AI has been introduced into the work of EuroChem-VolgaKaliy

2024: Introduction of the recommendation system

EuroChem has introduced a recommendation system in the production of potassium chloride that helps maintain the quality of the finished product. The solution using artificial intelligence is deployed at EuroChem-VolgaKaliy and provides an increase in the KCl recovery factor by 0.2% or 3.1 thousand tons in terms of the finished product. The expected economic effect of more than 90 million rubles per year will be achieved through a more accurate leaching process at a sylvinite enrichment plant.

At EuroChem-VolgaKaliy, when water is supplied to the leach unit, a situation periodically arises when the finished product is excessive or, conversely, insufficient potassium chloride content, and this is the target product of the enterprise. Such cases are the result of suboptimal leaching - dissolution of sodium and potassium salts in water, which affects the final quality of fertilizer. Correction of the water flow rate is not always carried out in a timely manner and in the required amount, and as a rule, is determined by experience.

In order to increase the production of the finished product of potassium chloride (KCl), the production digitalization team together with the enterprise team automated the process of adjusting the supply of water for leaching. To implement the idea, a large array of data was analyzed, including data from the MAYA in-line calimeter calibrated in-house. A recommendation system based on artificial intelligence helps to maintain the quality of the product at the proper level by timely regulating the supply of leaching water in the right ratios, as well as reducing the flow rate of water on the production line.

𝓲

EuroChem

Virtual quality analyzers, which were implemented as part of the recommendation system, were developed at the inlet and outlet of the leach apparatus. The data from the sensors are supplied to the model, which calculates the optimal required water consumption for maximum dissolution of sodium chloride and preservation in the potassium chloride system. Then the results are displayed on the operator's monitor, he can only indicate the required water flow rate in the input window, according to the recommendations received.

The project is at the MVP stage (minimum viable product), pilot tests and pilot operation have already been carried out, the project is being prepared for transfer to commercial operation by factory employees. The project team is improving the system, they are working to improve the accuracy of KCl measurements. The result of the analysis is highly dependent on the water content of the sample. Digitalizers are finalizing the current solution model in order to automatically calibrate the calimeter and determine the remaining elements in the ore, namely calcium and magnesium, "said Ivan Biro, head of digitalization at EuroChem mining enterprises.

The introduction of a recommendation system based on artificial intelligence was a significant step to optimize the technological process of enrichment. It greatly simplifies the work of operators and technologists. All data is supplied to the monitor in a convenient form, and the operator can only make the necessary water flow in accordance with the recommendations. This not only speeds up the process, but also allows employees to focus on quality control rather than manual calculations. The technology helps minimize routine operations and increases the efficiency of the entire team, "said Sergey Wolf, director of digitalization and automation at EuroChem-VolgaKaliy.