Abstract:

In order to get the optimal technological parameters of manufacturing meltblown nonwovens, the BP neural networks prediction model was established by using the first four data of every group as training input and output vectors in the experiment, while others as testing vectors to check the model. The results showed that the relative errors were less than 9% when the BP neural networks model was applied to predict the average values of filtration efficiency and permeability, and the correlation coefficients were both close to 1. It showed that the fitting effect of the BP networks model was very ideal. On this basis, the optimal technological parameters were obtained by computer simulation, that is, temperature, 218 ℃; DCD, 18 cm, and surface density of 469 g/m2. Under these conditions, the filtration efficiency reached 94% and the air permeability was above 400 L/(m2•s) .