摘要
结合BP网络计算机实验研究 ,建立了保护渣化学组成与性能的预测模型 ,并利用化学组成与性能的关系 ,对网络的实用性进行了检验 ,结果符合保护渣化学组成与性能的关系。保护渣粘度随着保护渣碱度的增大而减小 ,而半球点温度随碱度增大而增大 ;保护渣的半球点温度和粘度都随着渣中CaF2 含量的增加而减小。用BP网络的误差反向传播算法建立的保护渣的化学组成与性能的预测模型 ,得出的预测值与实际值的误差小 ,对保护渣的设计与应用都有一定的指导作用。
Combining with the BP nets computer aided experimental study, a comprehensive model for predicting properties and verification of mold powder was established. We also use the relationship between the chemical components and properties of mold powder to verify practicability of this model. The result accorded with the relationship of it. Following the increase of alkalinity, the viscosity of mold powder decreased but the hemisphere point temperature increased. Following the increase of CaF 2, the viscosity and hemisphere point temperature of mold powder all decreased. We use the error back propagation algorithm, a comprehensive model for predicting properties and verification of mold powder has been established. The result of this model indicates that the error between the forecast data and actual data is very little. This model is very useful for designing and using mold powder.
出处
《耐火材料》
CAS
北大核心
2004年第2期115-117,共3页
Refractories
