The adoption of Building Information Modeling (BIM) and model-based design within the Architecture, Engineering, and Construction (AEC) industry has been hindered by the perception that using BIM authoring tools demands more effort than conventional 2D drafting. To enhance design efficiency, this paper proposes a BIM command recommendation framework that predicts the optimal next actions in real-time based on users’ historical interactions. We propose a comprehensive filtering and enhancement method for large-scale raw BIM log data and introduce a novel command recommendation model. Our model builds upon the state-of-the-art Transformer backbones originally developed for large language models (LLMs), incorporating a custom feature fusion module, dedicated loss function, and targeted learning strategy. In a case study, the proposed method is applied to over 32 billion rows of real-world log data collected globally from the BIM authoring software Vectorworks. Experimental results demonstrate that our method can learn universal and generalizable modeling patterns from anonymous user interaction sequences across different countries, disciplines, and projects. When generating recommendations for the next command, our approach achieves a Recall@10 of approximately 84%. The code is available at: https://github.com/dcy0577/BIM-Command-Recommendation.git .

Road infrastructure constitutes a complex system characterized by multiple interacting subsystems. A comprehensive understanding of the correlations and dependencies among various road infrastructure elements is essential for enhancing infrastructure management and maintenance by providing a robust foundation for decision support. Digital Twins (DT) are recognized as effective tools for facilitating such decision-making. However, the development of comprehensive DTs considering road infrastructure in its entirety is still in its early stages. Hence, this paper focuses on the conceptualization of a digital representation of road infrastructure that enables the evaluation of relationships between the various heterogeneous subsystems. To accomplish this, we apply the systems-of-systems principle to road infrastructure. At its core, Labeled Property Graphs (LPG) are employed to capture intra-subsystem relationships and inter-system linkages, facilitating a holistic representation of interactions. Furthermore, we acknowledge the current organizational status of distributed responsibilities resulting in distributed data storage and maintenance by using the concept of federated databases. The presented approach enables multi-scale evaluations of relations among road infrastructure elements while preserving the system’s scalability and the distributed management of infrastructure data. Thus, previously separate data sets can be evaluated in relation to each other on a big scale. Doing so, the presented concept provides a foundation for extensive correlation studies between different heterogeneous infrastructure datasets. The concept is validated by applying it to a large-scale real-world data set stemming from multiple Bavarian road authorities, transferring into the proposed graph structure, and demonstrating the gained capabilities through cross-domain queries and analysis. • Concept for representing road infrastructure as systems-of-systems using graph-based methods. • A proposal for context-based linking of heterogeneous infrastructure data models • Concept for evaluation of spatial relations between different infrastructure assets on multiple scales. • Consideration of distributed data storage and maintenance of road infrastructure data. • Validation through large-scale real-world dataset spanning multiple road authorities in Bavaria.

This paper proposes an automated method for creating semantic digital building models using dense point clouds and images. The method employs a hybrid bottom-up, top-down approach, integrating artificial intelligence capabilities in scene understanding with domain engineering knowledge to overcome challenges in indoor 3D reconstruction. The pre-trained PointTransformer semantic segmentation model extracts thirteen building objects, where the wall and ceiling segments are utilized in a 3D space parsing algorithm. The parameterized floor plan map is then generated using a data-driven approach, enabling the creation of an extruded volumetric digital model. Additionally, the YOLO V 8 object detection network recognizes doors and windows in images derived from projected points of the wall instances. The validation results for six building datasets with different layouts showcase the effectiveness of the proposed model reconstruction algorithm, with a mean error of about 7 cm between the parameters of elements in digital reference models and reconstructed models. This highlights AI’s potential in automating the creation of digital models for the real world.

The provision of geometric and semantic information is among the most fundamental tasks in BIM-based building design. As the design is constantly developing along with the design phases, there is a need for a formalism to define its maturity and detailing. In practice, the concept of Level of Development (LOD) is used to specify what information must be available at which time. Such information is contractually binding and crucial for different kinds of evaluations. Numerous commercial and open-source BIM tools currently support the automatic validation of semantic information. However, the automatic validation of the modeled geometry for fulfilling the expected detailing requirements is a complex and still unsolved task. In current practice, domain experts evaluate the models manually based on their experience. Hence, this paper presents a framework for formally analyzing and automatically checking the Level Of Geometry (LOG) of building information models. The proposed framework first focuses on generating a LOG dataset according to the popular LOD specifications. Afterwards, multiple geometric features representing the elements’ complexity are extracted. Finally, two tree-based ensemble models are trained on the extracted features and compared according to their accuracy in classifying building elements with the correct LOG. Measuring the modeling time showed a 1.88–2.80-fold increase between subsequent LOGs, with an 8–15-fold increase for LOG 400 compared to LOG 200. The results of classifying the LOG indicated that the combination of 16 features can represent the LOG complexity. They also indicated that the trained ensemble models are capable of classifying building elements with an accuracy between 83% and 85%.

