模式识别在硫酸镍精制研究中的应用

Application of Pattern Recognition in the Study of Refining Nickel Sulfate

采用模式识别方法对硫酸镍精制实验数据进行处理,通过最优判别平面法将训练样本集投影到二维映照平面,使好、坏两类样本在模式空间中的分布区域可视化,从而可以方便地确定优化方向。结果表明,在所研究的实验条件范围内,提高反应温度、氟化钠过量系数和反应时间均能使实验结果更好。

Pattern recognition was used to analyze the experimental data of refining NiSO4 solution by optimal discrimination plan method（ODPM）, and all the training samples were projected onto a two-dimension mapping plan. Good and bad samples were found to be located separately in different districts of the mapping plan and so the classification results could be demonstrated visibly and easy to idicate the optimal direction. The results suggested that among the ranges of experimental conditions, better results could be got by increasing reaction time, the excess ratio of NaF and reaction temperature.