Practical application of learning classifier system for downcoiler decision support in steel hot strip mill
暂无分享,去创建一个
The aim of this project is to improve the quality and consistency of coiling in a hot strip mill at British Steel Strip Products' integrated works. The artificial intelligence technique of learning classifier systems (LCSs) is proposed for the processing of plant data. The stochastic computational technique of LCSs will produce offline rules to aid operator and engineering decision making. These rules link the plant inputs (plant condition, strip properties, and associated variables) to coil outputs (presentation -including telescoping and pinching) in a form that is capable of being verified and validated. This is central to the initial operation, where online data will produce offline rules that are critically evaluated by a human operator before implementation. Improvements to a basic LCS that allow operation on industrial data are detailed. Initial experimental results show that the technique of LCSs has the potential to become a very useful tool for processing industrial data. Improvements in availability, coil presentation, and ultimately customer satisfaction will result in a cost benefit to British Steel plc.