论文部分内容阅读
为了提高轧制力自学习模型的预报精度,将传统自学习模型的预报轧制力及影响轧制力的主要因素作为网络的输入,利用权值更新次数的倒数与单个样本本次激活的地址更新次数倒数和的比作为网络权值更新的信度,建立了基于信度分配的小脑模型CA-CAMC网络与轧制力自学习相结合的轧制力预报模型。通过大量在线数据实验分析,结果表明基于CA-CAMC网络模型的轧制力预报模型的精度高、稳定性好,能够更好地满足实际生产中越来越高的控制精度需求。
In order to improve the prediction accuracy of the self-learning model of rolling force, the main factors that affect the rolling force and the rolling force of the traditional self-learning model are regarded as the input of the network. The reciprocal of the number of updating of the weight and the address of the activation of the single sample The ratio of the number of updates to the countdown is used as the reliability of the network weight renewal to establish the rolling force prediction model based on the distribution of the cerebellum CA-CAMC network with rolling force self-learning. The experimental results show that the rolling force prediction model based on the CA-CAMC network model has high precision and good stability, which can better meet the increasingly higher control accuracy requirements in actual production.