No abstract

Control cooling is essential method for microstructure and mechanical properties control in hot rolling strip making. It is vital to realize high precises temperature distribution prediction and control in cooling process to ensure the industrial production. In this paper, a traditional mechanism model based on finite-difference method and combining with online cycle velocity calculation strategy was introduced as one of estimating temperature distribution baseline method. However, considering calculation time, variable-velocity rolling makes it difficult to rapidly realize temperature and water distribution modifying of each segment in cooling zone. Herein, a temperature distribution prediction method based on recurrent neural network was proposed, instead of only final cooling temperature prediction. And temperature distribution prediction performance of model with different recurrent cell and time steps were evaluated.The results indicated that the proposed model could realize temperature distribution prediction and the model based on bi-LSTM and 48 timesteps has the highest determination coefficient value of 0.976 and lowest root mean square error of 8.03 and mean absolute error of 5.7. Furthermore, compared with baseline model, the proposed model retained lower computational cost, making it applicable in industrial application by providing real-time temperature distribution prediction.

Large residual stress occurs during the quenching process of hot-rolled seamless steel tubes, which results in bending, cracking, and ellipticity exceeding standards and seriously affects the quality of hot-rolled seamless steel tubes. In addition, the stress generation mechanism of hot-rolled seamless steel tubes is different from that of steel plates due to the characteristics of annular section. In this research, the finite element simulation method was used to study the quenching residual stress of seamless steel tubes with different cooling intensities. The variation law of temperature and stress on the steel tube under different cooling intensities were analyzed. The results show that the radial stress was close to 0, and the circumferential and axial stresses were the main factors affecting the quality of the steel tube. With the increase in the cooling time, the magnitude and direction of each stress component of the steel tube changed simultaneously. Finally, a typical stress distribution state of “external compressive stress, internal tensile stress” was presented in the thickness direction of the steel tube. Furthermore, with the increase in the cooling intensity, the residual stress of the steel tube gradually increased and was mainly concentrated on the near wall of the steel tube.

No abstract

Control cooling is essential method for microstructure and mechanical properties control in hot rolling strip making. It is vital to realize high precises temperature distribution prediction and control in cooling process to ensure the industrial production. In this paper, a traditional mechanism model based on finite-difference method and combining with online cycle velocity calculation strategy was introduced as one of estimating temperature distribution baseline method. However, considering calculation time, variable-velocity rolling makes it difficult to rapidly realize temperature and water distribution modifying of each segment in cooling zone. Herein, a temperature distribution prediction method based on recurrent neural network was proposed, instead of only final cooling temperature prediction. And temperature distribution prediction performance of model with different recurrent cell and time steps were evaluated. The results indicated that the proposed model could realize temperature distribution prediction and the model based on bi-LSTM and 48 timesteps has the highest determination coefficient value of 0.976 and lowest root mean square error of 8.03 and mean absolute error of 5.7. Furthermore, compared with baseline model, the proposed model retained lower computational cost, making it applicable in industrial application by providing real-time temperature distribution prediction.