Key Takeaways & Executive Findings
- •• A novel CRNN model integrating ResNet18 with CBAM and Bayesian optimization achieves 97.01% accuracy and 99.85% AUC for smelting stage recognition from furnace mouth flame videos. • The model outperforms traditional static image methods by effectively extracting spatiotemporal features, addressing dynamic changes in the smelting process. • Real-time performance is demonstrated with an average recognition time of 5.49 ms, making it suitable for industrial online monitoring. • The approach offers a non-contact, cost-effective alternative to sub-lance and mass spectrometry methods for converter endpoint control.
Abstract
The converter steelmaking process represents a pivotal aspect of steel metallurgical production, with the characteristics of the flame at the furnace mouth serving as an indirect indicator of the internal smelting stage. Effectively identifying and predicting the smelting stage poses a significant challenge within industrial production. Traditional image-based methodologies, which rely on a single static flame image as input, demonstrate low recognition accuracy and inadequately extract the dynamic changes in smelting stage. To address this issue, the present study introduces an innovative recognition model that preprocesses flame video sequences from the furnace mouth and then employs a convolutional recurrent neural network (CRNN) to extract spatiotemporal features and derive recognition outputs. Additionally, we adopt feature layer visualization techniques to verify the model’s effectiveness and further enhance model performance by integrating the Bayesian optimization algorithm. The results indicate that the ResNet18 with convolutional block attention module (CBAM) in the convolutional layer demonstrates superior image feature extraction capabilities, achieving an accuracy of 90.70% and an area under the curve of 98.05%. The constructed Bayesian optimization-CRNN (BO-CRNN) model exhibits a significant improvement in comprehensive performance, with an accuracy of 97.01% and an area under the curve of 99.85%. Furthermore, statistics on the model’s average recognition time, computational complexity, and parameter quantity (Average recognition time: 5.49 ms, floating-point operations per second: 18260.21 M (1 M = 1 × 106), parameters: 11.58 M) demonstrate superior performance. Through extensive repeated experiments on real-world datasets, the proposed CRNN model is capable of rapidly and accurately identifying smelting stages, offering a novel approach for converter smelting endpoint control.
1. Introduction
The converter steelmaking process occupies a pivotal position in the metallurgical industry, with its intelligent manufacturing at the forefront of the sector and serving as a demonstration for other [1–4]. Globally, advanced large- and medium-sized steelmaking enterprises have extensively researched automatic control technologies for converters [5]. Endpoint control primarily focuses on the temperature and carbon content of molten steel, with its accuracy directly affecting product quality. Currently, traditional control models are categorized into static and dynamic types, with the former primarily based on energy and material balance calculations during the process to determine the addition of auxiliary materials and blowing conditions [6]. Nevertheless, owing to the inherent complexity and variability of the smelting process, these models exhibit significant deviations in effectively tracking and adjusting the process for optimization.
Dynamic control models address the shortcomings of static models and are crucial for improving the accuracy of endpoint carbon and temperature control, as well as enhancing the quality of the steel [7–9]. The advancement of dynamic model control technology has largely been driven by the implementation of sub-lance detection and exhaust gas mass spectrometry, which have attained a relatively mature level of application [10–11]. However, challenges related to detection accuracy, equipment maintenance, and associated costs highlight the necessity for the development of a novel endpoint prediction model.
The optical information present in the furnace mouth flame provides a direct and real-time representation of the decarburization reaction progress occurring within the furnace [12]. In light of the rapid advancement of industrial big data, it is critically important to develop non-contact intelligent prediction models for smelting that leverage the characteristics of the furnace mouth flame and associated optical information. Current research mainly focuses on two directions: radiation spectral information methods [9,13–14] and flame image information methods [15–16]. Zhang et al. [13] collected spectral information of the converter flame using a USB2000 and a spectrometer, and simultaneously obtained continuous carbon content changes during the later stages of smelting using a flue gas analysis mass spectrometer, constructing a large sample dataset and establishing a prediction model.
Loading authentic research manuscript (Pages 1–5)...
Zhangjie Dai, Ye Sun, Wei Liu, Shufeng Yang, Jingshe Li (2025). Smelting stage recognition for converter steelmaking based on the convolutional recurrent neural network. Int. Journal of Minerals, Metallurgy and Materials (矿物冶金与材料学报). https://doi.org/10.1007/s12613-024-3086-2
Research & Educational Purpose Only:The translations, structured abstracts, analytical annotations, and data reports provided by SinoTechIntel are intended exclusively for academic research, internal corporate R&D, and educational benchmarking. They do not constitute formal engineering, chemical safety, legal, or professional advice.
