Traditional onefold data-driven methods for fault detection in complex process industrial systems with high-dimensional, linear, nonlinear, Gaussian, and non-Gaussian coexistence often have less than satisfactory monitoring performance because only a single distribution of variables is considered. To address this problem, hybrid model based on PCA-KPCA-ICA-KICA-BI (Bayesian inference) proposed,...

Causal relations among variables may change significantly due to different control strategies and fault types. Off line-based knowledge is not adequate for diagnosis, existing causal models obtained from data driven methods are mostly based on historical only. However, variable correlation would remain identical, could be very under certain industrial operation conditions. To deal with this pro...

The proliferation of automated data collection schemes and the advances in sensorics are increasing amount we able to monitor real-time. However, given high annotation costs time required by quality inspections, is often available an unlabeled form. This fostering use active learning for development soft sensors predictive models. In production, instead performing random inspections obtain prod...

During the past few decades, a number of methods for selection of input-output pairings for decentralized control have been proposed. Most of these available methods require evaluation of every alternative in order to find the optimal pairings. As the number of alternatives grows rapidly with problem size, pairing selection through an exhaustive search can be computationally forbidding for larg...

Bayesian methods are a kind of data-driven methods developed in recent years and have played an important role in fault detection and diagnosis. Nevertheless, traditional Bayesian fault detection methods cannot deal with the case where some underlying modes are ambiguous in the historical data. To cope with this problem, a new method is presented in this paper. The modes with uncertainty in his...

This work demonstrates for the first time application of network topology variance decompositions in analyzing connectedness chemical plant process variable oscillations arising from disturbances and faults. Specifically, time-based frequency-based variables can be used to compute net pairwise dynamic (NPDC), which originated as a volatility spillover index financial markets studies field econo...

Dynamic Time Warping (DTW) can be used to minimize distance between two sequences which display the same trends but are not perfectly aligned with each other. This feature makes DTW an effective pattern classification method for sequence matching. In Ref. [1], a fault diagnosis method using pattern classification based on adaptive rank-order morphological transform has been proposed. It uses Eu...

Statistics pattern analysis (SPA) is a new multivariate statistical monitoring framework proposed by the authors recently. It addresses some challenges that cannot be readily addressed by the commonly used multivariate statistical methods such as principal component analysis (PCA) in monitoring batch processes in the semiconductor industry. It was later extended to the monitoring of continuous ...

Nonlinearity in industrial processes such as chemical and biological processes is still a significant problem. Kernel principal component analysis (KPCA) has recently proven to be a powerful tool for monitoring nonlinear processes with numerous mutually correlated measured variables. One of the drawbacks of original KPCA is that computation time increases with the number of samples. In this art...

When detecting cyberattacks in Industrial settings, it is not sufficient to determine whether the system suffering a cyberattack. It also fundamental explain why under cyberattack and which are assets affected. In this context, Anomaly Detection based on Machine Learning (ML) Deep (DL) techniques showed great performance when industrial scenarios. However, two main limitations hinder using them...