Comparison of the grey theory with neural network in the rigidity prediction of linear motion guide

The purpose of this paper is to compare the prediction models constructed through neural network and grey theory, and to apply the prediction model established to study of correlation between linear motion guide rigidity under the stress of tension and compression. Strain data of tension and compression are simultaneously obtained by the computer that is linked with the Universal testing machine and translated into rigidity values through the formula of δ k F = . Through this study we can understand the differences in prediction of rigidity between neural network and grey theory. Experiment results will serve as reference for manufacturers and users, with the hope that based on fewer measurement data testing time can be reduced and the outcome can be more accurately predicted. Based on fewer measurement data, the outcome can be more accurately predicted, and that with a nondestructive test can accurately predict the rigidity of the linear motion guide. The outcome indicates that the prediction model established through neural network is superior to the prediction model established through the grey theory, and that the neural network model can accurately predict the result. Key-words: grey prediction model;rigidity;linear motion guide;neural network;tension;compression