Steel manufacturing is one of the key industries in the development of the modern economy. For many years, steel production has been manual, labor-intensive industry. The emergence of artificial intelligence (AI) has brought about significant changes, offering effective cost control and efficiency gains.
We will explore how AI is being used in steel manufacturing and how it can help improve efficiency and accuracy.
In the past, automated optical inspection (AOI) was introduced to improve the leakage rate due to the fatigue of manual visual inspection.
However, there are many types of steel roll surface defects. It is difficult for rules-based AOI machine to find out the problem comprehensively. In addition, it often struggle with the problem of overkill/underkill.
By applying the image recognition technology of deep learning, it enables AOI to detect surface defects of steel coils more accurately.
Solution： Nilvana Vision Studio + Nilvana Vision Inference Runtime + AIoT Edge
On site, each bundle of rebar is counted manually to ensure the correct quantity. However, this process is tedious and labor-intensive, requiring multiple checks to ensure consistent counts.
Deep learning in the field of image analysis can automate rebar counting and reduce the possibility of errors and increase the speed of inventory in the field.
Solution： Nilvana Vision Studio + Nilvana Vision Inference Runtime + AIoT Edge
In the past, someone had to be on site to make adjustments based on data and experience during the production process. As the number of skilled blast furnace operators decreases, the skills are gradually lost.
By using machine learning to automatically optimize the operating conditions for the amount of iron ore and other raw materials and the amount of hot air in the furnace, the blast furnace operation is stabilized and the pressure on operators is reduced, resulting nonstop production.
If you are troubled with the shortage of blast furnace operators, you can learn to simulate process results and adjust parameters in real time with machine learning.
Solution：AI Starter Kit + AIoT Edge
AI Model Development Tool. Powerful features such as machine annotation and data augmentation are included to save 80% of your annotation time.
The Best Helper for Model Optimization, Application Deployment.
No need to write and modify model inference code, simply activate model API endpoints through GUI settings.
