A model of predicting end phosphorus in converter steelmaking was established based on the PLS-BP dimensionality reduction net using SPR and ANN to avoid the defaults of PLS and BP individually.

• 摘要用模式识别和神经网络建立PLS-BP的转炉终点磷含量降维网络模型，克服了PLS法和BP法各自的缺陷。