Transport in stochastic fibrous networks

Some fundamental issues concerning the design and performance of stochastic porous structures are examined, stemming from application of advanced fibrous electrode substrates in NiMH automotive cells. These electrodes must resist corrosion and local failures under hundreds of charge/discharge cycles. Such fibrous materials can be effectively used as substrates for chemical reactions because of their combinations of high surface area and high conductivity. Key questions concerning the relationships among connectivity and conductivity, scale and variability in material response are addressed. Two techniques are developed and compared for use in predicting these materials' conductivity. The first approach uses a statistical technique in conjunction with an adaptation of classic micromechanical models. The second approach uses the statistical generation technique, followed by an exact calculation of 2D network conductivity. The two techniques are compared with one another and with classic results. Several important conclusions about the design of these materials are presented, including the importance of use of fibers with aspect ratios greater than at least 50, the weak effect of moderate alignment for unidirectional conductivity, and the weak power-law behavior of conductivity versus volume fraction over the range of possible behaviors.