Aluminium electrolytic cells: A computer simulator for training and supervision

Tikasz, L.; Bui, R. T.; Potocnik, Vinko

doi:10.1007/bf01206536

Cited by 12 publications

(8 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[8]), and increasing them gradually for lower concentrations. The last part of Table II Similarly to the preceding cases, Figure 7 shows that, for lower concentrations, the sensitivity of the cell resistance increases, and, as a consequence, the quality of cell control improves.…”

Section: Control Of Cell In Statementioning

confidence: 93%

“…Note that the bands identified as lvq (3a), lvq (3b), and lvq (3c) of Figure 7 are generated by the criteria in Eqs. [8], [9], and [10], respectively. In this case, due to the parabolic form of the characteristic curve ( Figure 2) four operation zones can be clearly recognized in Figure 7: standard, lvq (3a), lvq (3b), and lvq (3c).…”

Section: Control Of Cell In Statementioning

confidence: 99%

“…The simulator, built recently by Tikasz et al, [7,8] is based on a mathematical model made of 18 differential equations and several algebraic relations describing the time-dependent heat and mass balances inside the cell. For the purpose of establishing the heat and mass balances, the cell is divided into 11 components.…”

Section: Neural Control Of Concentrationmentioning

confidence: 99%

See 2 more Smart Citations

Intelligent control of the feeding of aluminum electrolytic cells using neural networks

Meghlaoui

Bui

Tikasz

et al. 1997

Metall Mater Trans B

Self Cite

View full text Add to dashboard Cite

To be efficient, the control of alumina feeding of the electrolytic cell must be based on cell resistance, alumina concentration, and cell state. Most control schemes now in use are based on cell resistance only, and, thus, constitute an open-loop control that lacks robustness because their decision criteria are not explicitly tied to concentration nor to cell state. This results in the cell operating at nonoptimal concentrations, and cell efficiency is diminished. An optimal operation requires a knowledge of concentration and an adjustment of the decision criteria as a function of concentration. A learning vector quantization (LVQ) type of neural network was built and trained to recognize the cell state. Knowing the state of the cell and its resistance, concentration can be estimated using predetermined regression functions. The decision criteria for the control logic are then consequently adapted. A closed-loop control scheme is thus obtained. Results show that, with its control so structured, the cell can operate at or near optimal concentrations independently of its state. This flexible and intelligent character of the neural control can provide a considerable advantage as compared to the standard control.

show abstract

Section: Control Of Cell In Statementioning

confidence: 93%

Section: Control Of Cell In Statementioning

confidence: 99%

See 1 more Smart Citation

Intelligent control of the feeding of aluminum electrolytic cells using neural networks

Meghlaoui

Bui

Tikasz

et al. 1997

Metall Mater Trans B

Self Cite

View full text Add to dashboard Cite

show abstract

“…Hashimoto and Ikeuchi [5] split the cell into eight components and developed a nonlinear model that included ledge formation. More recently, Tikasz et al [6,7] built a model based on an 11-component cell and made of 18 differential equations and several algebraic relations. Conceived as a user-interactive computer simulator of the cell, it is currently used as a tool for research, process analysis, process supervision, and operator training.…”

Section: Previous Workmentioning

confidence: 99%

Predictive control of aluminum electrolytic cells using neural networks

Meghlaoui¹,

Bui

Tikasz

et al. 1998

Metall Mater Trans B

Self Cite

View full text Add to dashboard Cite

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.

show abstract

“…The flow pattern and velocity of the electrolyte are induced by the combined effects of the gas release pattern, the drag force from the metal pad, the thermal convection, and the magnetohydrodynamic force arising from the interaction between the magnetic fields and the electric current (Feng et al, 2010;Grjotheim and Welch, 1980;Doheim et al, 2008). Tikasz et al (1994) Incorporate 3 types of alumina for alumina balance calculation Biedler (2003) and Jessen (2008) Include heat loss from anode rods, deck plate into computation Apply first order dissolution rate equation Yurkov and Mann (2005) Incorporate 3 transformation phases of alumina for heat balance computation Kolås and Støre (2009) Consider emission and neutralization for the computation of AlF 3 mass balance Gusberti et al (2011) Apply extended cell control volume for heat balance computation different additional process features into consideration, most of them only compute the global temperature of important variables and overall ledge thickness. (Significant features of these models are listed in Table 1.)…”

Section: Introductionmentioning

confidence: 97%