SFU scientists automated electrolyte composition analysis for aluminium production

December 12, 2018

A team from Siberian Federal University (SFU) suggested a new method for automatic composition analysis of electrolyte samples from electrolysis baths. It will provide for more accurate technological control and increase the efficiency of aluminium production. The article of the scientists was published in the Crystals journal.

Metal aluminium is obtained in the course of electrolysis - passing of current through an alumina-cryolite melt at around 950°?. The main component of a melt is cryolite (a salt containing sodium, aluminium, and fluorine atoms) to which aluminium oxide (alumina) is added. To improve the technological properties of the electrolyte, aluminium fluoride, calcium fluoride, and sometimes magnesium and potassium fluoride can be added to it. In the course of electrolysis the composition of the substance in the baths constantly changes, and the ratio of components shifts as well. Preservation of optimal bath composition is a key element of electrolysis technology.

To maintain the optimal composition, one has to constantly take and analyze electrolyte samples. Such analysis should be as accurate as possible and take little time. An express control method used in the industry is called X-ray diffraction quantitative phase analysis. It is based on studying X-ray images formed by the X-rays reflected from samples. In its traditional variant, such analysis has one considerable disadvantage: it requires regular calibration using control materials with accurately determined phase compositions, and it also fails to take the actual crystal structure of the phases into account. An alternative option is called the Rietveld method. It provides for quantitative analysis on the basis of specifying the atomic and crystal structure of component phases without using control samples. However, this method is interactive and difficult to automate, as one has to manually set up to one hundred of initial system parameters and to manage the order of their programmed adjustment.

A team from SFU modernized the Rietveld method to make it applicable for automated analysis. To do so, they developed a self-configuring evolutionary genetic algorithm - a program that uses the principle of biological natural selection to find optimal parameter values when modelling an X-ray image. First, a genetic algorithm uses random values, but then it optimizes the vast range of X-ray image and phase crystal structure parameters and manages the adjustment of only the best of them using the Rietveld method. Therefore, the algorithm is able to work without human involvement and self-learn.

"Generally, our results meet the technological criteria for the accuracy of electrolyte analysis that are used at aluminium production facilities. We can recommend our genetic algorithm to express control of electrolyte composition. The analysis shows a minor system error caused by over-estimation of cryolite concentration due to its uneven crystallization in the course of sampling. Before this method is implemented by the industry, we have to eliminate this error and also to improve the efficiency of the method," says Igor Yakimov, the head of the project, PhD in physics and mathematics, and professor of the Institute of Non-Ferrous Metals and Material Studies of SFU.
The study was carried out together with scientists from Siberian State Aerospace University.

Siberian Federal University

Related Algorithm Articles from Brightsurf:

CCNY & partners in quantum algorithm breakthrough
Researchers led by City College of New York physicist Pouyan Ghaemi report the development of a quantum algorithm with the potential to study a class of many-electron quantums system using quantum computers.

Machine learning algorithm could provide Soldiers feedback
A new machine learning algorithm, developed with Army funding, can isolate patterns in brain signals that relate to a specific behavior and then decode it, potentially providing Soldiers with behavioral-based feedback.

New algorithm predicts likelihood of acute kidney injury
In a recent study, a new algorithm outperformed the standard method for predicting which hospitalized patients will develop acute kidney injury.

New algorithm could unleash the power of quantum computers
A new algorithm that fast forwards simulations could bring greater use ability to current and near-term quantum computers, opening the way for applications to run past strict time limits that hamper many quantum calculations.

QUT algorithm could quash Twitter abuse of women
Online abuse targeting women, including threats of harm or sexual violence, has proliferated across all social media platforms but QUT researchers have developed a sophisticated statistical model to identify misogynistic content and help drum it out of the Twittersphere.

New learning algorithm should significantly expand the possible applications of AI
The e-prop learning method developed at Graz University of Technology forms the basis for drastically more energy-efficient hardware implementations of Artificial Intelligence.

Algorithm predicts risk for PTSD after traumatic injury
With high precision, a new algorithm predicts which patients treated for traumatic injuries in the emergency department will later develop posttraumatic stress disorder.

New algorithm uses artificial intelligence to help manage type 1 diabetes
Researchers and physicians at Oregon Health & Science University have designed a method to help people with type 1 diabetes better manage their glucose levels.

A new algorithm predicts the difficulty in fighting fire
The tool completes previous studies with new variables and could improve the ability to respond to forest fires.

New algorithm predicts optimal materials among all possible compounds
Skoltech researchers have offered a solution to the problem of searching for materials with required properties among all possible combinations of chemical elements.

Read More: Algorithm News and Algorithm Current Events
Brightsurf.com is a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com.