基于BP 神经网络的汽车镁合金轮毂低压铸造工艺优化.pdfVIP

《热加工工艺》2023年4月第52卷第7期

基于BP神经网络的汽车镁合金轮毂

低压铸造工艺优化

程诚，宁萍，李玲玲3

(1.

学院，重庆400045)

摘要：基于BP神经网络建立的4×24x12x2四层结构汽车镁合金轮毂低压铸造优化模型，输人层为铸造模具预

热温度、浇注温度、浇注速度及充型压力，输出层为晶粒尺寸、屈服强度。通过BP神经网络优化模型对汽车镁合金轮毂

低压铸造成型工艺进行优化。结果表明：优化后的汽车镁合金轮毂铸件晶粒尺寸由146.8μm减小到123.2μm，晶粒尺

寸减小率为16.0%；屈服强度由169.7MPa提升到185.6MPa，屈服强度提升率为9.4%。汽车镁合金轮毂低压铸造最优

工艺参数为模具预热温度380℃、浇注温度690℃、浇注速度0.50m/s、充型压力6.5kPa。

关键词：BP神经网络；镁合金轮毂；低压铸造；工艺优化

D0I:10.14158/j.cnki.1001-3814

中图分类号：TG292文献标识码：A文章编号：1001-3814(2023)07-0077-04

OptimizationofLowPressureCastingProcessforAutomobileMgAlloy

WheelHubBasedonBPNeuralNetwork

CHENGCheng,NINGPing,LILingling

(1.ChongqingCreationVocationalCollege,Chongqing402160,China;2.SchoolofShippingShipEngineering,

ChongqingJiaotongUniversity,Chongqing400060,China;3.SchoolofMaterialsScienceandEngineering,Chongqing

University,Chongqing400045,China)

Abstract:BasedontheBPneuralnetwork,the4x24x12x2four-layerstructureautomobilemagnesiumalloywheelhubs

low-pressurecastingoptimizationmodelwasestablished,theinputlayerwasthecastingmoldpreheatingtemperature,pouring

temperature,pouringspeedandfillingpressure,andtheoutputlayerwasthegrainsizeandyieldstrength.Thelow-pressure

castingmoldingprocessofautomobilemagnesiumalloywheelhubswasoptimizedthroughtheBPneuralnetwork

optimizationmodel.Theresultsshowthatthegrainsizeoftheoptimizedautomobilemagnesiumalloywheelhubcastingsis

reducedfrom146.8μmto123.2μm,thegrainsizereductionrateis16.0%,theyieldstrengthisincreasedfrom169.7MPato

185.6MPa,andtheyieldstrengthimprovementrateis9.4%.Theoptimalprocessparameters