A new class of structural composite materials has been developed whose members contain aligned fibers with aspect ratios on the order of 10^, diameters on the order of 10 microns, and volume fractions around 0.6. These materials contain a viscous, shear-thinning matrix, one with a melt viscosity above 300 Pa-s, and may experience a local shear rate of the order 10^ /sec under typical forming conditions. A controllable spectrum of fiber lengths is the dominant microstructural feature in these highly concentrated suspensions. A spectral decomposition method and supporting software were developed to determine the fiber length distribution from the real-time process data of a common fiber production method. Fiber array mechanical responses were measured for a range of process parameters and compared with predictions from a tangent tensile compliance model. Experimental evidence suggests that inter-fiber contacts greatly influence the transient tensile response. These contacts are believed to form an elastic shear network throughout the array which must be disrupted before relative fiber motion may occur. An impregnation model was formulated using a convecting finite difference technique to solve for the coupled fluid and porous media flows developed during calender impregnation. Fiber arrays were hot-mell impregnated both to verify the model and to produce towpreg for further testing. A unique, long-section furnace was constructed and hot tensile tests were performed on towpreg both to highlight the role of the previously studied fiber response and to provide specimens for the further study of array stability at large strains. A macroscopic test cell was constructed and tests were performed to examine, so called, “screening” of the transverse momentum transport in concentrated suspensions. Monte Carlo methods were employed to simulate the local fiber packing environment and, with the screening data, to improve on previous models of the steady tensile viscosity. A model combining the dry tensile compliance and stochastic steady viscous response was also constructed. A stochastic simulation of array linear density variability was combined with a simple steady-state viscosity model to explore tensile stability and compare with experiment.

Conclusions drawn from this work are that the array tensile stress is sufficient to result in local fiber fracture at typical forming rates, that mixed stress states will cause significant, but contra-intuitive changes in axial tensile response, that the initial linear density variability, particularly the distribution of shorter fibers, determines the maximum forming strain, and that gage-length and clamping effects cannot safely be ignored in future simulation and practice of the sheetforming of ALFA materials.