基于支持向量回归的二次冷轧轧制力预报模型

A rolling force prediction model based on support vector regression for double cold reduction

为提高二次冷轧兼平整机组在二次冷轧模式下轧制力的预报精度,建立了一种基于摩擦系数自学习的轧制力预报模型。考虑到摩擦系数自学习模型的不足,为进一步提高轧制力的预报精度,提出了一种支持向量回归预测轧制力的计算误差与摩擦系数自学习相结合的轧制力预报方法。结果表明,该模型的计算值与实际值吻合较好,误差控制在±7%以内,满足现场生产要求,具有较高的工程应用价值。

Double cold reduction strip featured by thin thickness and high strength, is widely used in electron, can-making fields and so on. In order to improve the precision of rolling force prediction of a certain double cold reduction and temper rolling mill, a friction coefficient self-learning model is established to predict the rolling force. Considering the deficiency of the friction coefficient self-learning model, the support vector regression is adopted to predict the calculation error of the roiling force, and then the friction coefficient self-learning model combined with the error is used to obtain the high accurate rolling force. The results show that the predicted rolling force is in agreement with the measured values with error within ±7 %. the model satisfies the actual conditions and is valuable to engineering application.