GA-BP algorithm based on burst pressure prediction of steam generator tubing

In order to investigate the effect of flaws on the strength of steam generator (SG) tubes, a hybrid algorithm, GA-BP algorithm, was developed to establish a general model for predicting the burst pressure of Inconel 690 SG tubes. The genetic algorithm (GA) was used to initialize the weights of the back propagation artificial neural network (BP-ANN) to overcome the shortcomings of the BP algorithm such as slow convergence speed and local minimum. The model established was applied to explore the effect of the geometric parameters of flaws on the burst pressure of SG tubes. The simulation results show that the depth of the flaw has the most detrimental effect on the burst pressure. The burst pressure decreases rapidly with the increase in the depth of the flaw, and the burst pressure drops more clearly with the increase in the length of the flaw when the wrap angle of the flaw is small. In addition, the burst pressure of the SG tube with a deeper flaw decreases more significantly with the increase in the wrap angle.