Scholarlink

<p>The article presents a comprehensive review and artificial neural network (ANN) analysis of empirical methods for the estimation of the fundamental period of vibration in multistory reinforced concrete moment resisting frames. Nine international building code-based empirical equations and existing research studies are evaluated for accuracy using a large dataset of building designs with varying heights, widths, number of bays, and weights. Statistical comparison reveals that code-based formulations always under-predict periods with the error increasing for taller structures, while a number of research-based formulations over-predict periods significantly. Extensive parametric analyses confirm building height as the primary controlling parameter but also confirm building width as the second most significant parameter, although overlooked in all code formulations. Number of bays and overall building weight exhibit smaller but significant influence on the fundamental period. The negative correlation between building width and period was a power function, with wider buildings having shorter periods due to increased lateral stiffness. An improved ANN model with feed-forward backpropagation architecture was generated and performed extremely well compared to all the empirical equations for a large variety of building geometries. The ANN method captured coupled nonlinear relationships between more than one structural parameter more accurately than traditional power-law relationships. With this ANN model, a new simplified equation was formulated involving height, width, weight, and number of bays and proved to be very precise with the benefits of practical convenience in design application. The study provides detailed recommendations for practical implementation, including a design stage selection technique and guidelines for building code implementation. The outcomes provide useful information for seismic design practice and building code preparation to enable more economic and precise designing of RC moment resisting frames. The limitations of current analysis and future research are also discussed, including potential ANN model improvements, experimental validation, and advanced machine learning techniques.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A comprehensive review and ANN analysis of empirical formulae for determining the fundamental period of Multi-Story RC moment resisting frames

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年7月30日

查看全文
  • Hamdy A. Elgohary

摘要

The article presents a comprehensive review and artificial neural network (ANN) analysis of empirical methods for the estimation of the fundamental period of vibration in multistory reinforced concrete moment resisting frames. Nine international building code-based empirical equations and existing research studies are evaluated for accuracy using a large dataset of building designs with varying heights, widths, number of bays, and weights. Statistical comparison reveals that code-based formulations always under-predict periods with the error increasing for taller structures, while a number of research-based formulations over-predict periods significantly. Extensive parametric analyses confirm building height as the primary controlling parameter but also confirm building width as the second most significant parameter, although overlooked in all code formulations. Number of bays and overall building weight exhibit smaller but significant influence on the fundamental period. The negative correlation between building width and period was a power function, with wider buildings having shorter periods due to increased lateral stiffness. An improved ANN model with feed-forward backpropagation architecture was generated and performed extremely well compared to all the empirical equations for a large variety of building geometries. The ANN method captured coupled nonlinear relationships between more than one structural parameter more accurately than traditional power-law relationships. With this ANN model, a new simplified equation was formulated involving height, width, weight, and number of bays and proved to be very precise with the benefits of practical convenience in design application. The study provides detailed recommendations for practical implementation, including a design stage selection technique and guidelines for building code implementation. The outcomes provide useful information for seismic design practice and building code preparation to enable more economic and precise designing of RC moment resisting frames. The limitations of current analysis and future research are also discussed, including potential ANN model improvements, experimental validation, and advanced machine learning techniques.