Automatic Feature Extraction for Mold Flux Phase Changes: Analyzing Sequence Images for Optimized Steel Production

The phase changes of mold fluxes are difficult to directly observe due to the high temperature, transient fluid flow, complex phase transitions, chemical reactions, and opacity of the mold wall. To overcome this challenge, the SHTT II tester for melting and crystallization temperature is widely used to observe the crystallization behaviors of mold fluxes. However, after the experiment is completed, the process of manually analyzing and recording the images wastes valuable manpower and hinders the development of experimental process information. Therefore, there is an urgent need to develop automatic feature extraction and mathematical modeling technology for sequence images.

Attachment 1 contains 562 sequence images depicting the melting and crystallization of mold fluxes. These images were collected from the 110th to the 671st seconds after the start of the experiment. The file serial numbers follow the chronological order of collection, with images being captured every 1 second. The information is presented in digital image format in Attachment 1, with the corresponding time of each image and the temperature values of thermocouples No. 1 and No. 2 marked in the upper left corner (refer to Figure 1).

Automatic feature extraction and mathematical modeling technology of sequence images is crucial to effectively analyze and understand the phase distribution and behaviors of mold fluxes during melting and crystallization. By leveraging advanced image processing and machine learning algorithms, it is possible to automatically extract relevant features from the sequence images, such as temperature gradients, phase boundaries, and crystalline structures. These features can then be used to develop mathematical models that accurately describe the melting and crystallization processes of mold fluxes.

The development of such technology will not only save manpower and time in the analysis of sequence images but also provide valuable insights into the solidification requirements of different steel grades. By studying the phase distribution and behaviors of mold fluxes, researchers can optimize the design of mold fluxes to enhance their metallurgical functions and improve the efficiency and quality of the continuous casting process.

In conclusion, the application of automatic feature extraction and mathematical modeling technology for sequence images of mold fluxes' melting and crystallization is crucial for a comprehensive understanding of the phase transformations and behaviors of mold fluxes. This technology will enable more efficient analysis of sequence images and contribute to the optimization of mold flux design in the continuous casting process, ultimately enhancing the overall efficiency and quality of steel production.