Shear strength prediction of FRP reinforced concrete members using generalized regression neural network

• DOI: 10.1007/s00521-019-04107-x
• Publication Date: 2019-02-28T02:58:39Z
• Authors: Md Shah Alam; Uneb Gazder
• Keywords: 0211 other engineering and technologies; 02 engineering and technology

The behavior of FRP reinforced members in shear differs from that of steel reinforced members. Consequently, the design equations are continually changing and becoming more intricate. This paper presents a model based on generalized regression neural network (GRNN) for the predictions of shear strength of FRP reinforced concrete members with no transverse reinforcement. A database of 196 test specimens, failed in shear, is used to train and test the GRNN model. The results of training and testing set of the database are compared with the experimental results. It was observed that GRNN proves to be an effective method for predicting the shear strength of FRP reinforced concrete members without shear reinforcement. The database is also used to assess the accuracy of ACI 440.1R, CSA S806, JSCE, and BISE shear design procedures and to compare their predictions with the GRNN model. The GRNN model prediction shows more consistent and less scattered results in contrast to the four design codes and guidelines.