In the construction industry, technical, legal, and ecological requirements for buildings are predominantly available in textual form (e.g., standards, guidelines, or legislation). While these requirements are interpretable by humans, they cannot be directly utilized for digital, automated processes. At the same time, openBIM is becoming increasingly important as a foundation for model-based planning and validation. However, a consistent and methodical connection between textual requirements and their machine-readable translation and execution is still lacking.

This initial situation is further intensified by several structural challenges: today, the creation and implementation of model-based validation processes require both domain-specific and technical expertise, which effectively excludes non-programmers and non-specialized openBIM users from actively contributing to the process. At the same time, there is no practical and transparent methodology for systematically transforming textual provisions into machine-readable logic. Existing solutions are also predominantly proprietary, creating dependencies on software vendors and only partially fulfilling key requirements of public clients, such as transparency, version control, traceability, and long-term availability. Furthermore, the actual potential of existing information requirements for automated openBIM-based validation has so far only been insufficiently explored.

The RCC4ÖAG research project addresses this gap through the development of an open, transparent, and reproducible methodology for transforming natural language requirements into structured, machine-readable validation logics. The objective is to establish a generic process by which textual specifications can be systematically transformed into model-based requirements and executable logic. For methodological validation and practical testing of this approach, the developed concepts will be implemented within an open source–based system environment that enables visual modeling of the resulting logics. This ensures that the methodology is not only theoretically described, but also applicable, verifiable, and extendable under real-world conditions.

The diagram illustrates the transformation chain, from textual requirements through pseudocode and modular blocks to automated testing (RCC check).

At the core of the project is a multi-stage transformation chain:

• Prioritization and evaluation of provisions with regard to model relevance, automability, and adaptation potential
• Derivation of pseudocode that structures and describes the logic
• Transformation into modular building blocks (nodes) that can be combined within a visual environment
• Execution of the logics (RCC Check) within an open, version-controlled, and traceable system architecture 

A central aspect of the project is the combination of Large Language Models (LLMs) with a Human-in-the-Loop (HITL) approach to support this transformation chain. This enables domain experts to create and adapt complex logic without programming knowledge. A visual modeling environment (visual scripting) serves as the interface between domain-specific descriptions and technical implementation.

The project consistently follows an open source approach: all developed methods, data structures, and software components will be made freely accessible. This creates a vendor-independent, maintainable, and transparent foundation that particularly meets the requirements of public clients. Another focus lies on the open storage and reusability of results, for example through the generation of IFC-based data structures, ensuring that newly derived information remains interoperable and available in the long term.

In the long term, RCC4ÖAG will make a significant contribution to standardization, digitalization, and automation within the construction industry. By openly providing the project results, a shared foundation for administration, planning, and execution will be established, sustainably improving the efficiency, traceability, and quality of digital construction processes.

In addition to the project partners, the initiative is supported by so-called supporting partners. These contribute additional professional requirements, use cases, and practical feedback without being directly involved in the research and development work. This ensures that the developed methods are broadly aligned, practically validated, and transferable across different application domains.

Austrian Association of Engineers and Architects (ÖIAV) (Consortium Lead)

• ASFINAG
• Bundesimmobiliengesellschaft (BIG)
• ÖBB Infrastruktur AG
• Federal State of Lower Austria
• Federal State of Styria
• City of Vienna
• Wiener Linien
• TU Wien – Research Unit for Digital Building Processes 

Supporting Partner:

• ODE

Supporting Partners officially back the project. They gain early access to the open-source code for testing purposes and help with its further development. If you are interested in becoming a Supporting Partner, please feel free to contact us: harald.urban@tuwien.ac.at