Multi-Channel Graph Convolutional Network based End-Point Element Composition Prediction of Converter Steelmaking

Chang, Shuchao; Zhao, Chunhui; Li, Yuanlong; Zhou, Min; Fu, Chuan; Qiao, Huiting

doi:10.1016/j.ifacol.2021.08.234

Cited by 9 publications

(6 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Multitask learning can effectively leverage these correlations, improving prediction accuracy and efficiency. Notably, graph neural networks provide a suitable model architecture for this approach [52] …”

Section: Conclusion and Future Research Directionsmentioning

confidence: 99%

“…Recently, there has been an increasing interest in incorporating novel deep learning (DL) methodologies such as transfer learning, [51] graph neural networks (GNNs), [52] CNNs, [46] auto‐encoder Bayesian network, [24] and reinforcement learning [36] . The application of recurrent neural networks (RNNs), specifically long short‐term memory (LSTM), has been explored for handling time‐series data [53] .…”

Section: Data‐driven Modeling Workflow For Bof Processmentioning

confidence: 99%

“…integrated CBR and LSTM models to improve endpoint predictions, showing higher accuracy than traditional SVR and ANN models. Furthermore, some studies [52,76] have expanded the ML application scope to simultaneous prediction of multiple endpoint targets, departing from the more common singular focus approach. However, potential data bias from substantial sample reduction in certain studies calls for a careful interpretation of the results [41,61,82] …”

Section: Analysis Of ML Use Cases In Bofmentioning

confidence: 99%

See 2 more Smart Citations

Toward learning steelmaking—A review on machine learning for basic oxygen furnace process

Ghalati,

Zhang,

El‐Fallah

et al. 2023

Materials Genome Engineering Advances

View full text Add to dashboard Cite

Basic oxygen furnace (BOF) steelmaking is the most widely used process in global steel production today, accounting for around 70% of the industry's output. Due to the physical, mechanical, and chemical complexities involved in the process, conventional monitoring and control methods are often pushed to their limits. The increasing global competition has created a demand for new methods to monitor and control the BOF steelmaking process. Over the past decade, Machine Learning (ML) techniques have garnered substantial attention, offering a promising pathway to enhance efficiency and suitability of steel production. This paper presents the first comprehensive review of ML applications in the BOF steelmaking process. We provide an introduction to both fields: an overview of the BOF steelmaking process and ML. We analyze the existing work on ML applications in BOF steelmaking and synthesize common concepts into categories, supporting the identification of common use cases and approaches. This analysis concludes with the elaboration of challenges, potential solutions, and a future outlook for further research directions.

show abstract

Section: Conclusion and Future Research Directionsmentioning

confidence: 99%

Section: Data‐driven Modeling Workflow For Bof Processmentioning

confidence: 99%

Section: Analysis Of ML Use Cases In Bofmentioning

confidence: 99%

See 1 more Smart Citation

Toward learning steelmaking—A review on machine learning for basic oxygen furnace process

Ghalati,

Zhang,

El‐Fallah

et al. 2023

Materials Genome Engineering Advances

View full text Add to dashboard Cite

show abstract

“…erefore, the accuracy of information transmission can be guaranteed when the input adjacency matrix remains unchanged [26].…”

Section: Mobile Information Systemsmentioning

confidence: 99%

Unsupervised Community Detection Algorithm Based on Graph Convolution Network and Social Media

Zhou

Zhang

2022

Mobile Information Systems

View full text Add to dashboard Cite

In view of the difficulty and low efficiency of most existing algorithms in detecting large-scale community networks, an unsupervised community detection algorithm based on graph convolution networks and social media is proposed. First, some positive and negative sample nodes are labeled according to the node similarity to complete the graph segmentation. Then, the improved graph convolution network model is used for training to obtain the local community where the given starting node is located. Finally, the local community is optimized by setting the threshold of membership degree, so as to further screen the nodes outside the community and obtain accurate community detection results. The experimental analysis of the proposed algorithm based on Flixster, Douban, and Yelp datasets shows that when the number of community divisions is 12, the modularity values on the three datasets are 0.59, 0.62, and 0.69, respectively, and the standard deviations of F1 are 0.044, 0.048, and 0.040, respectively. Overall, the proposed unsupervised community detection algorithm has better robustness.

show abstract

“…[24] A convolutional neural network is one of the main representatives of deep learning, and its structure better simulates the actual biological neural system, which has great advantages in the fields of pattern recognition and statistical prediction and has achieved good application results. [25][26][27][28][29] Thus, in this study, the CNN was used to establish a converter end point oxygen content prediction model, and the optimal network structure was determined to obtain better prediction results than BP neural network.…”

Section: Introductionmentioning

confidence: 99%

Convolutional Neural Network‐Based Method for Predicting Oxygen Content at the End Point of Converter

Wang

Bao

2022

steel research int.

View full text Add to dashboard Cite

The main factor affecting the cleanliness of steel is inclusions, most of which are deoxidation products. The greater the oxygen content fluctuation at the converter's end point, the more unstable the inclusion control. A large amount of charging, operation, and sampling data is collected during the converter smelting, but it is not fully utilized, and the deoxidizing alloy addition still relies on manual experience. Herein, the representative convolutional neural network (CNN) of deep learning is adopted, and the number of convolutional layers, convolutional kernel size, and the number of convolutional kernels are optimized to establish the best prediction model of the converter end point. Its root mean square error, mean absolute error, and mean absolute percentage error are 35.29%, 25.59%, and 7.30%, respectively, which are superior to the back propagation neural network. This CNN prediction model uses 1300 sets of preprocessed data containing 23 indicator variables, of which 1200 sets of data are used for training and 100 sets for model validation. Within the ±50 ppm error scope, the model prediction hit rate can reach 87%, and when within the ±70 ppm error scope, the hit rate is 93%.

show abstract

Multi-Channel Graph Convolutional Network based End-Point Element Composition Prediction of Converter Steelmaking

Cited by 9 publications

References 12 publications

Toward learning steelmaking—A review on machine learning for basic oxygen furnace process

Toward learning steelmaking—A review on machine learning for basic oxygen furnace process

Unsupervised Community Detection Algorithm Based on Graph Convolution Network and Social Media

Convolutional Neural Network‐Based Method for Predicting Oxygen Content at the End Point of Converter

Contact Info

Product

Resources

About