Study On Temperature Field Model Of Medium And Heavy Plate During Rolling And Laminar Cooling Based On CUDA

Rolling and laminar cooling processes are critical to the final shape and quality of the medium and heavy plate.And the prerequisite for improving the final performance of the plate is to establish an accurate temperature field model.Most of the current temperature field models are modeled separately for rolling and laminar cooling processes,and there are some problems such as oversimplification of the model,excessively ideal parameter setting,and excessive calculation time.However,in the actual production process,The final temperature of plate rolling process directly affects the accuracy of the laminar cooling process model,complex environmental changes also have a great impact on the parameters in the model.And most of the current models improve the speed of the model by reducing the grid points and simplifying the calculation formula of the heat transfer coefficients,although this will have a great impact on the accuracy of the model.In the view of above problems,in this thesis,the rolling-laminar cooling process is integrated into a temperature field model,according to the temperature data collected by the experiment,the heat transfer coefficients in the model are inversely solved,and under the CUDA programming architecture,the parallel computing of the GPU is used to improve the calculation speed of the model.The main research contents of this thesis are as follows:(1)Establishment of temperature field model for rolling-laminar cooling process.In this thesis,based on the thickness and width directions of the plate,a two-dimensional temperature field model integrating the rolling-laminar cooling process is established.And according to the heat transfer characteristics of different stages in the process,different boundary conditions and initial conditions are used.The model is solved by the finite volume method,and the temperature distribution inside the plate is obtained.(2)Solution of heat transfer coefficients based on inverse problem algorithm.According to the research characteristics of the subject,the inverse calculation results of genetic algorithm and particle swarm optimization algorithm applied to the temperature field model are analyzed,and the shortcomings of the particle swarm optimization algorithm when the search range is expanded are improved.And because the heat transfer intensity of different surfaces of the plate is different,the heat transfer coefficients of the upper and lower surfaces of the plate are inversely calculated.(3)Parallel design and implementation of the temperature field model.In this thesis,under the CUDA programming framework,the different computing tasks of the temperature field model are divided.The GPU is used to calculate the temperature field model in parallel,and the parallel program is optimized by changing the GPU memory usage and the configuration of the kernel function thread.(4)Verification and analysis of the temperature field model.In this thesis,the calculation results of the temperature field model at different stages are compared with the experimental data to verify the accuracy of the temperature field model,and the calculation speed of the parallel model and the serial model are compared to verify the speed performance of the parallel model.Finally,the acceleration ratio of the parallel temperature field model is 2.058,and compared with the experimentally measured temperature values,the calculation error of the model is within 10℃,achieves good results in terms of speed and accuracy,laying a good foundation for subsequent related research and application.