FIGURE 3 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Source publication
This study is concerned with theoretical and practices approach for overall quality-related fault detection and identification in process industries. Fault detection and fault tracing can help engineers to take correct actions and recover the process operations. A novel diagnostic method is proposed based on stacked automatic encoder—canonical corr...
Context in source publication
Context 1
... β 1 is reduced to zero and β 2 is compressed. Compared with the variable selection of ridge in Figure 3, the cylinder in (a) represents the part of β 1 2 + β 2 2 ≤ t . In the twodimensional projection (b), the intersection of curve and l 2 -regular circular is the coefficient solution of ridge regression model. ...
Similar publications
The paper describes an automatic system for controlling and assessing the technological properties of grain processing products during milling by computer vision methods. Optimal technological characteristics of the system are empirically established ensuring its maximum efficiency. The combined action of electrostatic fields (with voltage of 24 kV...
Citations
... The emphasis on quality responsibility in the digital era provides a framework for tracing and analysing all aspects contributing to quality improvements, failures, and waste, including root cause diagnosis (Dong et al., 2019). The concept of traceability significantly contributes to sustainability throughout the supply chain management (Garcia-Torres et al., 2019) while system-based approach such as blockchain technology and comprehensive quality information system throughout the entire production system can also be made (Ali et al., 2022;Casino et al., 2019). ...
... Prior studies have aimed to develop DL models with advanced network architectures for resilient sensor self-validation across various applications [4,15,16]. A summary of previous studies on the development of soft sensor models for the detection, reconstruction, and diagnosis of faulty sensors in WWTP is provided in Table S1 in the supplementary information (SI). ...
Sensor malfunctions in wastewater treatment plants (WWTPs) significantly disrupt process control and energy usage, highlighting the critical need for effective sensor fault diagnosis and reconstruction. This study aims to introduce a novel application of a multi-task learning network in WWTPs to address the challenge of sensor malfunction by enabling simultaneous fault diagnosis and reconstruction. The proposed approach introduces an explainable deep multi-task learning autoencoder network (DMTL-UNet), which effectively allows sharing information among tasks through attention gates and residual connections. The effectiveness of the DMTL-UNet model is validated using a real-world dataset from a WWTP in South Korea. The results demonstrate the remarkable capability of the DMTL-UNet model in accurately diagnosing multiple faults (F1-score = 99.08 %) and achieving superior reconstruction performance (RMSE = 31.1175 mg/L) for faulty WWTP sensors. Moreover , implementing the DMTL-UNet model offers significant energy savings of 37.44 %, corresponding to a reduction in the aeriation cost by 1154.91 USD. Therefore, implementing the DMTL-UNet model for calibrating faulty sensors can enhance sensor reliability, improve maintenance practices, and contribute to the sustainable operation of WWTPs.
... A few existing deep architectures for FDD include convolutional neural networks (CNNs) [9], [10], deep belief networks (DBNs) [11], long short-term memory (LSTM) [12], stacked auto-encoders (SAE) [13], [14], variational auto-encoders (VAE) [15], and generative adversarial networks (GANs) [16]. By contrast, a recent study combined the advantages of two deep learning models to achieve a better detection performance using an adversarial auto-encoder (AAE) [17]. ...
Process monitoring is important for ensuring operational reliability and preventing occupational accidents. In recent years, data-driven methods such as machine learning and deep learning have been preferred for fault detection and diagnosis. In particular, unsupervised learning algorithms, such as auto-encoders, exhibit good detection performance, even for unlabeled data from complex processes. However, decisions generated from deep-neural-network-based models are difficult to interpret and cannot provide explanatory insight to users. We address this issue by proposing a new fault diagnosis method using explainable artificial intelligence to break the traditional trade-off between the accuracy and interpretability of deep learning model. First, an adversarial auto-encoder model for fault detection is built and then interpreted through the integration of Shapley additive explanations (SHAP) with a combined monitoring index. Using SHAP values, a diagnosis is conducted by allocating credit for detected faults, deviations from a normal state, among its input variables. The proposed diagnosis method can consider not only reconstruction space but also latent space unlike conventional method, which evaluate only reconstruction error. The proposed method was applied to two chemical process systems and compared with conventional diagnosis methods. The results highlight that the proposed method achieves the exact fault diagnosis for single and multiple faults and, also, distinguishes the global pattern of various fault types.
... In addition, some studies utilized the variable selection method to realize fault variable identification. Dong et al. [150] utilized LASSO to identify the fault relevant variable. Yu et al. [60] proposed an elastic network-based method to improve the identification performance. ...
Process monitoring technologies play a key role in maintaining the steady state of industrial processes. However, with the increasing complexity of modern industrial processes, traditional monitoring methods cannot provide satisfactory performance. In the past decades, deep learning models have achieved rapid development in industrial data analysis, especially autoencoder (AE), which has been widely used to deal with various challenges of process monitoring, and a number of related works have been proposed. This paper aims to present a comprehensive review of AE-based industrial applications, which mainly includes two parts: AE-based representation learning and monitoring strategies, which illustrate the entire design process of AE-based monitoring methods. In particular, AE, AE variants, and the encoder-decoder framework are briefly introduced first. Secondly, AE-based representation learning is comprehensively reviewed from the aspects of industrial data characteristics. Then, the state-of-the-art studies of monitoring strategies, including fault detection strategies and fault diagnosis strategies, are reviewed and discussed. Finally, some prospects for future research are explored.
