Assessment of shear capacity of adhesive anchors for structures using neural network based model

Abstract

In this study, an artificial neural network (NN) based explicit formulation for predicting the edge breakout shear capacity of single adhesive anchors post-installed into concrete member was proposed. To this aim, a comprehensive experimental database of 98 specimens tested in shear was used to train and test NN model as well as to assess the accuracy of the existing equations given by American Concrete Institute and prestressed/precast concrete Institute. Moreover, the proposed NN model was compared with another existing model which had been derived from gene expression programming by the authors in a previous study. The prediction parameters utilized for derivation of the model were anchor diameter, type of anchor, edge distance, embedment depth, clear clearance of the anchor, type of chemical adhesive, method of injection of the chemical, and compressive strength of the concrete. The proposed model yielded correlation coefficients of 0.983 and 0.984 for training and testing data sets, respectively. It was found that the predictions obtained from NN agreed well with experimental observations, yielding approximately 5 % mean absolute percent error. Moreover, in comparison to the existing models, the proposed NN model had all of the predicted values in ±20 % error bands while the others estimated up to %160 error.