Research On The Regulation Of PFC Emission During The Non-anode Effect Process Of Aluminum Electrolysis Base On Wavelet Analysis Of The Cell Voltage

Hall-Héroult electrolysis process is used in industrial aluminum electrolysis,Normal electrolysis produces carbon dioxide,and PFC gas occurs when the anodic effect occurs,PFC has found when there is no anodic effect,is called non-anode effect process.PFC gases in non-anodic effect processes produce low tank voltages and low emission rates,which are difficult to identify during industrial aluminum electrolysis.In this paper,the PFC emission mechanism,the anode behavior of graphite electrode in cryolite melt,and the characteristic value extraction of slot voltage signal were studied in the Na₃Al F₆-Al₂O₃ electrolyte system,aiming to establish a diagnostic model of alumina concentration,so as to timely adjust the level of alumina concentration in the tank to reduce the generation of PFC in the process of non-anode effect.The study found that:（1）Using electrochemical methods,it is found that the discharge of fluorine ions on the surface of the carbon anode has nothing to do with the concentration of alumina in the tank,and the anode will also be passivated to varying degrees under the condition of high alumina solubility.In the cyclic voltammetry scan with Pt as the reference electrode,the alumina concentration<2.0 wt%,the carbon anode is passivated at about 2.4 V,the alumina concentration is>2.0 wt%,the peak current density of the cyclic voltammetry anode is shifted to 3.0 V,and the peak current density potential shifts to positive direction as the alumina concentration rises,and the peak current density also increases.The concentration of alumina in the tank>2.0 wt%,which can effectively inhibit the discharge of fluoride ions;（2）X-ray photoelectron method was used to study the different chemical valence structure and content of each element in the graphite anode surface product sample,and it was found that when the concentration of oxygen-containing complexed ions was low,fluoride ions were discharged on the surface of graphite anode to form C-F chemical bonds,and CF₄ gas was produced during the non-anode effect.（3）The spectral analysis method was used to study the molecular ratios of 2.2,2.4 and2.6,and the solubility of alumina was 0wt%,1.0 wt%,1.5 wt%,2.0 wt%and 3.0 wt%of the spectrum characteristics of the slot voltage signal,respectively,the results showed that the main frequency of the slot voltage signal was distributed within 0.5 Hz,and the slot voltage signal contained noise signal,the wavelet transformation method was used to reduce the noise of the slot voltage signal,and the local mean decomposition was used to extract the characteristics of the reduced slot voltage signal to obtain the product functions E₁ and E₂,E₃ and other 9 energy characteristic values,the extracted energy characteristic values were used to construct the characteristic signal corresponding to the solubility of alumina,and the results showed that the constructed characteristic signal could distinguish the solubility of alumina in different tank conditions.（4）Based on the characteristic signal extracted by the voltage of the anode groove,the BP neural network alumina solubility identification model is established.The results show that the accuracy of the training set of the BP neural network model is above 93.00%,the accuracy of the test set is above 90.00%,and the comprehensive correlation coefficient of the model is greater than 0.9.After the model is established,the newly acquired slot voltage is input into the corresponding BP neural network model recognition verification,and the verification results show that the recognition accuracy of the newly acquired slot voltage signal is above 85%.