Flatness Prediction of Cold Rolled Strip Based on Deep Neural Network with Improved Activation Function

With the improvement of industrial requirements for the quality of cold rolled strips, flatness has become one of the most important indicators for measuring the quality of cold rolled strips. In this paper, the strip production data of a 1250 mm tandem cold mill in a steel plant is modeled by an improved deep neural network (the improved DNN) to improve the accuracy of strip shape prediction. Firstly, the type of activation function is analyzed, and the monotonicity of the activation function is deemed independent of the convexity of the loss function in the deep network. Regardless of whether the activation function is monotonic, the loss function is not strictly convex. Secondly, the non-convex optimization of the loss functionextended from the deep linear network to the deep nonlinear network, is discussed, and the critical point of the deep nonlinear network is identified as the global minimum point. Finally, an improved Swish activation function based on batch normalization is proposed, and its performance is evaluated on the MNIST dataset. The experimental results show that the loss of an improved Swish function is lower than that of other activation functions. The prediction accuracy of a deep neural network (DNN) with an improved Swish function is 0.38% more than that of a deep neural network (DNN) with a regular Swish function. For the DNN with the improved Swish function, the mean square error of the prediction for the flatness of cold rolled strip is reduced to 65% of the regular DNN. The accuracy of the improved DNN is up to and higher than the industrial requirements. The shape prediction of the improved DNN will assist and guide the industrial production process, reducing the scrap yield and industrial cost.

Evolution of Microstructure, Mechanical Properties and Residual Stress of a Cold Rolled Invar Sheet Due to Heat Treatment

Invar alloy possesses a uniquely low coefficient of thermal expansion, making it an ideal material for fine metal masks. To manufacture fine metal masks, Invar alloys are often cold-rolled, during which residual stress develops. Heat treatment is an effective means to control residual stress that develops within Invar sheets after cold rolling, but the treatment should be carried out with care. In this article, a comprehensive study on the effect of heat treatment on the residual stress, microstructure, and mechanical properties of a cold-rolled Invar sheet is reported. We show that while both recovery and recrystallization are effective means of reducing residual stress, substantial microstructural changes and, therefore, notable changes in mechanical properties and residual stress, occur after recrystallization. Moreover, residual stress release due to recrystallization can be affected by microstructure and texture prior to heat treatment as these factors play a significant role in recrystallization.

Effect of Cold Rolling on Cluster(1) Dissolvability during Artificial Aging and Formability during Natural Aging in Al-0.6Mg-1.0Si-0.5Cu Alloy

Natural aging after solution treatment has a negative effect on the precipitation strengthening of Al–Mg–Si alloys since Cluster(1) formed at a room temperature cannot be dissolved or transformed into precipitates during artificial aging at 170 °C. In this study, cold rolling is focused on as an alternative solution to pre-aging, which is a conventional method to prevent Cluster(1) formation. It is known that excess vacancies are necessary for cluster formation. Cold rolling suppresses cluster formation because excess vacancies disappear at dislocations introduced by cold rolling. In addition, it is expected that cold rolling accelerates the precipitation behavior because the diffusion of solute atoms is promoted by introduced lattice defects. The transition of Cluster(1) was evaluated by Micro Vickers hardness tests, tensile tests, electrical conductivity measurements and differential scanning calorimetry analyses. Results showed the negative effect of natural aging was almost suppressed in 10% cold-rolled samples and completely suppressed in 30% cold-rolled samples since Cluster(1) dissolved during artificial aging at 170 °C due to lowering of the temperature of Cluster(1) dissolution by cold rolling. It was found that the precipitation in cold-rolled samples was accelerated since the hardness peak of 10% cold-rolled samples appeared earlier than T6 and pre-aged samples.

Texture Evolution during Recrystallization and Grain Growth in Non-Oriented Electrical Steel Produced by Compact Strip Production Process

Evolution of texture and α*-fiber texture formation mechanism of Fe-0.65%Si non-oriented electrical steel produced by Compact Strip Production (CSP) process during all the thermo-mechanical processing steps were investigated using electron backscatter diffraction (EBSD) and X-ray diffraction (XRD) techniques. Columnar crystal structure of cast slab is fine and well-developed. Textures of the hot-rolled band are quite different in the thickness direction. During annealing of cold-rolled sheet, γ-fiber texture grains would nucleate and grow preferentially, and α*-fiber texture grains mainly nucleate and grow in the shear zone of α-fiber texture of cold-rolled sheet. During the recrystallization process, γ-fiber texture gradually concentrated to {111}<112>, and γ and α*-fiber texture increased significantly. {111}<112> texture priority nucleation at the initial stage of recrystallization. Due to the advantages of nucleation position and quantity, the content of α*-fiber texture is greater than {111}<112> texture in the mid-recrystallization. During grain growth process, {111}<112> oriented grains would grow selectively by virtue of higher mobility, sizes and quantity advantages than that of {411}<148 > and {100}<120>, resulting in the gradual increase of γ-fiber texture and the decline of α *-fiber texture.