稀土分离联产品质量控制研究

发布时间：2018-04-11 03:13

本文选题：统计过程控制 + EWMA控制图　；参考：《运筹与管理》2017年03期

【摘要】：以稀土分离企业为背景,抽取联产品特点及质量属性,绘制单一产品的指数加权移动平均控制图和联产品的多元残差T~2控制图,并将两类控制图进行对比分析,分析表明和EWMA控制图相比,联产品多元残差T~2控制图能降低控制图虚发警报的概率。针对多元残差T~2控制图发现的异常模式,采用支持向量机模型对异常模式进行分类处理,寻找分类规则,构造PSO-SVM分类器,运用粒子群算法对SVM参数寻优,并对得出的结果进行对比分析。分析表明该分类器能提高分类正确率,模式识别可以用于诊断稀土企业引起联产品多元残差T~2控制图出现异常的原因,从而提高过程质量管理水平。
[Abstract]:Under the background of rare earth separation enterprise, the characteristics and quality attributes of the combined products are extracted, the exponential weighted moving average control chart of a single product and the multivariate residuals T-2 control chart of the combined product are drawn, and the two control charts are compared and analyzed.The analysis shows that compared with the EWMA control chart, the combined product multivariate residuals control chart can reduce the probability of false alarm in the control chart.Aiming at the abnormal pattern found in the control chart of multivariate residuals, the support vector machine (SVM) model is used to classify the abnormal patterns, to find the classification rules, to construct the PSO-SVM classifier, and to use particle swarm optimization algorithm to optimize the SVM parameters.The results are compared and analyzed.The analysis shows that the classifier can improve the accuracy of classification, and the pattern recognition can be used to diagnose the abnormal causes of the multivariate residuals Tn-2 control chart in rare earth enterprises, thus improving the quality management level of the process.
【作者单位】：内蒙古科技大学经济与管理学院;内蒙古科技大学白云鄂博矿多金属资源综合利用重点实验室;
【基金】：内蒙古自然科学基金资助项目(2014MS0710) 国家自然科学基金资助项目(71162027) 内蒙古哲学社会科学规划资助项目(2014B030)
【分类号】：TQ133.3

【参考文献】