基于半监督深度网络的冷连轧轧制力预报

Prediction of cold continuous rolling force based on semi-supervised deep network

针对冷连轧生产中难以建立准确的轧制力数学模型的问题,提出了基于半监督深度网络的轧制力预报模型。首先,使用堆叠去噪自编码器逐层提取输入数据的高阶特征表示。为提高特征提取的有效性,根据输入值与目标值的相关性程度,对其各维度特征损失函数施加不同比例,构成比例损失堆叠去噪自编码器。然后,使用比例损失堆叠去噪自编码器提取的高阶特征初始化深度网络,对目标值进行预测。仿真结果表明,该模型预测精度可控制在3%以内,实现了轧制力的高精度预测。

In order to solve the problem of establishing mathematical model of rolling force accurately,a rolling force prediction model based on semi-supervised deep network was proposed. Firstly,a high level feature representation of input data was extracted layer by layer using a stacked denoise autoencoder. To improve the validity of feature extraction,according to the degree of correlation between the input value and the target value,different proportions were applied to the feature loss function of each dimension to form a proportional loss stack denoise autoencoder. Then,the deep network was initialized using the features extracted by proportional loss stack denoise autoencoder to predict the target value. The simulation results show that the prediction accuracy of the model can be controlled within 3%,and the high precision prediction of rolling force is realized.