Neural network for structural stress concentration factors in reliability-based optimization

Stress concentration damage is one of the most troublesome problems in practical engineering structures. In order to reduce damage, it is very important to determine stress concentration factors accurately. An artificial neural network (NN) technique has been used in this article to simulate the relationship between basic random variables and stress concentration factors, and the explicit expression of the stress concentration factors can be obtained directly. The expression of the stress concentration and the reliability theory have been combined to solve the structural reliability problems; then, the corresponding methods of reliability analysis and optimization design have been presented.