Chaoyang Zheng, Ye Liu, Yi Liao, Wenhai Yu, Feng Xiong
{"title":"Flexural behavior of bolted precast concrete floors at the serviceability limit state","authors":"Chaoyang Zheng, Ye Liu, Yi Liao, Wenhai Yu, Feng Xiong","doi":"10.1016/j.jobe.2025.113215","DOIUrl":null,"url":null,"abstract":"The traditional wet connection strategy in precast concrete floor systems impedes the further improvement of construction speed and the replacement of components when damaged or their recycling after service life. A novel bolted precast concrete floor has recently been proposed to cater the increasing demand for both rapid assembly and demountable features in precast structures. Excessive deformations that exceed expectations resulting in aesthetic defects and economic losses underscore the importance of accurately predicting the serviceability limit state for floors. Flexural tests were conducted to investigate the flexural behavior of the bolted precast floors up to the serviceability limit state. The tested bolted floors exhibited two-way flexural behavior, which was specifically reflected in similar cracking patterns and displacement distributions compared to those of monolithic two-way slabs. The combination of bolted double-steel-plate (BD) joint connectors and bolted angle (BA) anchorage connectors realized a 17.1 %–26.4 % improvement in the floors’ serviceability limit load. A simplified general simulation strategy for bolted connectors was employed in the finite element (FE) study, utilizing built-in connector elements to accommodate different types of bolted connectors, which enhanced the general applicability of the FE models. Two machine learning (ML) models were established to assess their feasibility of predicting the serviceability limit load of the bolted floors. Both ML models demonstrated favorable predictive performance, with their R<ce:sup loc=\"post\">2</ce:sup> exceeding 0.95 and MAPE under 0.025. Furthermore, an explainable model was created by combining SHapley Additive exPlanations (SHAP) analysis with the ML models. Through the interpretation of the model, the effect of input features on the serviceability limit load was studied at both global and local levels.","PeriodicalId":15064,"journal":{"name":"Journal of building engineering","volume":"16 1","pages":""},"PeriodicalIF":7.4000,"publicationDate":"2025-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of building engineering","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1016/j.jobe.2025.113215","RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"CONSTRUCTION & BUILDING TECHNOLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
The traditional wet connection strategy in precast concrete floor systems impedes the further improvement of construction speed and the replacement of components when damaged or their recycling after service life. A novel bolted precast concrete floor has recently been proposed to cater the increasing demand for both rapid assembly and demountable features in precast structures. Excessive deformations that exceed expectations resulting in aesthetic defects and economic losses underscore the importance of accurately predicting the serviceability limit state for floors. Flexural tests were conducted to investigate the flexural behavior of the bolted precast floors up to the serviceability limit state. The tested bolted floors exhibited two-way flexural behavior, which was specifically reflected in similar cracking patterns and displacement distributions compared to those of monolithic two-way slabs. The combination of bolted double-steel-plate (BD) joint connectors and bolted angle (BA) anchorage connectors realized a 17.1 %–26.4 % improvement in the floors’ serviceability limit load. A simplified general simulation strategy for bolted connectors was employed in the finite element (FE) study, utilizing built-in connector elements to accommodate different types of bolted connectors, which enhanced the general applicability of the FE models. Two machine learning (ML) models were established to assess their feasibility of predicting the serviceability limit load of the bolted floors. Both ML models demonstrated favorable predictive performance, with their R2 exceeding 0.95 and MAPE under 0.025. Furthermore, an explainable model was created by combining SHapley Additive exPlanations (SHAP) analysis with the ML models. Through the interpretation of the model, the effect of input features on the serviceability limit load was studied at both global and local levels.
期刊介绍:
The Journal of Building Engineering is an interdisciplinary journal that covers all aspects of science and technology concerned with the whole life cycle of the built environment; from the design phase through to construction, operation, performance, maintenance and its deterioration.