lead market for technologies in the area of smart living. |
| Heating homes causes high CO2 emissions. Incentives for environmentally conscious heating are currently almost exclusively financial in nature and are limited to turning the heating down, often to a temperature that tenants find uncomfortable and therefore fail to maintain. In addition, many renovations of buildings have so far been limited to insulation, overlooking potential opportunities for CO2 savings. | SECAI makes it possible to reduce CO2 consumption by demand-oriented regulation of room and building temperatures together with suggestions for more energy-efficient behaviour and the visualisation of potential savings when such suggestions are followed. |
| Smart living services in the heating sector are usually limited to individual manufacturers; these are proprietary systems lacking interoperability, which also often only address isolated points of the heating system. | SECAI enables the interoperability of smart living services through alignment with standards such as Gaia-X and open-source technologies and integration in existing data ecosystems (ForeSight platform). The manufacturer-independent system opens up new business models to manufacturers and service providers, at all heating levels. |
| Concerns about data protection prevent the use of self-optimising AI solutions to increase the efficiency of heating systems in private homes. | Thanks to the combination of edge technologies and federated learning strategies, SECAI allows AI to be used according to the privacy-by-design principle: the data remains in the home, and data sovereignty remains with the tenants. |
| Smart living solutions in the heating sector are currently limited to the control of individual rooms or individual housing units. Although it is technically possible to control heating centrally as part of a building management system, this has not been addressed so far. | SECAI is developing a smart living solution that can also be used to control central heating systems in order to realise gains in efficiency across the individual housing units in the entire building or building complex. |
