Abstract Civil Engineering Design has seen a surge in AI‐driven analyses of concrete compressive strength, yet this proliferation has obscured a methodological crisis in supervised machine learning for sustainable materials research. Existing studies conflate two accuracy dimensions: target prediction accuracy, validatable against known labels, and feature importance accuracy, which lacks validation benchmarks. Establishing true variable associations requires two criteria from causal inference: consistency (stable feature rankings across perturbations) and dose–response relationships (monotonic changes in model output upon feature removal), criteria prior studies have failed to meet. We introduce a leave‐top‐feature‐out validation framework applied to a concrete compressive strength dataset, wherein top‐ranked features are systematically removed and perturbations quantified. Evaluation of supervised models (Random Forest, XGBoost), unsupervised methods (Feature Agglomeration, Highly Variable Gene Selection), and nonparametric statistics (Spearman correlation) reveals that supervised approaches produce volatile rankings susceptible to label‐driven biases, while unsupervised methods yield more consistent hierarchies, with SHapley Additive exPlanations (SHAP) explanations shown to amplify model biases. Our framework establishes the first theoretically grounded benchmark for distinguishing genuine from spurious associations in concrete strength prediction.
Reliable experimental data is essential for the development and validation of structural design models. However, existing bond databases often lack consistency, traceability, and overall data quality. To address this, the Technical University of Munich (TUM) developed the TUM Bond Database (TUMBD), a curated collection of experimental test data focusing on lap splices and anchorages under tensile loading. TUMBD includes a variety of full-scale beam tests from the primary literature and tests conducted at TUM. The TUMBD currently comprises 1033 lap splice and 189 anchorage tests, with detailed information on geometry, test setup, materials, and experimental results. This paper discusses the database's origin and development, its structure and content, and presents a statistical analysis of the distribution of key parameters. The open-access database supports model calibration and data-driven research, including the use of artificial intelligence and advanced computational methods. Given the complexity of bond behavior, the emergence of new resource-efficient materials, and the potential of such advanced statistical techniques, TUMBD enhances the reliability, sustainability, and efficiency of reinforced concrete construction. Ongoing updates and strict criteria ensure data quality.
In the face of increasing construction project complexity and the widespread adoption of Building Information Modeling (BIM), the efficient planning and execution of workspaces have become critical for enhancing productivity, safety, and overall project success. This study presents a BIM‐based workspace design tool that addresses the challenges of confined workspaces, concurrent activities, and ergonomically demanding tasks. The tool seamlessly integrates with BIM models to visualize, analyze, and optimize workspace configurations, ensuring that anchor installations, among other tasks, can be completed safely and efficiently. Through systematic research and the application of various use cases, the tool's effectiveness is demonstrated and validated, highlighting its potential to alleviate project managers' workload and significantly improve construction outcomes. The study concludes by outlining future enhancements and providing valuable insights for researchers and practitioners.
Infrastructure management along highways and railways requires inventories of critical structures like retaining walls, which traditionally rely on manual inspection and documentation. Unfortunately, data in infrastructure databases is often incomplete. This study investigates the feasibility of automating retaining wall inventories using public aerial lidar data from the Swiss Federal Office of Topography (Swisstopo) combined with deep learning. We develop a pipeline for data processing and apply the state‐of‐the‐art Superpoint Transformer architecture with SuperCluster for panoptic segmentation. Three distinct approaches are evaluated: transfer learning from general Swisstopo lidar data, transfer learning from the DALES aerial lidar dataset, and training a specialized model from scratch. The specialized model achieves the best performance with 44% Intersection over Union (IoU) for semantic segmentation and 24% panoptic quality on test data. Our findings reveal that the primary challenges stem from data characteristics—like sparse sampling of vertical surfaces due to oblique scanning angles—rather than model architecture limitations. This work provides insights into the development of automated infrastructure inventories and identifies areas for improvement, including the need for expanded training data and robust augmentation techniques.
The sensitivity to reduced installation effort of post‐installed fastening systems is determined during the product qualification through robustness tests conducted under controlled laboratory conditions. Using the concept of required α ‐factors, the installation safety factor γ inst is determined and provided as an essential characteristic in the corresponding European Technical Assessment (ETA) of the fastener. The purpose of robustness tests is to simulate reduced installation conditions. To gain more insight on the variability of the installation conditions in practice, an in‐situ verification of the installation safety factor is carried out on four different construction sites in Austria using four anchorage types to cover all working principles. For this in‐situ verification, the obtained ultimate loads from unconfined pull‐out tests on‐site are compared with corresponding ultimate loads from reference tests under laboratory conditions. Based on the in‐situ tests, it is verified if the designed robustness tests are sufficiently appropriate to cover in‐situ variations during the installation. The investigation shows that for all tested fastening products, the specified installation safety factor from the ETA could be confirmed with the in‐situ tests based on mean values. As expected, the coefficient of variation of the in‐situ tests is for torque‐controlled mechanical fasteners around two to three times higher than in the laboratory tests, for concrete screws and bonded anchors, the value 1.5 applies.
