Data-driven prediction and optimization of fire performance for cold-formed steel walls with board joints

Journal of Building Engineering, Год журнала: 2025, Номер unknown, С. 112797 - 112797

Опубликована: Апрель 1, 2025

Язык: Английский

---

Hybrid machine learning approach with FHO algorithm and WERCS method for predicting fire resistance of timber columns

Materials Today Communications, Год журнала: 2025, Номер unknown, С. 112679 - 112679

Опубликована: Апрель 1, 2025

Язык: Английский

---

Fire resistance rating prediction of timber-to-steel connections and design optimization informed by explainable machine learning

Engineering Applications of Artificial Intelligence, Год журнала: 2025, Номер 156, С. 111127 - 111127

Опубликована: Май 29, 2025

Язык: Английский

---

Efficiency and explainability of design‐oriented machine learning models to estimate seismic response, fragility, and loss of a steel building inventory

Earthquake Engineering & Structural Dynamics, Год журнала: 2024, Номер unknown

Опубликована: Ноя. 20, 2024

Abstract Machine learning (ML) has recently been used as an efficient surrogate to estimate different steps of performance‐based earthquake engineering (PBEE), from dynamic structural analysis fragility and loss assessments. However, due the varied data, models, features in existing literature, relative efficiency ML models across PBEE remains unclear. Additionally, black‐box nature advanced algorithms limits their ability provide design‐oriented insights, hindering broader application PBEE‐based design. This study provides a comprehensive comparison accuracy explainability using consistent database 621 steel moment frames with varying designs geometry. Eight were careful training workflow comprising feature selection, hyperparameter tuning, cross‐validation, model inference. The sensitivity representative outputs—maximum responses, median fragility, expected annual loss—was assessed statistical measures. In addition, best for each step was examined explore relationship between design parameters corresponding output. results show that while can reasonably map all outputs, higher drift fragilities, component‐based metrics. optimal algorithm remained same steps, where support vector machines random forests provided highest average R 2 0.93 0.91 over outputs on test set. Although selected sets algorithms, height, number stories, fundamental period, minimum beams’ inertia influential both notably affected outputs.

Язык: Английский

---

Reliability analysis of timber columns under fire load using numerical models with equivalent section temperature

Engineering Structures, Год журнала: 2024, Номер 324, С. 119345 - 119345

Опубликована: Ноя. 27, 2024

Язык: Английский

---