A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Prediction of load-carrying capacity of adhesive anchors by soft computing techniques
Abstract The paper presented herein aims at providing explicit formulations for predicting the load-carrying capacity of the single adhesive anchors post-installed into uncracked hardened concrete. For this purpose, the worldwide database for the adhesive anchors compiled by the ACI Committee 355 was obtained and reduced to generate the training and the testing data sets to construct the closed-form solutions by means of the soft computing techniques such as Neural Network (NN) and Genetic Programming (GEP). Therefore, the NN and GEP models were developed with a correlation coefficient of as high as 0.99 and 0.95, respectively. Moreover, the mean absolute percentages of errors for the proposed models were fairly reasonable considering the noisy nature of the database. The prediction performance of the developed models improved with increase in the embedment depth, anchor diameter, and the concrete strength. The ratio of predicted to observed values for the developed NN and GEP models were much smaller than that of the well known Concrete Capacity Design (CCD) method which generally overpredicted the experimental values.
Prediction of load-carrying capacity of adhesive anchors by soft computing techniques
Abstract The paper presented herein aims at providing explicit formulations for predicting the load-carrying capacity of the single adhesive anchors post-installed into uncracked hardened concrete. For this purpose, the worldwide database for the adhesive anchors compiled by the ACI Committee 355 was obtained and reduced to generate the training and the testing data sets to construct the closed-form solutions by means of the soft computing techniques such as Neural Network (NN) and Genetic Programming (GEP). Therefore, the NN and GEP models were developed with a correlation coefficient of as high as 0.99 and 0.95, respectively. Moreover, the mean absolute percentages of errors for the proposed models were fairly reasonable considering the noisy nature of the database. The prediction performance of the developed models improved with increase in the embedment depth, anchor diameter, and the concrete strength. The ratio of predicted to observed values for the developed NN and GEP models were much smaller than that of the well known Concrete Capacity Design (CCD) method which generally overpredicted the experimental values.
Prediction of load-carrying capacity of adhesive anchors by soft computing techniques
Gesoğlu, Mehmet (author) / Güneyisi, Erhan (author)
2007
Article (Journal)
Electronic Resource
English
Prediction of load-carrying capacity of adhesive anchors by soft computing techniques
| British Library Online Contents | 2007
Prediction of load-carrying capacity of adhesive anchors by soft computing techniques
| Online Contents | 2007
Prediction of load-carrying capacity of adhesive anchors by soft computing techniques
| Springer Verlag | 2007
Prediction of load-carrying capacity of adhesive anchors by soft computing techniques
| Online Contents | 2007
Prediction of load-carrying capacity of adhesive anchors by soft computing techniques
| Online Contents | 2007