EUVIMEDEUVIMED entdecken
Crossref

A Data-Driven Computational Framework for Predicting the Fundamental Natural Frequency of Thin Rectangular Plates

Shehryar Ahmed, Mehran Sudheer, Tahsin Ullah, Kefayat Ullah

Journal of Machine Learning Advances · 2026

DOI: 10.66879/jmla.2026.64 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Machine Learning Advances, 2026. Verfasst von Shehryar Ahmed und 3 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Structural designers need the fundamental natural frequency of thin rectangular plates, but even small changes in geometry, material, or edge restraint typically require new eigenvalue or finite-element analysis. This study instead builds six supervised machine-learning surrogates to predict the nondimensional first-mode frequency parameter of thin plates with complex boundary conditions. The training data is produced using the finite integral transform method. The data set formations across seven input variables. Six machine-learning models including linear regression, support vector regression, random forest, gradient boosting, XGBoost, and LightGBM, were trained and evaluated using a held-out test set. Gradient boosting out performed (R² = 0.966), with random forest and LightGBM close behind at 0.944 and 0.939; linear regression, support vector regression, and XGBoost lagged at 0.927, 0.911, and 0.891. The SHAP, SHAP-interaction, and Shapash analyses reliably recognized the plate aspect ratio and the two rotational-restraint coefficients as the dominant predictors, while elastic stiffness and density mattered comparatively little. When tested on restraint-coefficient combinations never seen during training, the surrogate tracked the Finite Integral Transform solution to within 2.3% mean absolute percentage error, evidence that a properly trained machine-learning model can stand in for nonclassical plate-vibration analysis in rapid parametric studies and preliminary design, without sacrificing much accuracy.

Zitieren

APA

Ahmed, S., Sudheer, M., Ullah, T., & Ullah, K. (2026). A Data-Driven Computational Framework for Predicting the Fundamental Natural Frequency of Thin Rectangular Plates. Journal of Machine Learning Advances. https://doi.org/10.66879/jmla.2026.64

Vancouver

Ahmed S, Sudheer M, Ullah T, Ullah K. A Data-Driven Computational Framework for Predicting the Fundamental Natural Frequency of Thin Rectangular Plates. Journal of Machine Learning Advances. 2026. doi:10.66879/jmla.2026.64

BibTeX
@article{Ahmed2026,
  author = {Ahmed, Shehryar and Sudheer, Mehran and Ullah, Tahsin and Ullah, Kefayat},
  title = {A Data-Driven Computational Framework for Predicting the Fundamental Natural Frequency of Thin Rectangular Plates},
  journal = {Journal of Machine Learning Advances},
  year = {2026},
  doi = {10.66879/jmla.2026.64}
}
Kostenlos starten →