根据灰色理论 ,采用新陈代谢灰色建模法对铜锍品位的历史数据建立了GM (1 ,1 )模型群 ,并对各维模型进行了精度检验 ,计算表明维数为 4～ 6时模型精度达A级 ,维数继续增大则模型的精度变差。选出精度高的模型对当前加入连续吹炼炉的铜锍的品位进行预测并做均值化处理 ,采用此法对现场 90余班次的数据进行了预测计算 ,与化验值相比 ,预测值的平均绝对误差在± 0 .5%以内。 According to the gray theory, GM (1, 1) model group was established on the historical data of copper matte grade using the metabolic gray modeling method, and the accuracy of each model was tested. The results showed that the accuracy of model reached to 4 ~ 6 A-class, the number of dimensions continues to increase the accuracy of the model worse. Select the model with high precision to predict the grade of copper matte which is added into the continuous converter and do the average value treatment. The method is used to predict the data of more than 90 shifts in the field. Compared with the experimental value, The average absolute error is within ± 0.5%.