Multi-classification of ball mill load status using vibration frequency spectrum.

Lijie Zhao; Decheng Yuan

No Detail Found!

Multi-classification of ball mill load status using vibration frequency spectrum.

Author(s)

Lijie Zhao College of Information Engineering, Shenyang University of Chemical Technology, Shenyang 110142, Liaoning, China

Decheng Yuan College of Information Engineering, Shenyang University of Chemical Technology, Shenyang 110142, Liaoning, China

Abstract

On-line monitoring and recognition of mill load status has significant effect on the operating efficiency, product quality and energy consumption for the milling circuit. Due to low reliability to recognize the operating states near the boundary region, a multi- classification model is built to identify the operating status of ball mill load. Spectrum features of shell vibration signals are extracted using kernel principal component analysis as input of the multi-classification model. Partial least square-based extreme learning machine model predicts the output coding of ball mill load status. Bayesian decision theory further enhances the reliability and accuracy of the classification model. The proposed method is compared with one-against-one multi-classification strategy and verified with the experimental ball mill. Experimental results show the accuracy and stability of the proposed multi-classification model of ball mill load status outperforms multi-classification with one-against-one strategy.

Publication Details

Page(s) 694-701

DOI DOI not available

Published Journal: Journal of Theoretical and Applied Information Technology, Volume: 45, Issue: 2, Year: 2012

Request the full text from PASTIC

Citations

Downloads

Views