Checking the coupling between overhead crane hook and steel ladle trunnion in Steelmaking Plant using convolutional neural networks

Resumo

Overhead cranes are equipment designed for the efficient and safe movement of large loads and are widely used in the industry. They play a very important role in logistics and production by allowing the flow of materials, machinery, and products from one location to another, whether within a factory, warehouse, shipyard, or any other industrial environment. These equipment have a simple mechanical structure, consisting of a main beam equipped with wheels that allow movement along tracks. A hoist is suspended under the main beam, which can also move along it, and is used to lift and lower loads. The element that couples the load to the lifting system varies according to the type of load.
In this article, we explore the coupling system commonly used in the steel industry for handling steel ladles through overhead cranes, called hooks and bails. This study proposes the use of artificial intelligence (AI) and machine learning techniques, through computer vision based on convolutional neural networks to detect the coupling, segmenting both the hook and the bail and thereby determining the precise alignment of the assembly before lifting. This contributes to validating the operator's visual information, reducing possible human errors, and promoting greater safety for both the process and the people involved.