A particle-level simulation methodology is proposed to study the squeeze flow behavior of model magnetorheological fluids. The simulation algorithm takes into account Brownian motion and local field corrections to magnetic interactions of the particles. Simulation results obtained from using different initial configurations, including one single-particle-width chain per simulation box, random or lattice arrangements of preassembled single-particle-width chains as well as randomly dispersed particle suspensions, are compared with experimental data and predictions of a recently developed microscopic model. The assumption of single-particle-width chain structures in the systems has been shown to generate normal stresses larger than those found in experiments and the micromechanical model. However, much better agreement between the simulation and experimental results have been reached when using random initial configurations in the simulations.

An extensive experimental and simulation study is carried out in conventional magnetorheological fluids formulated by dispersion of mixtures of carbonyl iron particles having different sizes in Newtonian carriers. Apparent yield stress data are reported for a wide range of polydispersity indexes (PDI) from PDI = 1.63 to PDI = 3.31, which for a log-normal distribution corresponds to the standard deviation ranging from 38. 0   to 76. 0  . These results demonstrate that the effect of polydispersity is negligible in this range in spite of exhibiting very different microstructures. Experimental data in the magnetic saturation regime are in quantitative good agreement with particle-level simulations under the assumption of dipolar magnetostatic forces. The insensitivity of the yield stresses to the polydispersity can be understood from the interplay between the particle cluster size distribution and the packing density of particles inside the clusters.