Periodic Reporting for period 1 - DigiBUILD (High-Quality Data-Driven Services for a Digital Built Environment towards a Climate-Neutral Building Stock)
Berichtszeitraum: 2022-06-01 bis 2023-11-30
Traditional silo approaches, where stakeholders manage their own data, could be replaced by digital and smart buildings, merging heterogeneous data sources, and placing the stakeholders as the core of these buildings. DigiBUILD will catalyse this much-needed transformation by making use of high-quality data and next generation digital building services, supporting the deployment of EU-wide Framework for a Digital Building Logbook.
An inclusive environment for multi-stakeholder knowledge exchange (based on European Bauhaus initiative) will be applied to co-design end-user-oriented services. DigiBUILD will provide an open, interoperable, and cloud-based toolbox to transform current ‘silo’ buildings into digital, interoperable, and smarter ones, based on consistent and reliable data, supporting better-informed decision-making for performance monitoring & assessment, planning of building infrastructure, policy making and de-risking investments. It will be built on top of existing platforms and common EU initiatives, towards an Energy Efficient Building Data Space, based on standard cloud-data platform frameworks (FIWARE) and Data Space initiatives (GAIA-X and IDSA). On top of this advanced data governance framework, we will create AI-based data analytics and Digital Building Twins based on high-quality data, aiming to facilitate transparency, trust, informed decision-making and information sharing within the built environment and construction sector, which will be deployed across 10 real-world conditions (TRL 8). DigiBUILD will contribute to the uptake of digital technologies in the building sector to better align the EU Member States’ long-term renovation strategies with the EPBD requirements on decarbonisation, and on a path towards a climate-neutral building stock by 2050.
An inclusive environment for multi-stakeholder knowledge exchange (based on European Bauhaus initiative) will be applied to co-design end-user-oriented services. DigiBUILD will provide an open, interoperable, and cloud-based toolbox to transform current ‘silo’ buildings into digital, interoperable, and smarter ones, based on consistent and reliable data, supporting better-informed decision-making for performance monitoring & assessment, planning of building infrastructure, policy making and de-risking investments. It will be built on top of existing platforms and common EU initiatives, towards an Energy Efficient Building Data Space, based on standard cloud-data platform frameworks (FIWARE) and Data Space initiatives (GAIA-X and IDSA). On top of this advanced data governance framework, we will create AI-based data analytics and Digital Building Twins based on high-quality data, aiming to facilitate transparency, trust, informed decision-making and information sharing within the built environment and construction sector, which will be deployed across 10 real-world conditions (TRL 8). DigiBUILD will contribute to the uptake of digital technologies in the building sector to better align the EU Member States’ long-term renovation strategies with the EPBD requirements on decarbonisation, and on a path towards a climate-neutral building stock by 2050.
The Consortium has identified most of user requirements, starting from use cases and user stories, by adopting a co-creation methodology together with DigiBUILD pilots and identified the most relevant system requirements. Interoperability was specified as well as the software architecture. Security and privacy aspects have been considered too. The data security and trust framework has been defined, based on the application of blockchain technology, to certify the real-time data monitoring and the existence and persistence of data in the Data Lake. A pilot audit and data inventory have been developed to help understand the data availability. Based on these datasets, ETLs and data quality methodology have been implemented to convert the data into high-quality data. Regarding static data, the adopted approach was to understand the information requirements and specify the DigiBUILD ontology supports the creation of knowledge graphs for unified data querying.
From the Data Lake point of view, the different components of the Data Lake have been identified and technology has been selected to approach the Data Lake. Also, an initial schema for data storage and the synchronisation methods are defined to keep coherence between dynamic and static repositories. Finally, the first technological approach for generating Data Marts and Intelligent Querying was elaborated.
Substantial analyses were performed on pilot data, facilitating the development of suitable data preprocessing procedures tailored to each use case. AI models were meticulously designed, developed, and trained to align with the project’s use cases, utilizing the established infrastructure.
A specific methodology was conceived with the main goal to refine the definition of DTs services.
High level architecture concerning DTs was designed and a first release of the Digital Twin Suite and cloud-based DigiBUILD data toolbox was released.
A comprehensive plan (M&V) was developed that defined the steps and procedures for the implementation and validation of the DigiBUILD services in the real-life pilot operation demonstrators.
