论文部分内容阅读
采用统计学方法分析CSP的轧制数据,发现存在严重的自相关现象。结合工艺分析了自相关的成因,采用主成分分析法和岭回归法进行了分析以消除自相关性。通过对志田茂模型的拟合,验证了自相关的不利影响,并应用统计学对数据进行了部分剔除。拟合了志田茂改进模型和多元回归模型,精度得到提高。引入再结晶分数模型后,轧制力模型得到了较好精度。采用新的30个数据对5个模型进行了对比验证,结果显示考虑再结晶的模型5精度较好。
Using statistical methods to analyze the CSP rolling data, we found that there was a serious auto-correlation phenomenon. The causes of autocorrelation were analyzed by the combination of techniques. Principal component analysis and ridge regression were used to eliminate the autocorrelation. By fitting Chi Shimao model, the adverse effects of autocorrelation were verified, and the data were partly eliminated by using statistics. Fitting Chi Tianmao improved model and multiple regression model, the accuracy is improved. The introduction of recrystallization fraction model, the rolling force model has been better accuracy. The new 30 data were used to verify the five models. The results show that the model 5 considering the recrystallization has better accuracy.