A neural network model is proposed to predict the crack growth and remaining life of corroded RC beams. Input and output vectors of the neural network are designed on the basis of fracture mechanics concepts. Corrosion effect has been accounted for in the form of reduction in the diameter and modulus of elasticity of reinforcement. Stress intensity factor (SIF) has been computed by using the principle of superposition. At each incremental crack length, net SIF has been computed as the difference of SIF of plain concrete and reinforcement. Further, remaining life has been predicted for corroded beams.

Keywords: Neural network, Plain concrete, Reinforced concrete, Stress intensity factor, Crack growth, Remaining life, Corrosion.