Predictive model based on artificial neural network for fatigue performance of prior-corroded aluminum alloys

A prediction model for corrosion and fatigue performance of the prior-corroded aluminum alloys under a varied corrosion environmental spectrum based on artificial neural network was developed and the nonlinear relationship among maximum corrosion depth, fatigue performance, corrosion temperature and time was established, based on BP (back propagation) learning algorithm analysis and convergence improvement. The maximum corrosion depth and fatigue performance of prior-corroded aluminum alloys can be predicted by means of the trained neural network from the testing data. By virtue of the prediction model, the future corrosion status and fatigue performance of aluminum alloys can be evaluated under random complicated environmental spectrum.