Numerical Simulation And Experimental Research On The Temperature Field Of High-Rigidity Coating Sphere Grinding

High-performance sphere valve, which can resists high temperature, high pressure, wearing, thermal impact and heat erosion, is one of key components used in coal liquefaction engineering, steel smelting, petroleum chemical engineering and nuclear power. In order to cope with such rigorous work requirements, the spool is generally made with steel 45 as the substrate, and coated with WC-Co powder as the enhanced layer by high-speed flame spraying, plasma spraying and other thermal spraying technologies. Then the spool is processed by grinding. However, compared to other processing methods, grinding needs more energy to remove per unit volume material, which may produce high temperature and affect the surface quality. Although the temperature field of surface grinding has a complete study, but the temperature field of sphere grinding, especially the temperature field of high-rigidity coating sphere grinding with bowl wheel has not been studied yet. Therefore according to the machining method of high-rigidity coating sphere grinding with bowl wheel, the temperature field is studied by both ANSYS simulation and experiments. Then the article tries to predict the grinding temperature by using the artificial neural network, and reduce the grinding temperature by improving the bowl wheel. The concrete content is as follows: The model of high-rigidity sphere grinding with bowl wheel is established by using ANSYS software. In order to obtain the sphere temperature field, rotate arc heat is loaded to the workpiece surface by APDL language to simulate the process of grinding with bowl wheel. At the same time, the influence of the workpiece rotation speed, the bowl wheel size and the heat-flow density to the sphere temperature field is studied by changing the input variables.The finite element model is verified by comparing the simulation results to the measured results which are obtained through the experiments of high-rigidity sphere grinding with bowl wheel. The influence of the grinding depth, the workpiece rotation speed and the bowl wheel rotation speed to the grinding temperature is also analyzed. In order to reduce the grinding temperature, the bowl wheel is divided to several grinding blocks which are fixed to the wheel evenly. It is found that the grinding temperature is lower in sphere grinding with blocked bowl wheel than with normal bowl wheel under the same grinding conditions.Considering the non-linear approximation characteristic of the artificial neural network, a BP neural network which has been improved by the batch training and Levenberg-Marquardt algorithm and trained by the orthogonal test results is established to predict the grinding temperature in high-rigidity sphere grinding with bowl wheel quickly as to guide the selection of parameters in practical grinding.Finally, the paper summarizes all the done work and views the study direction in future.