Optimization of superplastic forming parameters of railway bearing cage and forming test

The superplastic tension test of cast Aluminum bronze was performed and the test results were used as samples to establish the artificial neural network (ANN) model to predict the superplastic properties of cast Aluminum bronze using Levenberg-Marquardt algorithm. Based on the model, the superplastic processing parameters of the materials were optimized and the optimal parameters were achieved. According to the parameters, the test of superplastic extrusion of the railway bearing cage was performed. Results show that the model reflected well the relationship between the superplastic properties and the superplastic tension conditions. The less error between the test value and the predicted output of the network demonstrates that it is feasible and effective using ANN to predict the superplastic properties of cast Aluminum bronze. The parameters conform well to the superplastic extrusion of the cage. And the processing has remarkable economic benefits under the optimal superplastic condition.