Dynamic trend analysis is an important technique for fault detection and diagnosis. Trend analysis involves hierarchical representation of signal trends, extraction of the trends, and their comparison (estimation of similarity) to infer the state of the process. In this paper, an overview of some of the existing methods for trend extraction and similarity estimation is presented. A novel interv...

Ž . Principal component analysis PCA is a well-known data dimensionality technique that has been used to detect faults Ž . during the operation of industrial processes. Dynamic principal component analysis DPCA and canonical variate analysis Ž . CVA are data dimensionality techniques which take into account serial correlations, but their effectiveness in detecting faults in industrial processes...

In this paper, a new process monitoring methodology is presented to detect fault occurrence. The proposed methodology incorporates a wavelet de-noising approach based on the fast wavelet transform (FWT) to extract the embodied fault dynamics from the noisy measured data. A level dependent soft thresholding technique using Daubechies 3 with three levels of decomposition is utilized. An appropria...

This study describes a classification methodology based on support vector machines (SVMs), which offer superior classification performance for fault diagnosis in chemical process engineering. The method incorporates an efficient parameter tuning procedure (based on minimization of radiudmargin bound for SVM's leave-one-out errors) into a multi-class classification strategy using a fuzzy decisio...

Handling missing values and large-dimensional features are crucial requirements for data-driven fault diagnosis systems. However, most intelligent diagnostic systems not able to handle data. The presence of high-dimensional feature sets can also further complicate the process diagnosis. This paper aims devise a data imputation unit along with dimensionality reduction in pre-processing module sy...

Abstract The partial least squares (PLS) is a commonly applied multi-variate method in anomaly detection problems. PLS strategy has been amalgamated with $$T^{2}$$ T 2 and squared prediction error (SPE) based statistical indicators to detect anomalies process. These tra...

This paper proposes a fault diagnosis method based on an improved residual network (ResNet) for complex chemical processes. The can automatically and efficiently extract features from the extensive data generated by operation process. improvement is carried out in three aspects. Firstly, 1D convolution introduced construction of model to reduce number parameters training time, shortcut connecti...

The actual multimodal process data usually exhibit non-linear time correlation and non-Gaussian distribution accompanied by new modes. Existing fault diagnosis methods have difficulty adapting to the complex nature of modalities are unable train models based on small samples. Therefore, this paper proposes a modal method meta-learning (ML) neural architecture search (NAS), MetaNAS. Specifically...

This paper presents a novel mutual information (MI) matrix based method for fault detection. Given m-dimensional process, the MI is m×m in which (i,j)-th entry measures values between ith dimension and jth variables. We introduce recently proposed matrix-based Rényi’s α-entropy functional to estimate each of matrix. The new estimator avoids density estimation it operates on eigenspectrum (norma...

Recently, neural networks (NNs) have been proposed for the detection of cyber attacks targeting industrial control systems (ICSs). Such detectors are often retrained, using data collected during system operation, to cope with evolution monitored signals over time. However, by exploiting this mechanism, an attacker can fake provided corrupted sensors at training time and poison learning process ...