Multi-working condition performance assessment based on knowledge extraction of optimal operating states for continuous annealing processes

Abstract

Performance assessment is a key to strip quality improvement and energy consumption reduction of Continuous Annealing Processes (CAP). However, existing methods focus on performing the assessment under a single working condition, and the assessment accuracy must be improved. This study proposes a new multi-working-condition performance assessment method based on the knowledge extraction of the optimal operating states for CAP. First, a mechanism–data fusion-based assessment index construction method is proposed for the key parameter selection. Second, a knowledge extraction strategy for the optimal operating states under multiple working conditions is proposed to construct a benchmark library. Third, a knowledge-enhanced assessment model is built to achieve qualitative performance evaluation and quantitative non-optimal traceability. The experiment based on the process data shows the effectiveness of assessing the operating performance, providing decision guidance for strip quality improvement and energy consumption reduction.

Keywords:

• Continuous annealing process
• key parameter selection
• multiple working conditions
• benchmark library
• performance assessment

Acknowledgments

The authors extend their gratitude to Hunan Lianyuan Steel for their contributions to this project. Their provision of industrial data and numerous valuable suggestions greatly enhanced this endeavour. We are also grateful to Leyu Bi and Linwei Guo from China University of Geosciences for providing many useful suggestions.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data not available due to commercial restrictions.

Additional information

Funding

This work was supported by the National Natural Science Foundation of China [grant number 61773354], the Natural Science Foundation of Hubei Province [grant number 2020CFA031], the 111 project [grant number B17040], and the Fundamental Research Funds for National Universities [grant number CUGDCJJ202210].