Study On The Theoretical Permeability Of Needle-Punched Geotextiles

During the manufacture process of needle-punched geotextile, a thin fiber-net is formed by air vortex or physical combing. Then net layers are stacked randomly or unidirectional. Finally, the layers are tangled with needles by needle-punched machine and to be tight. Therefore, in order to create the models to estimate the permeability of needle-punched nonwoven, we should consider fibers which are distributed not only on fabric surface, but also along fabric's thickness direction. In this paper, three kinds of models depended on different theories were established to predict the permeability of needle-punched geotextile, that is, fractal model, random network model and drag-froce model. Comparison was made between experimental and theoretical results to verify the accuracy of the four models, and also made between the models created in this paper and existing models. Finally, the applicability of every model were presented, respectively.In fractal model, the effective porosity, surface fractal dimension and tortuous fractal dimension were obtained by using image analyze technology, and then applied the capillary water transport knowledge to permeability model. For the thin geotextiles, the theorecital results have quite good agreement with experimental ones. For thick geotextiles, the error is mainly caused by our presupposition that the fractal model only analyze pores distributing in the geotextile surface. Besides, there were limitations in image analyze process such as the depth of field, resolution, binarization.Random network model assumed that the pores inside fabric were tortuous capillaries with different diameters. And the possibility of pores formed by crossed fibers on the surface met Poisson Polyhedron distribution. In this paper, the possibility of pore distribution was cooperated with fractal theory and image-analyze technology. The result turned out that Random network model can predict permeability well when geotextile had low thickness. However, just like fractal model, the random network model ignored the distribution of fibers in the deeper thickness. So this model was more likely suit for low-thickness nonwovens.Drag-force model was based upon N.Mao and S.J.Russell's model. It presumed fibers were isotropic in the plane of fabric, but along the thickness direction (Z axis) fibers surround the punch hole with open-package way. The result of predict made by drag model shown that it was a little higher than experiment's. Because fluid had less resistance when flow along with the fiber than the resistance when flow against facade of fiber. However, the real structure of needle-punched nonwoven was more complicated than drag model's presumption. There were oblique fibers in real geotextile structure which make the result of predications larger. On the other hand, the drag model was fit for needle-punched geotextile with large thickness.The innovation of this paper were (1) the image processing program was optimized in the establishment of fractal model, so the program can automatically determine the selected binary threshold; (2) the fabric pore structure developed from Poisson Polyhedron theory was combined with the surface pore morphology of fabric created by image processing technology in order to improve the model practicability when established the random network model; (3) the fiber distribution model in perpendicular direction was developed to improve the accuracy of drag-force model.This solution can get as much information of fabric as possible, and made the model get closer to real condition. Those models which were provided to predict the permeability of needle-punched geotextile can get rid of time-consuming lab tests when geotextile was used for drainage system.