基于核偏最小二乘法的热轧板凸度预测

Prediction of Hot Rolled Plate Crown Based on Kernel Partial Least Squares

以某2160 mm热连轧生产线为研究对象,结合生产数据和数据驱动算法,提出并建立基于核偏最小二乘法和偏最小二乘法的热轧板凸度预测模型,并对模型的参数进行调整和优化。结果表明:偏最小二乘法模型的预测值与实测值误差为-25~30μm,88.74%的板凸度预测值绝对误差小于10μm;核偏最小二乘法模型的预测值与实测值误差为-25~25μm,91.57%的板凸度预测值绝对误差小于10μm。因此,优化后的核偏最小二乘法模型具有更好的预测效果。

Taking a 2160 mm hot rolling production line as research object,combined with production data and data-driven algorithms,a prediction model for hot rolled plate crown based on kernel partial least squares(KPLS)method and partial least squares(PLS)method was proposed and established,and the parameters of the model were adjusted and optimized.The results indicated that the error between the predicted and measured values of the PLS model was -25 to 30μm,and 88.74% of the prediction data of plate crown had an absolute error of less than 10μm;the error between the predicted and measured values of the KPLS model was-25 to 25μm,and 91.57% of the prediction data of plate crown had an absolute error of less than 10μm.Therefore,the optimized KPLS model had better prediction effect.