A worker is monitoring the ‘AI heat treatment predictive control system’ in the continuous hot dip galvanizing (CGL) operation room at KG Steel’s Dangjin plant. (Photo = KG Steel)

KG Steel will begin operating its process temperature artificial intelligence (AI) predictive control system in earnest from the beginning of this month. Through this, it is expected to secure a high level of product material prediction accuracy and speed up smart factory construction, while also securing quality competitiveness and reducing costs.

KG Steel (016380) announced on the 7th that it will operate the 'AI heat treatment predictive control system' in the continuous hot dip galvanizing (CGL) process furnace section of the Dangjin plant from the beginning of this month. Previously, KG Steel developed CGL heating furnace AI optimal control technology last year and recently completed pilot operation and performance enhancement work for about half a year.

In the system, AI is designed to implement better patterns by learning the behavior of skilled workers. AI receives new work instructions and process status information every 15 seconds and predicts product quality up to 2 hours from the current point to maintain optimal operating conditions.

This is possible by AI controlling the internal temperature of the heating furnace. Once the heat treatment predictive control system is in operation, it becomes possible to predict the physical properties such as △yield strength △tensile strength △elongation △hardness of galvanized steel sheets before the product is put into the heating furnace. This is the first attempt in the surface-treated steel sheet industry.

Accuracy is also excellent. The accuracy of predicting physical properties and set temperature is expected to reach 98%, ensuring quality competitiveness for manufactured products. Before introducing the AI ​​system, the final stage of the CGL process, the product at the exit, was sampled and the material was analyzed, which took about 2 hours.

A KG Steel official said, “If workers control the equipment, it is difficult to check product properties in real time, so we had no choice but to set process conditions conservatively. After AI learns from past work on its own, it presents an optimal pattern with an accuracy approaching 100%.” “Because of this, more efficient operations will be possible,” he explained.

In addition, the introduction of AI technology is expected to predict and control product properties, improving the production quality of KG Steel products and reducing the use of liquefied natural gas (LNG), resulting in significant cost savings.

Meanwhile, on the 31st of last month, KG Steel held a '3CGL heating furnace predictive control system construction presentation' for executives and employees at the Dangjin Plant Technology Research Center and explained the construction progress and expected effects. In the future, KG Steel plans to gradually expand and apply AI technology to △real-time plating adhesion prediction model △construction of air knife automatic control system.