Improved K-means algorithm for manufacturing process anomaly detection and recognition

被引：0

作者：

Zhou Xiaomin ^{[1
]}

Peng Wei ^{[1
]}

Shi Haibo ^{[1
]}

机构：

[1] Chinese Acad Sci, Shenyang Inst Automat, Shenyang 110016, Peoples R China

来源：

1st International Symposium on Digital Manufacture, Vols 1-3 | 2006年

关键词：

data mining; clustering; quality management; anomaly detection and recognition;

D O I：

暂无

中图分类号：

T [工业技术];

学科分类号：

08 ;

摘要：

Anomaly detection and recognition are of prime importance in process industries. Faults are usually rare, and, therefore, predicting them is difficult . In this paper, a new greedy initialization method for the K-means algorithm is proposed to improve traditional K-means clustering techniques. The new initialization method tries to choose suitable initial points, which are well separated and have the potential to form high-quality clusters. Based on the clustering result of historical disqualification product data in manufacturing process which generated by the Improved-K-means algorithm, a prediction model which is used to detect and recognize the abnormal trend of the quality problems is constructed. This simple and robust alarm-system architecture for predicting incoming faults realizes the transition of quality problems from diagnosis afterward to prevention beforehand indeed. In the end, the alarm model was applied for prediction and avoidance of gear-wheel assembly faults at a gear-plant.

引用

页码：1036 / 1041

页数：6