According to the characteristics of strong coupling of variables for universal rolling line of high speed rail,taking the rolling precision of rail section as the goal, it established an optimization model of rolling parameter of rail by combining the optimization-range algorithms with neural network.Corresponding relation among rolling course data and rail section gather data was set up, the dominant interzone was utilized to obtain initial value of roll gap and rolling strength under the particular working condition after preconditioning of the data, and the neural network was used to revise initial value. It has proved that this model can reduce the influence of the subjective factors and improve the rolling precision of the rail. 

针对高速钢轨万能轧制线多变量强耦合的特征,以钢轨断面轧制精度为目标,将优势区间控制算法与神经网络相结合,建立了钢轨轧制参数优化模型。将生产现场采集的大量轧制过程数据与钢轨断面检测数据建立对应关系,数据预处理后利用优势区间控制算法求出特定生产条件下的辊缝及轧制力的初始值,再利用神经网络对初始值进行偏差修正。实践证明,该模型能够减少主观因素的影响,提高钢轨的轧制精度。

钢轨, 优势区间, 神经网络, 轧制参数