Russian Federation aluminium titan UC Rusal announced this week the implementation of a new neural network at Irkutsk Aluminium Smelter that will monitor the quality of the plant’s aluminium ingot.
Rusal’s new neural network utilizes a video monitoring system to check the surface quality of small aluminium ingots during production, checking for defects as the ingots pass by on the casting conveyor. Among the defects the system checks for are cracks, bumps, and foreign inclusions.
The neural network saves the information gathered and displays it as an analytical diagram showing the number of defects detected per day and the number of sub-standard aluminium ingots found in each shift. The neural network can be adjusted by plant personnel based upon the product then in production.
Rusal’s Chief Technical Officer Victor Mann said in a press release that the neural network would boost the quality of the product shipped to its clients.
“This new solution enables production line personnel to track the quality of aluminium ingots in real time, while they’re being cast. It will guarantee that the products which get shipped to our customers meet all the applicable requirements.”
Rusal says it plans to implement similar neural networks at all of its aluminium smelters in the future. Ultimately such systems will make the aluminium smelting process a largely automated process. In addition, several of Rusal’s sites are also incorporating automated video monitoring systems to detect losses in hermetic seals in the potlines’ reduction cells.