Application of BP neural network to semi-solid apparent viscosity simulation

1　INTRODUCTIONSincethethixotropic propertiesofvigorouslystirredtin leadslurrieswerediscoveredbySpenceretal[1] atMIT ,investigationandapplicationofthese mi solidtechnologyhavebeendevelopedquickly .Thestudyontheapparentviscosityisnotonlyoftheoreti calbutalsoofutilizationsignificance .Manyinvesti gatorshavecarriedoutagreatdealofexperimentalresearchandputforwardmanymathematicalmodelsofthesemi solidapparentviscosity .Someofthesemodels ,whicharerelatedto popularexponentialfunctionandaclassica…

Application of BP neural network to semi-solid apparent viscosity simulation

Two-layer BP neural network was designed for the semi-solid apparent viscosity simulation. The apparent viscosity simulations of Sn-15%Pb alloy and Al-4.5%Cu-1.5%Mg alloy stirred slurries were carried out. The trained BP neural network forecast the curve of the apparent viscosity versus solid volume fraction of Sn-15%Pb alloy, under the condition of shear rate, =150 s-1, and cooling rate of G=0.33 ℃/min. The simulation results are well agreement with the experimental values given in references The fitted mathematical formula of Sn-15%Pb alloy apparent viscosity, under the condition of the cooling rate of G=0.33 ℃/min, was obtained by optimization method. The results show that the precision of apparent viscosity simulation value by neural network is much better than that of its calculation value by fitted mathematical formula.