Fault detection technique is essential for improving overall equipment efficiency of semiconductor manufacturing industry. It has been recognized that fault detection based on k nearest neighbor rule (kNN) can effectively deal with some characteristics of semiconductor processes, such as multimode batch trajectories and nonlinearity. However, the computation complexity and storage space involved in neighbors searching of k NN prevent it from online monitoring, especially for high dimensional cases. To deal with this difficulty, principal component-based k NN has also been presented in literature, in which dimension reduction is done by principal component analysis (PCA) before k NN rule implemented to fault detection. However, dimension reduction by PCA may distort the distances between pairs of samples (trajectories). Thus the false alarm and missing detection of k NN for fault detection may increase in principal component subspace because PCA fails to preserve pairwise distances in subspace. To overcome this drawback, we propose a new fault detection method based on random projection and k NN rule, which combines the advantages of random projection in distance preservation (in the expectation) and k NN rule in dealing with the problems of multimodality and nonlinearity that often coexist in semiconductor manufacturing processes. An industrial example illustrates the performance of the proposed method.

Recently, the well-known k-nearest neighbor (kNN) rule has been successfully applied to the fault detection of industrial processes with nonlinear, multimode, and non-Gaussian distributed data. Once a fault is detected, how to investigate the root causes of the fault by isolating the true faulty variables without any historical fault information is a challenging problem. Inspired by the idea of the contribution analysis (CA) methods developed in the frame of the principal component analysis (PCA), in this paper, a novel isolation index will be provided by decomposing the kNN distance used as the detection index in kNN-based fault detection method. The commonly used CA-based isolation methods suffer from smearing effect due to the correlation among the defined isolation indices, thus prone to misdiagnosis. The proposed isolation index is defined in the original measurement space without correlation. Moreover, theoretical analysis of the isolability for the proposed fault isolation method shows that it can isolate multiple sensor faults under a less strict condition than that used in the analysis of those CA-based fault isolation methods. The numerical examples and Tennessee Eastman (TE) benchmark process are used to illustrate the effectiveness of the proposed method.

Dear Editor, This letter provides a simple framework to generalized zero-shot learning for fault diagnosis. For industrial process monitoring, supervised learning and zero-shot learning (ZSL) can only deal with seen and unseen faults, respectively. However, in the online monitoring stage of the actual industrial process, both seen and unseen faults may occur. This makes supervised learning and zero-shot learning impractical in industrial process monitoring. Generalized zero-shot learning (GZSL) can handle this problem, but its implementation process is too complicated. This letter introduces GZSL into industrial process fault diagnosis, and a simple end-to-end framework is provided to implement GZSL-based fault diagnosis. In this framework, GZSL-based fault diagnosis can be realized by using only a binary classification algorithm. Experimental results show that the proposed framework can accomplish this challenging task of GZSL for fault diagnosis.

The k-nearest neighbor (kNN) method only uses samples’ paired distance to perform fault detection. It can overcome the nonlinearity, multimodality, and non-Gaussianity of process data. However, the nearest neighbors found by kNN on a data set containing a lot of outliers or noises may not be actual or trustworthy neighbors but a kind of pseudo neighbor, which will degrade process monitoring performance. This paper presents a new fault detection scheme using the mutual k-nearest neighbor (MkNN) method to solve this problem. The primary characteristic of our approach is that the calculation of the distance statistics for process monitoring uses MkNN rule instead of kNN. The advantage of the proposed approach is that the influence of outliers in the training data is eliminated, and the fault samples without MkNNs can be directly detected, which improves the performance of fault detection. In addition, the mutual protection phenomenon of outliers is explored. The numerical examples and Tenessee Eastman process illustrate the effectiveness of the proposed method.