Based on Spectral Information Fusion Soft Sensor Modeling of Wet Ball Mill Load

摘要

Ball mill load measurement is essential to control and operational optimization for the wet grinding process, which affects the production efficiency and energy consumption. A frequency spectral information fusion soft sensor model is proposed to estimation the mill load in the paper. PCA is used to extract the feature of frequency spectrum to deal with many, noise and collinear variables. FFT is used to estimate the power spectral density (PSD)of the vibration and acoustic signal. PLS were combined with PCA scores inputs to develop mill load. Principal component numbers are selected by an optimize model. A case study shows that the proposed frequency spectral information fusion soft sensor model is effective, and produces better predictive performance than single sensor model.