【摘要】 为了提高冷连轧过程中板形预设定和闭环反馈的控制效果，以1 450 mm五机架UCM冷连轧机组为研究对象，对1 742个实验数据进行分类和预处理，以74个工艺参数变量作为输入特征，20个不同位置的板形值作为输出结果，构建了反向传播(backpropagation, BP)神经网络模型，并采用遗传算法(genetic algorithm, GA)进行优化，得到了基于遗传算法的反向传播(GA-BP)神经网络模型.结果表明，所构建的GA-BP神经网络模型在拟合优度、预测精度和稳定性等方面均优于BP神经网络模型，其RMSE值从0.981 8 I降至0.447 6 I,MAE值从0.622 5 I降至0.219 3 I,R~2由0.745 4增至0.913 1.更多还原

【Abstract】 To enhance the preset and closed-loop feedback control of strip shape during the tandem cold rolling process, a study was conducted on a 1 450 mm five-stand UCM tandem cold rolling mill. A total of 1 742 experimental data points were classified and preprocessed, using 74 process parameter variables as inputs and strip shape values at 20 different positions as outputs. A backpropagation(BP) neural network model was constructed and optimized using a genetic algorithm(GA), to obtain a genetic algorithm-based backpropagation(GA-BP) neural network model. The results indicate that the GA-BP model outperforms the BP model in terms of goodness of fit, prediction accuracy, and stability. The RMSE decreases from 0.981 8 I to 0.447 6 I, the MAE decreases from 0.622 5 I to 0.219 3 I, and the R~2 increases from 0.745 4 to 0.913 1.更多还原