Graphical approach of equipment health monitoring using network analysis

Abstract

As the complexity of semiconductor devices increases, new approaches are required for monitoring equipment health. In this paper, we adopt the weighted gene co-expression network analysis (WGCNA) for sensor data following in three stages: 1) construction of networks, 2) identification of modules, and 3) detection of abnormal sensors by the graphical approach. Therefore, manufacturing process conditions are able to modify, the similarity between sensors was increased, and in return 0.3% yield rate gain is achieved.