Research on recognition method of liner wear stage based on operation and vibration information of semi-autogenous mil

Abstract

The wear stage of the liner is of great significance to the production and maintenance of the mill. However, it is difficult to monitor the wear stage of the liner in real time. Therefore, this study developed a method to recognize the wear stage of semi-autogenous(SAG) mill liner on line. The pulp density is an important characteristic variable for reflecting liner wear stage. In view of the difficulty of on-line detection, this paper presents a method to calculate the pulp density of SAG mill under the condition that the product fineness index is qualified and the filling level of the internal grinding medium(steel ball) is stable. On account of the characteristics of nonlinear and non-stable significant changes in mill vibration information, this paper uses mode decomposition method(EMD, VMD) and principal component analysis (PCA) method to screen and reduce for the important vibration modal characteristics. Combined with the important operating parameters of the mill (feed ore and water), multiple groups of deep learning algorithms based on the recognition model of liner wear stage are established. In response to the actual need to recognize the stage of severe liner wear, an accurate recognition method of critical interval of liner wear stage is proposed to achieve accurate recognition of severe wear stage of liner based on the performance analysis results of several liner wear stage recognition models.