In process industries, input fluctuations of essential raw material characteristics, e.g. particle size distribution, chemical and mineralogical composition, moisture, ash content, calorific value, etc., can greatly influence the efficiency of separation and other processes in particle technology. Furthermore, the efficiency of separation processes can decrease in the course of the starting-up period due to instationary states and transient phenomina. On the basis of the general model for mechanical macro-processes, the following examples are treated and compared with practical results and experience. (1) The transfer of stochastic fluctuations through a process unit; in this case, the dynamic behaviour is described by (a) the dispersion model and (b) the model of a cascade of ideally mixed vessels. As a result, the reduction of fluctuations (characterized, for example, by means of the standard deviation) can be calculated. (2) Applying the general model for mechanical macroprocesses to a special separation process (turbulent cross-flow classifier, pulp division model) it is possible to show that instationary states impair the separation efficiency, for example, to the separation sharpness derived from the grade efficiency curve (Tromp distribution curve).