Analysis and visualization of aluminum reduction cell noise based on wavelet transform

Abstract

The unsteady part of a cell pseudo-resistance signal called noise is closely related to electro-chemical and magneto-hydrodynamics processes, which is consequently reflected in its non-stationary and transient character, particularly in cases of different disturbances. Therefore wavelet transformation has been introduced in order to examine cell noise simultaneously within time-frequency space, thus enabling a deeper view of the process. A special data acquisition system has been installed on the API8 Pechiney cell, with the capability of conducting continuous measurements of the cell's operating data, performing wavelet analysis and evaluating a result using the data obtained from the plant control system. In this paper the analysis and visualization of cell resistance noise, based on continuous wavelet transformation and wavelet packet transformation is proposed, in order to obtain complementary data for more efficient process control. The research results showed promising potential for using the wavelets for the visualization of dynamics and the detections of different disturbances during the aluminum reduction process.