A Method for Anode Effect Prediction in Aluminum Electrolysis Cells Based on Multi-scale Time Series Modeling

The aluminumAluminum industry is moving toward intelligent and low-carbon development. Accurately predicting the anode effectAnode effect has always been a significant challenge in monitoring the aluminum electrolysisAluminum electrolysis process. However, due to the high-temperatureHigh-temperature and high-magnetic detection environment of aluminum electrolysisAluminum electrolysis cellCell, some critical parameters cannot be measured online. This inconsistency in data flow makes it challenging to apply traditional data-driven methodsMethod directly. In response to the characteristics of large data samples collected in actual production, we have proposed a multi-scale time series modelingModeling approach based on hybrid deep learningDeep learning. This methodMethod combines three advanced neural networkNeural networks models: BiLSTM, LSTM, and DNN. It enables the extraction of parameters that influence the anode effectsAnode effect from both short-term and long-term cyclic variables. Compared to traditional shallow machine learningMachine Learning (ML) methodsMethod, deep learningDeep learning methodsMethod, and hybrid learning methodsMethod, our proposed algorithm achieves the highest accuracy and F1 score, reaching 0.95 and 0.93, respectively. These results hold significant promise for reducing energyEnergy consumption and carbon emissions in actual production processes, paving the way for future applications.