The iterative and developing nature of designing a building involves the specification and handling of vague, imprecise, and incomplete information. A crucial factor for mitigating the impact of these uncertainties on the decision-making process is to effectively quantify and communicate them among the project stakeholders. The interactive visualization of 3D building models provides great support for evaluating building designs. However, the currently available visualization methods of the available authoring tools do not incorporate the potential uncertainties associated with the geometric and semantic information of building elements. Currently, building models appear precise and certain, even in the early design stages, which can lead to false assumptions and model evaluations, affecting the decisions made throughout the design stages. Hence, this paper presents a set of visualization approaches, including intrinsic, extrinsic, animation, and walkthroughs, that have been developed to present the uncertainties associated with the building elements’ information. The efficiency of the approaches developed in this study was evaluated through an online survey and interviews. More specifically, the approaches were compared in terms of intuitiveness, applicability, and acceptance. The evaluation results positively indicated the participants’ ability to understand the amount and impact of the uncertainties on the design by using the developed approaches.

For the design of large infrastructure projects such as inner-city subway tracks, it proves necessary to consider differing model scales, ranging from the scale of several kilometers down to a few millimeters. This challenge can be addressed by using multi-scale product models comprising multiple levels of detail (LoD). Ensuring consistency across the different LoDs can be achieved by applying procedural and parametric modeling techniques while creating the model. This results in a flexible multi-scale model that can be easily modified on one scale while other scales are automatically updated. However, the correct application of parametric constraints and procedural dependencies has shown to be a very complex and time-consuming process. To address this issue, this papers presents a semi-automated detailing mechanism, which is based on formal procedures based on graphs and graph transformations. This paper discusses how procedural parametric models based on two-dimensional sketches can be represented by graphs and how detailing steps in the form of parametric modeling operations can be formalized by using rule-based graph rewriting.

The planning of large infrastructure projects such as inner-city subway tracks is a highly collaborative process in which numerous experts from different domains are involved. While performing the planning task, widely differing scales have to be taken into consideration, ranging from the kilometer scale for the general routing of the track down to the centimeter scale for the detailed design of connection points. Currently there is no technology available which supports both the collaborative as well as the multi-scale aspect in an adequate manner. To fill this technological gap and better support the collaborative design and engineering activities involved with infrastructure planning, this paper introduces a new methodology which allows engineers to simultaneously manipulate a shared multi-scale tunnel model. This methodology comprises two main aspects. The first aspect is a multi-scale model for shield tunnels, which provides five different levels of detail (LoD) representing the different levels of abstraction required throughout the planning progress. The second aspect is a conceived collaboration platform, which enables simultaneous modifications of the multi-scale model by multiple users. In existing multi-scale approaches, where the individual representations are stored independently from each other, there is a high risk of creating inconsistencies, in particular in the highly dynamic collaborative planning context. To overcome this issue, the concept presented in this paper makes use of procedural modeling techniques for creating explicit dependencies between the geometric entities on the different LoDs. This results in a highly flexible, yet inherently consistent multi-scale model where the manipulation of elements on coarser LoDs results in an automated update of all dependent elements on finer LoDs. The proposed multi-scale model forms a well-suited basis for realizing the collaboration concept, which allows several experts to simultaneously manipulate a shared infrastructure model on various scales while using the different design tools they are accustomed to. The paper discusses in detail the principles and advantages of the proposed multi-scale modeling approach as well as its application in the context of collaborative tunnel design. The paper concludes with a case study of a large infrastructure project: a new inner-city subway tunnel in Munich, Germany.

Building Information Models (BIM) are comprehensive digital representations of buildings, which provide a large set of information originating from the different disciplines involved in the design, construction and operation processes. Moreover, accessing the data needed for a specific downstream application scenario is a challenging task in large-scale BIM projects. Several researchers recently proposed using formal query languages for specifying the desired information in a concise, well-defined manner. One of the main limitations of the languages introduced so far, however, is the inadequate treatment of geometric information. This is a significant drawback, as buildings are inherently spatial objects and qualitative spatial relationships accordingly play an important role in the analysis and verification of building models. In addition, the filters needed in specific data exchange scenarios for selecting the information required can be built by spatial objects and their relations. The lack of spatial functionality in BIM query languages is filled by the Query Language for Building Information Models (QL4BIM) which provides metric, directional and topological operators for defining filter expressions with qualitative spatial semantics. This paper focuses on the topological operators provided by the language. In particular, it presents a new implementation method based on the boundary representation of the operands which outperforms the previously presented octree-based approaches. The paper discusses the developed algorithms in detail and presents extensive performance tests.