Abstract
Purpose: The purpose of this paper is to reduce the crop losses at two ends and both sides of a plate in the rolling process, to produce a rectangular plan view pattern plate and to enhance the total product yield of the plates. Design/methodology/approach: Based on sample data and the unchangeable principle of the slab volume in the rolling process, the predictive MAS control models were set-up. They are width broad MAS predictive model, width broad MAS control model, gaugemeter automatic gauge control (GM-AGC) model, and plates tracking model. After the models were tuned, the rolling test was implemented at LinFen Iron & Steel Co., Ltd. Findings: It is found that by accurately predictive online modeling of MAS methods, the plan view pattern control can be applied in plate mills. Research limitations/implications: As the rolling process is in high temperature, the plan view pattern is difficult to be detected. Normally, the real-time abnormity distortion cannot be obtained. Practical implications: The test results showed that the crop losses are reduced and the product yield is greatly increased. Originality/value: This paper presents an accurately predictive online modeling of MAS method.
Original language | English |
---|---|
Pages (from-to) | 1351-1358 |
Number of pages | 8 |
Journal | Kybernetes |
Volume | 39 |
Issue number | 8 |
DOIs | |
Publication status | Published - 2010 |
Keywords
- Control systems
- Cybernetics
- Metal refining
- Modeling
- Pattern recognition