EUVIMEDEUVIMED entdecken
Crossref

Machine Learning‐Based Assessment of Manufacturing Tolerances on the Initial Stiffness of Bolt‐less Beam‐to‐Upright Joints

Marcelo Calispa, Jon Larrañaga, Jon Elguezabal, David Santamaria, Ibai Ulacia

ce/papers · 2026-09 · 9 (2-3) · 583-588

DOI: 10.1002/cepa.70389 ↗

Kurz gefasst

Fachartikel in der Zeitschrift ce/papers, online erschienen am 14.09.2026. Verfasst von Marcelo Calispa und 4 weiteren Autor:innen. Die Arbeit nennt 17 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The structural performance of steel storage racks is governed by the mechanical behaviour of beam‐to‐upright joints (BUJ), which control global stiffness and stability. Traditional stiffness prediction relies on experimental testing or numerical models, both constrained by cost and data availability. This study proposes a data‐driven framework to predict the initial stiffness of bolt‐less BUJ joints, with emphasis on evaluating the effectiveness of synthetic data augmentation. An experimental database is used as the reference dataset, and synthetic samples are generated using four augmentation techniques that represent realistic variations in geometry, material properties, and manufacturing tolerances. The original and augmented data are used to train and evaluate four machine learning models (Artificial Neural Network, Multilinear Regression, Gaussian Process Regression, and Support Vector Machine). The main novelty of the research lies in systematically quantifying how synthetic data affect model accuracy, robustness, and generalization in data‐scarce conditions, with manufacturing tolerances explicitly embedded in the augmented data. From the comparative analysis feature‐space interpolation combined with ANN provides the best balance between accuracy and generalization.

Zitieren

APA

Calispa, M., Larrañaga, J., Elguezabal, J., Santamaria, D., & Ulacia, I. (2026). Machine Learning‐Based Assessment of Manufacturing Tolerances on the Initial Stiffness of Bolt‐less Beam‐to‐Upright Joints. ce/papers, 9(2-3), 583-588. https://doi.org/10.1002/cepa.70389

Vancouver

Calispa M, Larrañaga J, Elguezabal J, Santamaria D, Ulacia I. Machine Learning‐Based Assessment of Manufacturing Tolerances on the Initial Stiffness of Bolt‐less Beam‐to‐Upright Joints. ce/papers. 2026;9(2-3):583-588. doi:10.1002/cepa.70389

BibTeX
@article{Calispa2026,
  author = {Calispa, Marcelo and Larrañaga, Jon and Elguezabal, Jon and Santamaria, David and Ulacia, Ibai},
  title = {Machine Learning‐Based Assessment of Manufacturing Tolerances on the Initial Stiffness of Bolt‐less Beam‐to‐Upright Joints},
  journal = {ce/papers},
  year = {2026},
  volume = {9},
  number = {2-3},
  pages = {583--588},
  doi = {10.1002/cepa.70389}
}
Kostenlos starten →