Research On Optimization Of Copper Powder Preparation Process Based On Neural Network Atomization

With the rapid development of new technologies,the requirements for material performance in various fields are becoming higher and higher,and the material preparation industry urgently needs a breakthrough.This article uses the atomization equipment of Anhui Dingheng Industrial Group Co.,Ltd.,This article discusses and analyzes the control parameters of the atomizing equipment,such as the position of the molten metal outlet,the temperature of the molten metal,the diameter of the conduit and the pressure of the atomizing inlet on the performance of the atomizer.Optimization of atomized copper powder and atomization preparation process parameters.However,due to the high cost of the experiment,the "ANSYS Fluent 19.2" fluid dynamics software was used to simulate the atomized airflow field Obtain the parameter range of the atomizing intake pressure,and then use multi-factor multi-level visual design to design three-factor multi-level 10 sets of experiments,Establish an artificial neural network analysis method to process experimental data,provide the best process parameters for the preparation of atomized copper powder,and verify the preparation of atomized copper powder to optimize the process parameters,conclusion as below:(1)Use "ANSYS Fluent 19.2" fluid dynamics software to conduct computer simulation analysis on the flow field structure of the dual-flow atomizing ring-hole nozzle,When the exit distance of the molten metal is 5 mm,the structure of the gas flow field inside the atomization chamber has a good influence on the crushing of the molten metal and the molding of the copper powder.When the air pressure value of the atomization inlet is 0.4MPa,the powder atomization interval is small,and the powder atomization interval between 0.5MPa～0.9MPa gradually becomes larger,However,the internal structure of the atomization zone is complex and variable.After reaching 0.9MPa,the structure of the atomizing zone will not fluctuate significantly with the increase of the pressure value of the atomizing inlet.The actual pressure of the atomizing inlet is set to 0.5MPa～0.9MPa.(2)Using multi-factor and multi-level design methods,10 sets of different levels of atomization process parameters were designed according to the simulation data,and 10 sets of experimental test data were obtained.Using the "leave two method" for data processing,a 3-8-8-5 four-layer topology artificial neural network model was established,and the model error was as low as 10-7.(3)The artificial neural network model was used to explore the influence of different metal melt temperature,catheter diameter and atomization inlet air pressure on the performance of atomized copper powder and the optimization of atomization preparation process parameters.The best atomization process parameters:metal melt temperature is 1220℃,conduit diameter is 6.10mm,atomization inlet pressure is 0.82MPa,copper powder performance data simulated by the model during this process is oxygen content 273.04ppm;bulk density is 4.49 g/cm3;fluidity is 14.11s/50g;tap density is 5.38g/cm3;yield is 61.12%.(4)The optimized process parameters are used to verify the preparation of copper powder.Under this process condition,the atomized copper powder had an oxygen content of 271.61 ppm,a bulk density of 4.48 g/cm3,a fluidity of 14.92 s/50 g,a tap density of 5.40 g/cm3,and a yield of 59.51%.The prediction error value of the experimental data and artificial neural network simulation data is very small,and the optimized performance of the atomized copper powder fully meets the requirements of the metal 3D printing powder enterprise standard Q/DH102.090.