【摘要】 分析了连铸保护渣的化学成分和物理性能,提出了预测连铸保护渣性能的神经网络模型,根据保护渣的化学成分以该模型预测其粘度。结果表明,模型估计与观测值相当吻合。并与多元线性和非线性回归模型作了比较。更多还原

【Abstract】 The relation between physical property and chemical composition of the mold powder was analyzed. An artificial neural network model was developed to predict the physical property of the mold powder, and to estimate the viscosity of the smelted mold flux by its chemical composition . The result indicated that the model prediction was in agreement with the observed value. Also the NN model was compared with the models obtained respectively by multi-linear and nonlinear regression.更多还原