Design, Fabrication And Motor Application Of Fe-Based Amorphous/Nanocrystalline Alloy Based On Machine Learning

Fe-based amorphous alloy has the advantages of excellent soft magnetic performance,good thermal stability and great corrosion resistance,so the application trend is favorable in power electronic system,such as transformer,motor core,5G communication,wireless charging and other fields.The demand for high performance and miniaturized electronic components as well as the corresponding excellent performance Fe-based amorphous materials makes the design of new materials urgent.However,it is difficult to design new ferro-based amorphous alloys because of the complexity of grouping elements and unclear correlation between properties and composition.In this paper,the alloy composition with high saturation magnetic induction intensity（B_s）was designed by machine learning method,its soft magnetic properties,corrosion resistance and crystallization properties were prepared,characterized,tested and verified.Then,Ansoft finite element software is used to simulate the magnetic density distribution,groove torque and hysteresis loss characteristics of Fe₈₄Si₂P₃B₉C_1.5Cu_0.5 material designed in this work as motor core,and the influence of the material as motor core is calculated.The main work of this paper is as follows:（1）Collecting and constructing a set of 200 B_s characteristics and H_ccharacteristics data of Fe-based amorphous materials,and trained the B_s and H_c calculation of Fe-based amorphous materials using Linear model,neural network model（NN）,random forest regression model（RFR）and support vector machine model（SVR）respectively.The goodness of fit for B_s prediction was 0.725,0.952,0.964 and 0.793,and the goodness of fit for coercivity prediction was 0.255,0.857,0.964 and 0.508,respectively.The neural network model has the best prediction performance.Fe₈₄Si₂P₃B_10.5-xC_xCu_0.5（X=0,0.5,1,1.5,2,2.5）alloy with high Bs was designed by machine learning method in Fe-Si-P-B-C-Cu system.（2）The Fe-based amorphous alloy was prepared,and the amorphous structure was found by XRD and TEM tests.VSM test shows that B_s of amorphous alloy increased with the increase of C content.After 300 s isothermal annealed at 653 K and 773 K,B_sincreased first and then decreases,H_c increases.The experimental B_s results were close to the machine learning results.Through electrochemical test,the increased of C content is beneficial to improve the corrosion resistance of Fe-based amorphous.The activation energies of Fe₈₄Si₂P₃B₉C_1.5Cu_0.5 were calculated by Kissinger’s method and Ozawa’s method,KAS and FWO methods were used to calculate local activation energies,and JMAK formula was used to calculate Avrami index n.Ex>Ep,the energy required for nucleation during crystallization is greater than that required for grain growth,and nucleation is more difficult than grain growth.According to Avrami index n,there are two-dimensional and three-dimensional growth in the crystallization process.（3）Ansoft finite element software was used to simulate Fe₈₄Si₂P₃B₉C_1.5Cu_0.5 as motor core,explored the difference between Fe₈₄Si₂P₃B₉C_1.5Cu_0.5 and silicon steel motor,calculate and analyzed the influence of Fe₈₄Si₂P₃B₉C_1.5Cu_0.5 as motor core on motor magnetic density distribution,groove torque,hysteresis loss and core loss.The simulation results show that the Fe-based amorphous alloy can effectively reduce the loss of the motor and has better performance than the silicon steel motor.