Research on two fuzzy neural networks to predict mechanical properties of welded joints

Due to high nonlinear,complex interaction of many factors in welding process,it was difficult to predict the mechanical properties of welded joints. In this paper the adaptive neuro-fuzzy inference system(ANFIS) and the fuzzy radial basis function network model had been established based on TC4 titanium alloy in TIG welding to predicate the mechanical properties of welded joints. The welding process parameters were regarded as the input and mechanical property as output parameters of prediction models. 27 sets of experimental data were used to train the model and another 6 sets of experimental data were used to make simulation. The results showed that two fuzzy neural network models have high prediction accuracy and can be used to predict the mechanical properties of welded joints. But in terms of the structure,training speed,stability,generalization ability and reflection of the true situations of network model,the fuzzy RBF neural network is better than adaptive neuron-fuzzy neural network.