【摘要】 采用BP神经网络模型建立了铝合金的晶粒尺寸与细化剂的成分、含量、温度的关系，预测铝合金的晶粒尺寸，并对该模型进行了交叉验证。结果表明：神经网络成功建立了细化剂成分、添加量、保温时间与晶粒平均尺寸之间的预测模型，并且模型的均方根误差E_RMS为2.83,决定系数R²为0.99,表明该模型的拟合较好，对调节铝合金细化处理工艺参数有一定的指导意义。更多还原

【Abstract】 A BP neural network model was employed to correlate grain size of aluminum alloys with refiner composition, content and temperature. This model was used to predict the grain size of aluminum and its alloys and was validated through cross-validation. The results indicate that the neural network has successfully constructed a predictive model between the refiner composition, refiner addition amount, holding time, and the average grain size. The model has achieved an RMAE of 2.83 and an R~2 of 0.99, demonstrating good fitting performance and providing certain guidance significance for the process parameters of using refiners in aluminum alloys.更多还原