Time-frequency analysis method and equipment for normalized cell voltage data for aluminum electrolysis

Abstract

The invention discloses a method for time-frequency analysis of normalized cell voltage data for aluminum electrolysis. The method includes: performing time-frequency analysis on historical data of normalized cell voltage, and using energy spectral density on the basis of a scale map to obtain The frequency range corresponding to the alumina concentration and the frequency division of low-frequency tank noise and its frequency bandwidth related to tank stability. Implement the present invention, not only can overcome the above-mentioned deficiencies and defects of the prior art, and the normalization cell voltage that the present invention obtains has included multiple signal components and their corresponding frequency ranges, compared with the prior art, this frequency range is more narrow, the distribution of related signal components is more precise, and also covers the sub-low frequency noise caused by other factors not covered by the prior art.

Description

technical field [0001] The invention relates to the field of aluminum electrolytic cell process control, in particular to a method and equipment for time-frequency analysis of normalized cell voltage data for aluminum electrolysis. Background technique [0002] Cryolite-alumina molten salt electrolysis is the only way for the aluminum electrolysis industry to produce primary aluminum, and the aluminum electrolysis cell is the "heart" of the aluminum electrolysis production process. The raw material used in electrolysis is alumina, the electrolyte is molten cryolite, and a carbon anode is used. The result of electrolysis is molten aluminum on the cathode and carbon dioxide on the anode. Due to the complex electrochemical, physical and chemical reactions involved in the aluminum electrolytic cell, the cryolite-alumina molten salt has the characteristics of high temperature and strong corrosion, so many key parameters cannot be directly measured. At the same time, the "five l...

AI Technical Summary

Problems solved by technology

Therefore, the use of spectrum analysis cannot analyze the frequency composition and change law of the cell voltage signal well, and cannot clearly distinguish the correspondence between the components of the cell voltage signal and their physical meanings, which in turn affects the signal components related to the alumina concentration state information and cell noise. Accuracy of Acquiring Work[4]

[0009] In terms of the acquisition of alumina concentration state information and groove noise signal components, due to the limited ability of inertial filters, Kalman filters and Butterworth filters to deal with non-stationary signals, they are designed by their cascade method[4,5] The high-pass, band-pass and low-pass filters formed cannot well separate the information components of each frequency band in the tank voltage signal, which in turn affects the accuracy and real-time performance of key control parameters for alumina concentration control and tank stability

[0010] In summary, the existing cell voltage analysis method based on the combination of spectrum analysis and multi-filter cascading can obtain the alumina concentration state information and the corresponding frequency band of the cell noise, so that the separated from the cell voltage signal and alumina Concentration-related state information still contains a lot of low-frequency trough noise, and the signal component separation methods designed based on filter cascades such as inertial filters, Kalman filters, or Butterworth filters have large gaps in the selection of parameters. Subjectivity affects the accuracy and real-time performance of key parameters in alumina concentration control and tank stability control, which in turn affects the performance of the tank control machine control system

Images