A sequential resampling approach for imbalanced batch process fault detection in semiconductor manufacturing
- PDF / 2,611,124 Bytes
- 16 Pages / 595.276 x 790.866 pts Page_size
- 92 Downloads / 197 Views
A sequential resampling approach for imbalanced batch process fault detection in semiconductor manufacturing Yi Zhang1,2 · Peng Peng1 · Chongdang Liu1 · Yanyan Xu1,3 · Heming Zhang1 Received: 25 March 2020 / Accepted: 12 November 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Fault detection is one of the most important research topics to guarantee safe operation and product quality consistency especially in the batch process of semiconductor manufacturing. However, the imbalanced fault data bring great challenges to extract the high nonlinearity and inherently time-varying dynamics of the batch process. Motivated by these, we propose a sequential oversampling discrimination approach for imbalanced batch process fault detection. Especially, different from the traditional oversampling methods, which extract temporal features from the whole process, we transform a whole batch sequence into multiple fixed-length sequences each batch by a sliding window, to extract the robust time-varying dynamics features. Then, an oversampling neural network is performed to balance both sequences of minority and majority classes. The needed sequences of the minority class are generated by an improved combination model of variational auto-encoder and generative adversarial network. Finally, a simplified sequential neural network is learned by the balanced-class sequences to perform the discrimination. We conduct extensive experiments based on two datasets of semiconductor manufacturing. One is a benchmark dataset and the other is a dataset from a real production line. The results achieved significant improvement, compared with other state-of-art fault detection methods and oversampling techniques. Keywords Fault detection · Imbalanced classification · Variational auto-encoders (VAEs) · Batch process monitoring · Semiconductor manufacturing
Introduction The modern high-tech industry is characterized by flexible production that meets the needs of offering a variety of product types, a high level of product quality, and short lead times for customers. Consequently, batch production has gradu-
B
Heming Zhang [email protected] Yi Zhang [email protected] Peng Peng [email protected] Chongdang Liu [email protected] Yanyan Xu [email protected]
1
Department of Automation, Tsinghua University, Beijing 100091, China
2
Naval Research Academy, Beijing 100073, China
3
Unit 94926, Wuxi 214141, China
ally developed into a major production mode and made its irreplaceable market position in many industrial applications such as semiconductor, pharmaceutical, chemical, biology and so on. The study of batch process monitoring has a practical significance to ensure quality consistency and operational safety (Zhu and Gao 2018). Fault detection (FD) is the first main step of process monitoring, which aims to identify whether the abnormal status has occurred (Said et al. 2020). With the rapid development of storage devices and sensors, a large number of process measurements are collec
Data Loading...
