MULTIVARIATE STATISTICAL-ANALYSIS OF VIBRATION SIGNALS FROM INDUSTRIAL-SCALE BALL GRINDING

Multivariate statistical modelling based on vibration signal analysis was performed at commercial scale grinding. The source digital signals consist of three channels of mechanical vibrations obtained at the ax ial, horizontal and vertical directions. The feed rate, power draw, pu lp temperature were collected automatically by the control system whil e samples of the feed material and ground product of the ball mill wer e manually taken to determine the particle size distributions and pulp densities. Using projection to the latent structure (PLS) and/or prin ciple component regression (PCR), empirical models between grinding pa rameters of interests and the vibration signals were built based on th e training data set collected in two weeks, thus the new grinding para meters could be automatically predicted whenever the vibration signals were known. The modelling results show that both the PCR and PLS mode l can be used to predict grinding parameters online.