Predicting the Fiber diameter of Spunbonding Nonwovens Via Empirical Statistical methods and Neural Network Model

Abstract

In this paper, the empirical statistical and artificial neural network methods are established. We present a comparative study of two modeling methodological for predicting the fiber diameter of spunbonding nonwovens from the process parameters. The radial basis neural network, which has good approximation capability and fast convergence rate, is employed in this work, and it can provide quantitative predictions of fiber diameter. The effects of process parameters on fiber diameter are also determined by the ANN model. The results show the artificial neural network model yield more accurate and stable predictions than the statistical method, which reveals that artificial neural network technique is really an effective and viable modeling method.