GFRP wrapped concrete column compressive strength prediction through neural network

Wrapping of fibre over the concrete columns significantly improves the compressive strength and ductility behaviour of columns. In this study the compressive strength of the Glass Fibre Reinforced Concrete (GFRP) wrapped concrete columns were studied under axial compression. The parameters varied are types of GFRP (Surface Mat, Chopped Strand Mat and Woven Roving Mat), number of plies (0, 1 ply and 3 plies) and period of curing (7, 14 and 28 days). Twenty-one standard cylindrical specimens were tested to failure under compression testing machine. The GFRP confined concrete columns increases the compressive strength by 30% for different types of fibres. The change in the number of plies from one to three improves the strength six times. The Artificial Neural Network (ANN) is an alternate tool used to accurately estimate the confined strength of the wrapped columns. The multi-layer neural network has been used with back propagation training algorithm. The predicted strength was compared with the experimental results. A good correlation was obtained between the predicted strength by ANN model and experimental values with correlation coefficient R values of 0.992, 0.999 and 0.999 for training, validation and testing data sets.