To ensure that all the necessary details were clarified and defined, active input was asked from the pilots themselves. This collaborative approach allowed their specific needs and requirements to be taken into consideration during the development of the plan, aligning it with user requirements and associated services. The Key Performance Indicators (KPIs), properly selected, serve as measurable metrics to evaluate the performance and effectiveness of DigiBUILD services within each pilot project.
The baseline assessment has been successfully completed. During the pre-pilot phase, the first version of the services developed has been validated through all the pilots, utilizing high-quality data. The process involved identifying areas for improvement and outlining the necessary steps for their real implementation.
From the Data Lake point of view, the different components of the Data Lake have been identified and technology has been selected to approach the Data Lake. Also, an initial schema for data storage and the synchronisation methods are defined to keep coherence between dynamic and static repositories. Finally, the first technological approach for generating Data Marts and Intelligent Querying was elaborated.
Substantial analyses were performed on pilot data, facilitating the development of suitable data preprocessing procedures tailored to each use case. AI models were meticulously designed, developed, and trained to align with the project’s use cases, utilizing the established infrastructure.
A specific methodology was conceived with the main goal to refine the definition of DTs services.
High level architecture concerning DTs was designed and a first release of the Digital Twin Suite and cloud-based DigiBUILD data toolbox was released.
A comprehensive plan (M&V) was developed that defined the steps and procedures for the implementation and validation of the DigiBUILD services in the real-life pilot operation demonstrators.
To ensure that all the necessary details were clarified and defined, active input was asked from the pilots themselves. This collaborative approach allowed their specific needs and requirements to be taken into consideration during the development of the plan, aligning it with user requirements and associated services. The Key Performance Indicators (KPIs), properly selected, serve as measurable metrics to evaluate the performance and effectiveness of DigiBUILD services within each pilot project.
The baseline assessment has been successfully completed. During the pre-pilot phase, the first version of the services developed has been validated through all the pilots, utilizing high-quality data. The process involved identifying areas for improvement and outlining the necessary steps for their real implementation.
DigiBUILD proposes several advancements beyond that state of the art. Concerning the semantic interoperability for adaptive digital building services, DigiBUILD is proposing a dynamic and adaptable interoperability framework based on data modelling for data representation according to a representative set of use cases and user stories defined in the project. For data quality, DigiBUILD is deploying a methodology, supported by a federated data lake, that covers the data life-cycle of the building. It covers the different stages to reduce error propagation. The project is also working on the integration with existing energy data spaces (Omega-X, Data Cellar, ENERSHARE as well as the BSO) by properly adapting and deploying Connectors (starting from the ENG's TRUE/OneNet Connector). DigiBUILD proposes AI/ML analytics for smart building applications and advances the SOTA by developing ML/DL multi-functional analytics as microservices, with a view to match information privacy/security and low latency constraints and requirements. Reinforcement learning algorithms are being implemented and used for individual or combined electrical and thermal peak shaving to reduce network losses. ML and hybrid inspired heuristics are being defined and used for the delivery of combined energy services and with comfort AAL services and for optimal cross sectors services combination to reduce the transactions costs and capital expenses.
DigiBUILD is also adopting the ‘Comfort Performance Contract’ service by ensuring users an optimal level of thermal comfort for the entire duration of their stay inside the building. Personalised thermal comfort models are being developed through environmental, physiological, and personal parameters, that considers the quality of the collected data, using DL and ML algorithm (LSTM, CNN, SVM, etc.). Two methodological frameworks are being introduced to define robust solutions for decision-making under uncertainty for efficient and climate resilient buildings. Concerning Digital Twins for buildings and districts, the advancement consists in deploying DT in buildings to have real time insights and control energy consumption, the development of automated systems for construction of BIM, exploiting AI for advanced analytics combined with expertise on energy and comfort by partners involved.
DigiBUILD is also adopting the ‘Comfort Performance Contract’ service by ensuring users an optimal level of thermal comfort for the entire duration of their stay inside the building. Personalised thermal comfort models are being developed through environmental, physiological, and personal parameters, that considers the quality of the collected data, using DL and ML algorithm (LSTM, CNN, SVM, etc.). Two methodological frameworks are being introduced to define robust solutions for decision-making under uncertainty for efficient and climate resilient buildings. Concerning Digital Twins for buildings and districts, the advancement consists in deploying DT in buildings to have real time insights and control energy consumption, the development of automated systems for construction of BIM, exploiting AI for advanced analytics combined with expertise on energy and comfort by partners involved.