作　　者： ; ; ; ; ;

机构地区： 东北大学材料与冶金学院轧制技术及连轧自动化国家重点实验室

出　　处： 《钢铁》 2008年第1期58-63,共6页

摘　　要： 以SPHC为研究对象,通过遗传算法改善神经网络参数,获得最佳预测能力的神经网络,建立了带钢化学成分等因素对产品性能影响的定量关系。考虑了生产的条件和能力以及用户对力学性能的要求,通过遗传算法计算轧制过程中的温度制度参数,实现组织性能的柔性化轧制。研究结果表明,实际测量的3个温度制度参数包含在预测的温度制度参数范围内。在生产能力允许的条件下,温度制度可以控制的屈服强度波动范围约为±10%。采用遗传算法能高效地计算温度制度参数。 The SPHC steel was studied with a neural network optimized by genetic algorithm and quantitative relationship between mechanical properties of strip and influencing factors, such as chemical composition etc,was established. With the consideration of conditions and capacity of production line, consumers＇ demands for mechanical properties, the flexible rolling for microstructure property was realized by calculating temperature schedule accord- ing to genetic algorithm. It was shown that, three measured temperature schedules are within the predicted range. Yield strength of strip can be controlled within ±10%. It is efficiently to use genetic algorithm for calculating temperature schedule.

关 键 词： 热轧 柔性化轧制 精轧入口温度 精轧出口温度 卷取温度

领　　域： [金属学及工艺]