Detection of abnormal alumina feed rate in aluminium electrolysis cells using state and parameter estimation

Aimed to improve accuracy rate and time advance of the Anode Effect (AE) prediction in aluminum electrolysis cell, a new prediction method by means of the Generalized Regression Neural Network (GRNN) is proposed. The structure and advantages of the GRNN are introduced, then modeling the anode effect system of aluminum electrolysis cell by means of system identification based on the GRNN. The structure of samples is analyzed emphatically in the process of modeling, and the influence of sample size to model prediction accuracy is analyzed by experiment. The prediction model based on GRNN is trained and tested by sufficient samples extracted from the production data of the 400kA aluminum electrolysis cell. It's proved that the accuracy rate of the AE prediction is more than 90% on average in advance 30 minutes. In addition, results tested for different aluminum electrolysis cells show that the method has extensive applicability while keep high prediction accuracy and its effective.

The paper addresses the basic mechanisms leading to the low-voltage anode effect (LVAE) as well as the high-voltage anode effect (HVAE) in aluminium electrolysis cells. The root cause of any anode effect is too low alumina concentration at the anode. By statistical treatment based on a Gaussian distribution of the alumina concentration in the bath, it was found that the onset of a HVAE takes place at higher average alumina concentration if there is a large standard deviation. By calculating the anode potential using a pragmatic numerical model, it was found that a LVAE can take place at a part of the anode as small as 0.001 m2, while the rest of the anode is at the normal anode reaction, which makes it extremely difficult to discover. It was suggested that the on-and-off current pattern observed during LVAE can be explained by alternation between blocking of the surface (AE) and normal electrolysis, induced by non-stationary mass transfer of alumina.KeywordsAnode effectPFCAlumina distributionStatisticsCOF2

The stable operation of the aluminum electrolysis cell is the basis for the safe and efficient production of the electrolytic aluminum industry, and the cell condition identification technology is an important means to ensure the normal operation of the aluminum electrolysis cell. The cell condition identification technology based on anode current signal has played an increasingly important role in identify and fine control large aluminum electrolysis cells. This paper proposes an improved temporal convolutional network, which uses the time characteristics of the current signal to classify the current sequence for cell condition identification. The classification result can help people identify and monitor the conditions of the electrolysis cell. In this paper, the proposed method is verified on the real current signal, which can be used to identify various cell conditions such as anode effect and anode sliding.

In this work, neural networks are built and trained to be used in a predictive control scheme for the aluminum electrolytic cell. An efficient control of the cell requires the knowledge of predicted future values of the decision variables in order to enable the standard (nonpredictive) control logic to take anticipated actions to prevent the anode effect, a destabilizing event occurring during cell operation. The networks are first trained on data obtained from a computer simulator of the cell prior to undergoing further on-line learning. Trained to predict the cell resistance and the resistance’s trend indicators, the networks are applied to the control of cells in different cell states, with a view to preventing anode effects, the latter being deliberately induced by reducing the alumina feed rate or reducing the feeding frequency and duration. Results show that, with neural-predictive control, anode effects can be avoided, which should result in increased thermal stability, decreased power consumption, and reduced fluoride emissions. Further, the on-line learning capacity of the networks offers a good perspective for their application to other complex industrial processes as well.

Aluminum electrolytic cell voltage signal is the most important real-time data in the process of aluminum electrolysis. Under normal circumstances, the traditional LMD decomposition of the original voltage signal containing noise in the aluminum electrolytic cell will produce a large error and affect the diagnosis of the failure of the aluminum electrolytic cell. In this paper, an improved LMD decomposition method is used to decompose the cell voltage signal based on wavelet packet denoising. At the same time, according to the energy analysis, the extracted features are classified. The accuracy and reliability of feature extraction can be improved, and the faults existing in the aluminum electrolytic cell can be accurately diagnosed.

The global warming potential of perfluorocarbons (PFCs) is much higher than that of CO2 in the aluminum reduction industry; hence, the release of PFCs is becoming increasingly regulated. Mitigation of PFC release by reducing anode effects (AEs) occurring during aluminum reduction is challenging. In this study, online anode current measurements were conducted at the anode beam of 400-kA cells, following which a simple local AE detection model was developed for AE prediction. A satisfactory AE true positive prediction rate of 80.3% was obtained in an industrial test. An independently controlled feeder strategy was then developed and applied to cells integrated with the local AE detection and prevention. The results show that the process provides an improved current efficiency with low direct current power consumption. Finally, the immense potential to reduce AE is confirmed with twinkling AEs and normal AEs, and daily AE duration decreased by 42.8%, 60%, and 56.9%, respectively.

