Fault detection with moving window PCA using NIRS spectra for monitoring the anaerobic digestion process

Principal component analysis (PCA) is a popular method for process monitoring. However, most processes are time-varying, thus older samples are not representative of the current process status. This led to the introduction of adaptive-PCA based monitoring, such as moving window PCA (MWPCA). In this study, near-infrared spectroscopy (NIRS) responses to digester failures were evaluated to develop a spectral data processing tool. Tests were performed with a spectroscopic probe (350–2,500 nm), using a 35 L mesophilic continuously stirred tank reactor. Co-digestion experiments were performed with pig slurry mixed with several co-substrates. Different stresses were induced by abruptly increasing the organic load rate, changing the feedstock or stopping the stirring. Physicochemical parameters as well as NIRS spectra were acquired for lipid, organic and protein overloads experiments. MWPCA was then applied to the collected spectra for a multivariate statistical process control. MWPCA outputs, Hotelling T2 and residuals Q statistics showed that most of the induced dysfunctions can be detected with variations in these statistics according to a defined criterion based on spectroscopic principles and the process. MWPCA appears to be a multivariate statistical method that could help in decision support in industrial biogas plants.

Adaptive Principal Component Analysis Combined with Feature Extraction-Based Method for Feature Identification in Manufacturing

This paper developed a principal component analysis (PCA)-integrated algorithm for feature identification in manufacturing; this algorithm is based on an adaptive PCA-based scheme for identifying image features in vision-based inspection. PCA is a commonly used statistical method for pattern recognition tasks, but an effective PCA-based approach for identifying suitable image features in manufacturing has yet to be developed. Unsuitable image features tend to yield poor results when used in conventional visual inspections. Furthermore, research has revealed that the use of unsuitable or redundant features might influence the performance of object detection. To address these problems, the adaptive PCA-based algorithm developed in this study entails the identification of suitable image features using a support vector machine (SVM) model for inspecting of various object images; this approach can be used for solving the inherent problem of detection that occurs when the extraction contains challenging image features in manufacturing processes. The results of experiments indicated that the proposed algorithm can successfully be used to adaptively select appropriate image features. The algorithm combines image feature extraction and PCA/SVM classification to detect patterns in manufacturing. The algorithm was determined to achieve high-performance detection and to outperform the existing methods.