分析了连铸板坯裂纹的成因,并根据连铸生产的特点和工艺条件,结合国内外一些先进钢铁厂板坯裂纹的防止对策,引入BP神经元网络,采用神经网络与专家系统相结合的方法,运用热电偶温度数据,结合结晶器液面、保护渣、钢号以及钢液中C、S、P含量等数据,实现板坯表面裂纹预报。 Based on the characteristics of the continuous casting and the technological conditions, the paper also introduced the BP neural network by combining with the prevention measures of the slab cracks in some advanced steel mills at home and abroad. The neural network was combined with the expert system Method, using thermocouple temperature data, combined with the mold surface, mold flux, steel and C, S, P content in the molten steel and other data to achieve slab surface crack prediction.