Gas generated beneath anodes in aluminum electrolytic cells play an important role for the circulation of the bath, alumina mixing, and heat balance. Those bubbles cause an extra voltage drop, which is strongly affected by the amount and shape of the bubbles beneath anodes. Consequently, understanding the dynamic behavior of bubbles in aluminum electrolytic cells has been a major research focus worldwide in recent decades.

It is an important part in aluminum electrolysis production to control the anode effect (AE). Since there are some shortcomings in traditional methods of anode effect prediction in aluminum electrolysis, this paper combined two methods, the slope of cell resistance and learning vector quantization (LVQ) neural network, to predict anode effect. First of all, the first prediction of anode effect will be conducted based on the slope of cell resistance. Afterwards, the inaccurate data are supposed to be re-predicted. The second prediction consists of two steps, one is to estimate the power spectrum from the signal of cell resistance by means of periodogram, the other is to re-predict the anode effect with the LVQ neural network, since the energy of frequency bands are served as the input feature variables of neural network, so as to raise the accuracy of prediction. It turned out that the success rate of ten-minute in advance prediction for anode effect can be above 85%, though just cell resistance signal is studied.

Numerical simulations of the effects of different anode slot configurations on the gas–liquid multiscale flow characteristics in an aluminum electrolysis cell have been conducted based on a recently developed mathematical model. The results clearly showed that use of anode slots can significantly promote the bubble evacuation behavior and thus affect the overall flow pattern. Both the gas volume fraction and bubble size decreased obviously when transversal or especially longitudinal slots were used. Moreover, the greater the number of both kinds of slot, the lower the mentioned parameters. With increasing anode slot width, the gas volume fraction decreased slightly while there was almost no effect on the bubble size distribution (BSD). Reducing the anode slot height caused a higher gas volume fraction. Both the overall gas volume fraction and BSD in industrial-scale cells are apparently influenced by two large circulation vortices caused by electromagnetic forces (EMFs).

Improved CFD modeling of full dissolution of alumina particles in aluminum electrolysis cells considering agglomerate formation

At present, the aluminium electrolytic cell is developing in the direction of large-scale, but it still adopts the past ‘overunder’ discharging method for the control of alumina concentration, which has low current efficiency and frequent occurrence of anode effect. The development of fiber optic measurement technology provides new possibilities for the measurement of anode current in electrolytic tanks. Based on fiber optic current measurement technology and modern detection technology, this paper develops a real-time monitoring system for aluminium electrolytic tanks based on regional current, realizes accurate online measurement of regional current and tank voltage in aluminium electrolytic tanks and uploads data to the cloud, realizes fine feedback of alumina concentration, and provides data support and technical guarantee for the independent discharging of aluminium electrolytic tanks in different regions.

Sodium has a big affect on the air and CO2 reactivity of the carbon anodes. A high content of sodium in the anode will cause serious oxidation and falling dusts of the anode and affect the working conditions of the aluminium electrolysis cell. Sodium in the anode mainly comes from the petroleum coke and the recycled butts. The present butt cleaning technology cannot completely clean the attached bath on the butts, and the present equipment occupies too much space. We developed a special impact butt-cleaning system which has advantages of small volume, high capacity, low failure rate, low operation cost and can be used in the existing anode production line. The industrial application has proven that the new equipment can completely clean the attached bath and soft butts, improve the reactivity and avoid the falling dusts.

Carbon anode blocks, used in aluminum electrolysis cells, are usually baked in furnaces for several days, during which they release volatiles due to pitch pyrolysis. Therefore, numerical modeling of anode baking furnaces has to include some representation of pitch pyrolysis via a set of kinetic parameters. These kinetic parameters can vary with raw materials and baking parameters and are tedious to determine experimentally. In this work, we studied how the main outputs of an anode baking model are affected by the variance of the kinetic parameters. Results show that certain model outputs are not considerably influenced by changes in the kinetic parameters (e.g., spatial variation of anode porosity, maximum heating value from volatiles), while others are significantly affected (e.g., time evolution of anode porosity, time of maximum heating value of volatiles, internal pressure of anode), in particular by activation energy variability.

Proof of concept to recover thermal wastes from aluminum electrolysis cells using Stirling engines

This article describes the importance of temperature monitoring in aluminum electrolysis cells. The challenge of getting long-term performance from thermocouples is discussed, and appropriate examples are given of information of value to cell operations derived from bath temperature measurements.