Copyright & Intellectual Property Notice: Original copyright of the underlying source articles and experimental data remains with the respective authors, institutions, and original publishing journals. SinoTechIntel claims intellectual property only over its proprietary translations, analytical syntheses, and AEO structured enhancements in accordance with international fair use and academic citation principles.
Frequently Asked Questions
What is the main contribution of this paper?
The paper proposes a novel CRNN-based model that integrates ResNet18 with CBAM and Bayesian optimization to recognize smelting stages from furnace mouth flame videos, achieving high accuracy and real-time performance.
How does the proposed model differ from traditional image-based methods?
Traditional methods use single static flame images, which fail to capture dynamic changes. The proposed model processes flame video sequences to extract spatiotemporal features, improving recognition accuracy.
What are the key performance metrics of the BO-CRNN model?
The BO-CRNN model achieves an accuracy of 97.01% and an AUC of 99.85%, with an average recognition time of 5.49 ms, demonstrating superior performance.
What is the significance of this research for the steel industry?
It provides a non-contact, cost-effective method for real-time smelting stage recognition, which can enhance endpoint control and support intelligent manufacturing in converter steelmaking.
What datasets were used to validate the model?
The model was validated through extensive repeated experiments on real-world datasets of furnace mouth flame videos from converter steelmaking processes.
Related Technical Papers & Translations
Direct Repair of the Crystal Structure and Coating Surface of Spent LiFePO4 Materials Enables Superfast Li-Ion Migration
The rapid accumulation of spent LiFePO4 (LFP) cathodes from retired lithium-ion batteries necessitates the development of effective and environmental-friendly recycling strategies. In this context, direct regeneration has emerged as a promising approach for reclaiming LFP cathode materials, offering a streamlined pathway to restore their electrochemical functionality. We report an integrated regeneration protocol that simultaneously repairs the degraded crystal structure and reconstructs the damaged carbon coating in spent LFP. The regenerated cathode material had superfast lithium-ion diffusion kinetics and a stable cathode–electrolyte interface, giving a remarkable rate capability with specific capacities of 122 mAh g−1 at 5C and 106 mAh g−1 at 10C (1C = 170 mA g−1). It also maintained capacities of 110.7 mAh g−1 (5C) and 84.1 mAh g−1 (10C) after 400 cycles. It could be used in harsh environments and could be stably cycled at subzero temperatures (−10 and −20 °C) and in solid-state electrolyte batteries. Life cycle assessment combined with economic evaluation using the EverBatt model reveals that this direct regeneration approach has high economic and environmental benefits.
Oxide Semiconductor for Advanced Memory Architectures: Atomic Layer Deposition, Key Requirement and Challenges
Oxide semiconductors (OSs), introduced by the Hosono group in the early 2000s, have evolved from display backplane materials to promising candidates for advanced memory and logic devices. The exceptionally low leakage current of OSs and compatibility with three-dimensional (3D) architectures have recently sparked renewed interest in their use in semiconductor applications. This review begins by exploring the unique material properties of OSs, which fundamentally originate from their distinct electronic band structure. Subsequently, we focus on atomic layer deposition (ALD), a core technique for growing excellent OS films, covering both basic and advanced processes compatible with 3D scaling. The basic surface reaction mechanisms—adsorption and reaction—and their roles in film growth are introduced. Furthermore, material design strategies, such as cation selection, crystallinity control, anion doping, and heterostructure engineering, are discussed. We also highlight challenges in memory applications, including contact resistance, hydrogen instability, and lack of p-type materials, and discuss the feasibility of ALD-grown OSs as potential solutions. Lastly, we provide an outlook on the role of ALD-grown OSs in memory technologies. This review bridges material fundamentals and device-level requirements, offering a comprehensive perspective on the potential of ALD-driven OSs for next-generation semiconductor memory devices.
Laser powder bed fusion of biodegradable Zn-4Cu alloy: Processing, microstructure and properties
Zn's natural degradability and biocompatibility make it a promising candidate for implants, however, its mechanical properties remain insufficient for bone applications. In this study, the performance of Zn was enhanced by developing Zn-Cu alloys via laser powder bed fusion (LPBF). Optimal LPBF parameters for forming stable tracks were achieved by adjusting laser power and scanning speed. Under optimized conditions of 100 W and 100 mm/s, high-density (99.58%) Zn-Cu alloys with improved hardness (68.2HV) and yield strength (160 MPa) were achieved. These improvements are attributed to solid solution strengthening, segregation strengthening, and grain refinement. The Zn-Cu alloys also demonstrated favorable degradation behavior, with a rate of 0.16 mm/year. This degradation is primarily driven by micro-galvanic corrosion between the CuZn5 phase and Zn matrix, along with refined grains and increased grain boundary density. This work demonstrates a viable strategy for fabricating Zn-based implants with enhanced structural integrity and mechanical performance via LPBF.