Abstract: For the problem of humidity, temperature, PH value and their interaction can accelerate the aging of silk fabrics, the BP neural network model can predict the mechanical properties of silk fabrics in the different environmental factors. First, using L9(34) orthogonal experimental to design the interactive environment and begin the aging treatment of silk fibers for about 24h. Second, using the type 3365 Instron universal material testing machine to test the breaking strength of silk fabrics, and using the breaking strength value of silk fibers as the BP neural network training sample, then establish three-layer neural network to train and predict. The result show that the relative errors of the training output values and the actual experimental values are less than 0.005%, and BP neural network can predict the mechanical properties of silk fabrics effectively.

Key words: BP neural network, silk fabric, environmental factors, mechanical property