Application Of Neural Network In Research For Cushioning Property Of Honeycomb Paperboard

This paper mainly studies the application of neural network in research for cushioning property of honeycomb paperboard.At first, the traditional research method for cushioning property of honeycomb paperboard and the theory of neural network are presented, which indicates the impropriety of the traditional method. And the feasibility and necessity for the application of neural network in research for cushioning property of honeycomb paperboard are also discussed. Secondly, based on the experimental data from the dynamic compressing experiment and vibration transfer experiment, the Back Propagation neural network model, which can reflect the cushioning property of honeycomb paperboard is constructed. When the result of the Back Propagation neural network model is compared with that of the traditional approach-polynomial fit, the conclusion that the neural network method is better is drawn. At last, using the GUI(Graphical User Interface) of MATLAB6. 5, the software of the cushioning property of honeycomb paperboard research system based on the Back Propagation neural network is programmed, which can construct the corresponding Back Propagation neuralnetwork according to user' s experimental data. Furthermore, the maximal acceleration-static stress curves and the vibration transfer rate curves can be drawn, and the coordinates of every point in curves can also be displayed. This software has a friendly interface and easy to operate, and it' s a new tool for the researchers of honeycomb paperboard.