Zhenfeng Peng
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Zhenfeng Peng has a total of 5 co-authors. Here are the top co-authors by number of shared publications:
- Han Wu
School of Infrastructure Engineering, Nanchang University, No. 999 Xuefu Avenue, Nanchang 330047, China
1 shared publication - Hua Huang
China Construction Fourth Engineering Division Corp. Ltd., Guangzhou 511400, China
1 shared publication - Jinkun Hou
China Construction Fourth Engineering Division Corp. Ltd., Guangzhou 511400, China
1 shared publication - Xudong Zheng
China Construction Fourth Engineering Division Corp. Ltd., Guangzhou 511400, China
1 shared publication - Yuxi Ding
School of Infrastructure Engineering, Nanchang University, Nanchang 330031, China
1 shared publication
Latest Publications
- Open Access
Study on the Ultimate Load-Bearing Capacity of Disc Buckle Tall Formwork Support Considering Uncertain Factors
Disc buckle steel pipe brackets are widely used in building construction due to the advantages of its simple structure, large-bearing capacity, rapid assembling and disassembling, and strong versatility. In complex construction projects, the uncertainties affecting the stability of disc buckle steel pipe support need to be considered to ensure the safety of disc buckle steel pipe supports. A surrogate model based on a deep neural network is built and trained to predict the ultimate load-carrying capacity of a stent. The results of the finite element model calculations are used to form the sample set of the surrogate model. Then, we combined the computationally efficient DNN surrogate model with the Monte Carlo method to consider the distribution of the ultimate load capacity of the disc buckle bracket under the uncertainties of the bracket node pin wedge tightness, the wall thickness of the steel pipe, and the connection of the connecting wall member. At the same time, based on the DNN model, the SHapley Additive exPlanations (SHAP) interpretability analysis method was used to study the degree of influence of various uncertainty factors on the ultimate bearing capacity of the stent. In practical engineering, the stability analysis of a disc buckle tall formwork support has shown that a surrogate model based on a deep neural network is efficient in predicting the buckling characteristic value of the support. The error rate of the prediction is less than 2%. The buckling characteristic values of the bracket vary in the range of 17–25. Among the various factors that influence the buckling characteristic value of the bracket, the joint wedge tightness has the greatest impact, followed by the bottom and top wall-connecting parts.
Open AccessStudy on the Ultimate Load-Bearing Capacity of Disc Buckle Tall Formwork Support Considering Uncertain Factors
Disc buckle steel pipe brackets are widely used in building construction due to the advantages of its simple structure, large-bearing capacity, rapid assembling and disassembling, and strong versatility. In complex construction projects, the uncertainties affecting the stability of disc buckle steel pipe support need to be considered to ensure the safety of disc buckle steel pipe supports. A surrogate model based on a deep neural network is built and trained to predict the ultimate load-carrying capacity of a stent. The results of the finite element model calculations are used to form the sample set of the surrogate model. Then, we combined the computationally efficient DNN surrogate model with the Monte Carlo method to consider the distribution of the ultimate load capacity of the disc buckle bracket under the uncertainties of the bracket node pin wedge tightness, the wall thickness of the steel pipe, and the connection of the connecting wall member. At the same time, based on the DNN model, the SHapley Additive exPlanations (SHAP) interpretability analysis method was used to study the degree of influence of various uncertainty factors on the ultimate bearing capacity of the stent. In practical engineering, the stability analysis of a disc buckle tall formwork support has shown that a surrogate model based on a deep neural network is efficient in predicting the buckling characteristic value of the support. The error rate of the prediction is less than 2%. The buckling characteristic values of the bracket vary in the range of 17–25. Among the various factors that influence the buckling characteristic value of the bracket, the joint wedge tightness has the greatest impact, followed by the bottom and top wall-connecting parts.