Knowledge of parameters inside horizontal rotating drums is useful for process optimization and control in industrial applications including milling, agglomeration and coating. However, due to the harsh internal environment found in these units the internal parameters cannot be measured directly. Instead the parameters can be inferred from external measurements with the use of a soft sensor.

The intensity of impacts inside a drum is an internal parameter particularly useful in milling processes, where impact intensity dictates the rate of particle breakage. This paper illustrates how impact intensity data can be predicted by a discrete element method (DEM) model of a horizontal rotating drum, and the impact intensity quantitatively estimated from external measurements using a soft sensor created with multivariate statistical methods. The results will be verified with small-scale experiments.