Aluminum Reduction Technology: Cell Voltage and Pot Control
Sponsored by: TMS Light Metals Division, TMS: Aluminum Committee
Program Organizers: Mark Dorreen, Light Metals Research Centre, The University of Auckland

Tuesday 2:00 PM
February 28, 2017
Room: 2
Location: San Diego Convention Ctr

Session Chair: Daniel Whitfield, Emirates Global Aluminium

2:00 PM Introductory Comments

2:05 PM  
Application of Multivariate Statistical Process Control with STARprobeTM Measurements in Aluminium Electrolysis Cells: Jean-Pierre Gagné1; Pascal Lavoie2; Albert Mulder2; Rémi St-Pierre1; Pascal Côté1; 1STAS; 2Consultant
    Due to the multivariate nature of the aluminium electrolysis process, the usual univariate control algorithms coupled with the low visibility of the process used in the aluminium industry intrinsically causes control errors resulting in sub optimal process control. A trend in the industry consists in applying multivariate statistical process control with specific responses to the cause of variations. Such a system in the form of a Bath Temperature and Chemistry Control Module was developed to be used in conjunction with state of the-art STARprobeTM measurements of electrolyte properties, providing the ability to respond quickly to causes of abnormalities detected from the immediate synchronous bath measurements. This paper presents the system and results from a smelter’s pilot section.

2:30 PM  
Predicting Instability and Current Efficiency of Industrial Cells: Patrice Côté1; Olivier Martin1; Bertrand Allano1; Véronique Dassylva-Raymond2; 1Rio Tinto Alcan; 2Consultant, Reso-Lean Conseil
    New statistical models were developed to predict the behavior of industrial potlines. A first model, based on a multiple linear regression analysis, predicts the instability (cell noise) of P155 potlines and is used as a tool to investigate drifts. An example of the usefulness of such a model is presented. A second model that predicts the current efficiency was built using a similar approach. It is based on 10 years of data from 30 potlines. The non-linearity of the relationship between current efficiency (CE) and anode-cathode-distance (ACD) was investigated, and confirmed, during the course of this study. It has been taken into account in the CE model. These current efficiency and instability models provide new ways to anticipate potline performance and to conduct gap analysis.

2:55 PM  
Detecting, Identifying and Managing Systematic Potline Issues with Generation 3 Process Control: Nursiani Tjahyono1; Yashuang Gao1; David Wong1; Ron Etzion1; Albert Mulder2; 1University of Auckland, Light Metals Research Centre; 2IT Consultant
    Systematic issues in aluminium smelting process are those that occur on multiple pots simultaneously and if not identified early can result in prolonged issues across a potline. However identifying these issues early can be a difficult task for smelter engineers and management, particularly when information sources are not available in a centralised location. In addition, they are often not displayed in a way that is convenient for tracking and identifying systematic problems. For example, an increasing trend in cryolite ratio on all pots can be hard to spot if graphical trends are only displayed for individual pots, or worse, only in tabular format. Furthermore, the cause of these systematic issues are often unavoidable, such as a change in the supplier and properties of raw materials. This paper presents a number of cases of identifying and managing systematic issues in smelting process using Generation 3 process control system.

3:20 PM  
Integrating a New Smelter Supervision HMI in Existing Control Systems at ALBRAS: Vanderlei Fernandes1; Geir Sandnes2; Leonel Mota Ivo3; Rogério Labanca3; 1ALBRAS Aluminio Brasileiro S.A.; 2Norsk Hydro ASA; 3Accenture
    ALBRAS needed to align its operation with the follow up methodology used by Hydro Aluminium, its majority owner. To achieve this goal, ALBRAS decided to standardize its potline production information by installing Hydro’s APICS system. A tight integration with the underlying pot control systems was needed to replace their supervision HMI by APICS’s one. To design an interface to integrate both systems and install it in the running control systems without disturbing the smelters’ operations was a challenge. In this paper the interface solution provided to integrate APICS with two different versions of the SCORE supervision and control system, installed in all four smelters at ALBRAS, will be presented. Challenges will be discussed, the defined interface solution will be detailed and the results along with some lessons learnt will be presented.

3:45 PM Break

4:00 PM  
Clustering Aluminium Reduction Cells: Flavia Lima1; Alan Souza1; Fabio Soares1; Diego Lisboa1; Roberto Oliveira1; 1UFPA
    Aluminium Smelting Potlines usually have big number of cells, producing aluminium in a continuous and complex process. An analytical monitoring is essential to increase the industries' competitive advantage, however, during their operation, some cells share similar behaviours, therefore forming clusters of cells. These clusters rely on data patterns that are usually implicit or invisible to operation, but can be found by means data analysis. In this work we present two clustering techniques (Fuzzy C-Means and K-Means) to find and cluster the cells that present similar behaviours. The benefits of clustering are mainly in the simplification of potline analysis, since a large number of cells can be summarized in one single cluster, which can provide richer but compacted information for control and modelling.

4:25 PM  
Study of Impact of the Anode Slots on the Voltage Fluctuations of Aluminium Electrolysis Cell Using Bubble Layer Simulator: Sandor Poncsak1; László Kiss1; Sébastien Guérard2; Jean François Bilodeau2; 1Univeristy of Quebec at Chicoutimi; 2Rio Tinto Aluminium
    There is a constant effort from the aluminium producers to reduce energy consumption of the Hall-Héroult cells in order to decrease cost and environmental fingerprint. Among others, slotted anodes were introduced in order to promote faster evacuation of the electrically isolating anode gas bubbles and thus diminish their contribution to the total cell voltage. A bubble layer simulator was developed to reproduce cell voltage fluctuations, caused by the dynamics of anode bubbles. Results show that the slots cut in the right position and direction can reduce both the amplitude of fluctuations and the average cell voltage. This impact is even higher for new, almost horizontal flat anode bottom. It is also revealed that the slots are not only acting as bubble sink, but they contribute to the acceleration of the bubble layer and thus their role in the momentum exchange between liquid and gas must be taken into account.

4:50 PM  
Minimizing Cathode Voltage Drop by Optimizing Cathode Slot Design: Ralph Friedrich1; Frank Hiltmann1; Andreas Lützerath2; Richard Meier2; Markus Pfeffer1; Till Reek2; Oscar Vera Garcia1; 1SGL CFL CE GmbH; 2TRIMET Aluminium SE
    Contact resistance between cathode block and cast iron plays an important role for the ohmic resistance of the cathode system and hence for the cathode voltage drop (CVD). Investigations of the standard dove-tail cathode slot design showed poor electrical contact in the bottom of the slot and small contact area on the cathode wings. Different thermomechanical properties of carbon cathode block, cast iron, and steel collector bar resulted in displacements relative to each other after rodding, during cooling and handling. To mitigate this effect without changing materials or procedures, different slot geometries were modeled with respect to preheating and casting. By introducing anchoring grooves for the cast iron at the bottom corners of the slot, bottom contact was achieved: Lab tests proved that voltage drop between cathode slot bottom and cast iron decreased significantly. Subsequent field trials showed a CVD reduction of 25 mV.