基于性能预报技术减少热轧带钢取样频率的探索

Exploration of hot-rolled strip sampling frequency reduction based on mechanical property prediction technology

性能预报技术是一件具有深远影响的技术,但实现在线应用需系统考虑解决一些约束问题,目前在线应用的报道并不多见。宝钢研究了一种基于性能预报模型对现有热轧产品取样管理流程的优化方法,它通过建立带钢力学性能预报模型,将产品取样管理流程与带钢的实际生产过程相关联,对多数合格带钢实现预报替代取样,而对少数异常带钢进行有针对性地实际取样,从而达到减少取样频率的目的。基于现有生产执行系统(MES系统)的应用实践,探索出一套基于性能状态优化中间品采样频率的方案,实现了性能预报模型的大生产在线应用。该方法依托信息技术和统计技术,既实现了取样的精细化管理,又降低了运营成本。实际应用效果表明,基于性能预报模型的新取样管理方法不但没有降低可靠性,而且可以降低69%的中间品取样量;新方法在加快生产物流上效果也非常明显,可以节省23 h,这些都说明该方法是一项具有实际应用价值的技术。

＂Mechanical property prediction＂ is a far-reaching influence technology.It is difficult to carry out the online application because of some constraint conditions and it is also seldom reported.Baosteel technicians developed a method of optimizing the existing sampling management process for hot-rolled strip based on the mechanical property prediction.Establishing a model,the technicians associated the sampling process with the actual production process and substituted prediction for sampling to most qualified strips.But to other unqualified strips,it needs sampling with pertinence actually,so as to reduce sampling frequency.On the basis of the present manufacturing excutive system（MES）,a plan to optimize the sampling frequency of intermediate goods by their property states was developed and applied to the mass production online.According to the application,this method based on ＂mechanical property prediction＂ technology can obviously reduce about 69 percent of sampling quantity of intermediate goods and also speed up production logistics by 23 hours.It demonstrates that the new method has a practical application value in technology.