Second-Order Component Analysis for Fault Detection
This is an incremental improvement for fault detection in industrial process monitoring, addressing overfitting in neural network-based methods.
The paper tackles fault detection in process monitoring by proposing second-order component analysis (SCA), a method that uses a second-order autoencoder with orthogonal constraints to handle heteroscedastic data and reduce overfitting. Experimental results on the Tennessee-Eastman benchmark show SCA outperforms PCA, KPCA, and autoencoder in missed detection rate and false alarm rate.
Process monitoring based on neural networks is getting more and more attention. Compared with classical neural networks, high-order neural networks have natural advantages in dealing with heteroscedastic data. However, high-order neural networks might bring the risk of overfitting and learning both the key information from original data and noises or anomalies. Orthogonal constraints can greatly reduce correlations between extracted features, thereby reducing the overfitting risk. This paper proposes a novel fault detection method called second-order component analysis (SCA). SCA rules out the heteroscedasticity of pro-cess data by optimizing a second-order autoencoder with orthogonal constraints. In order to deal with this constrained optimization problem, a geometric conjugate gradient algorithm is adopted in this paper, which performs geometric optimization on the combination of Stiefel manifold and Euclidean manifold. Extensive experiments on the Tennessee-Eastman benchmark pro-cess show that SCA outperforms PCA, KPCA, and autoencoder in missed detection rate (MDR) and false alarm rate (FAR).