Abstract
The application of existing theoretical models for the computation of magnetic and hydraulic forces in a real oriented matrix consisting of regularly arranged rods and wires indicates that these models produce no exact results. The differences between computations and measurements of force effects documented by Maxwell lead to the conclusion that it is necessary to start with different physical assumptions when modelling a high–gradient separation process. First of all, the magnetic field of the rods or wires system differs from the magnetic field of a single rod. Second, the particle need not be attracted to the rod surface, it is brought there by the suspension stream and the magnetic force must hold it, so that it is not entrained by the streaming suspension. As the layer of attracted particles grows, the magnetic attractive force on the surface of the growing layer decreases until the magnetic attractive force is in equilibrium with the entraining force of suspension flow.