... LASSO is widely used in the literature and has proved its effectiveness in different applications on real industrial data. Several works proposed LASSO-based approaches for different purposes as: feature selection [12], prediction [13], [14], [15] and process monitoring [16], [17]. In this paper, it is adopted for health monitoring of batch processes. ...
Over the last few years, with the increasing worldwide competition, semiconductor industries have had to constantly innovate in order to enhance their performance, productivity and minimize the downtime. Monitoring the state of health of their equipment units is important to avoid machine failures and to plan maintenance actions.
For that, a novel approach for health indicator extraction named Significant Points combined to the Least Absolute Shrinkage and Selection Operator (SP-LASSO) is proposed in this paper. It deals with the problem of high dimensional data and the specificity of the health indicator in real industrial cases.
The proposed method performs feature selection and health indicator extraction and it is mainly based on LASSO. A numerical application on simulated data illustrates the accuracy of this approach.
... Manufacturing is highly competitive (Choudhary et al. 2009) and managing manufacturing operations can be very complex. This complexity is increasing (Dong et al. 2019) as new measures are adopted that lead to a data-intensive environment (Abdelrahman and Keikhosrokiani 2020;Sinha et al. 2021). Manufacturing companies should solve their operational problems efficiently and permanently in order to remain competitive. ...
... The second level focuses on which physical occurrences are the root cause (e.g., extraordinary increases in pressure, high voltages). Examples of studies that focus on this kind of data are Dong et al. (2019) and Saez et al. (2019). The third level focuses on the human and organizational characteristics of the root cause (e.g., equipment maintenance) to identify what triggered the physical occurrences. ...
Overlap has been identified in previous works as a significant obstacle to automated diagnosis using data mining algorithms, since it makes it impossible to discern how each machine influences product quality. Several solutions that handle overlap have been proposed, but the final result is a list of potential overlapped root causes. The goal of this paper is to develop a solution resilient to overlap that can determine the true root cause from a list of possible root causes, when possible, and determine the conditions in which it is possible to identify the root causes. This allows for a better understanding of overlap, and enables the development of a fully automatic root cause analysis for manufacturing. To do so, we propose an automatic root cause analysis approach that uses causal inference and do calculus to determine the true root cause. The proposed approach was validated on simulated and real case-study data, and allowed for an estimation of the effect of a product passing through a certain machine while disregarding the effect of overlap, in certain conditions. The results were on par with the state-of-the-art solutions capable of handling overlap. The contributions of this paper are a graphical definition of overlap, the identification of the conditions in which is possible to overcome the effect of overlap, and a solution that can present a single true root cause when such conditions are met.
... For quality-relevant tasks, Yuan et al. [17] applied the variable-wise weighted stack autoencoder (SAE) in the modeling of refinery process. Dong et al. [18] proposed a method based on stacked automatic encoder-canonical correlation analysis and least absolute shrinkage selection operator for fault diagnosis. However, the neural network (NN)-based methods model the quality variables by minimizing the predictive errors of samples that are sensitive to extreme values and local information. ...
... One includes IDV (1), (5), and (7), which indicate the faults that can recover due to the close-loop control strategies. The other type involves IDV (2), (6), (8), (12), (13), (18), and (21), which cannot recover after the faults occur. Figure 4 presents the normalized samples of IDV (1), (2), and (4). ...
Quality-relevant fault detection aims to reveal whether quality variables are affected when a fault is detected. For current industrial processes, kernel-based methods focus on the nonlinearity within process variables, which is insufficient for obtaining nonlinearities of quality variables. Alternatively, neural network is an option for nonlinear prediction. However, these models are driven by predictive errors on samples. For quality-relevant tasks, the key is to capture the trends of quality variables. Therefore, this study proposes a new model, namely, maximizing correlation neural network (MCNN), to predict the quality-relevant information intuitively. The MCNN is trained to maximize the linear correlation between quality variables and the combinations of nonlinear representations mapped by a multilayer feedforward network. As such, fault detection can be implemented in the quality-relevant and irrelevant subspaces on the basis of the deep most correlated representations of process variables. Considering that different variables have different sensitivities to quality at various locations due to their nonlinear relationship, fault backpropagation is designed in the MCNN to isolate the faulty variables on the basis of real-time faulty information. Finally, numerical example and Tennessee Eastman process are used to evaluate the proposed method, which exhibits a competitive performance.
For decades, manufacturers have been collecting and storing high amounts of data with the aim of better controlling and managing their processes. With the vast amount of information and hidden knowledge in all of these data, the challenge for these manufacturers to monitor their equipment units, is the extraction of an appropriate health indicator from these data that illustrates the actual state of their equipment units. In this paper, we are interested in extracting the health indicator of semiconductor equipment where manufacturing is performed by batch. For that, a novel automatic approach named Significant Points combined to the Least Absolute Shrinkage and Selection Operator (SP-LASSO) is proposed. This approach is mainly based on LASSO regression model. Its accuracy is illustrated by numerical application on simulated data.
The industrial process monitoring and operating performance assessment techniques are of great significance to ensure the safety and efficiency of the production and to improve the comprehensive economic benefits for the modern enterprises. In this paper, a new key performance indicator (KPI) oriented nonlinear process monitoring and operating performance assessment method is proposed based on the improved Hessian locally linear embedding (HLLE), in view of the problems of strong nonlinearity, high dimension and information redundancy in actual industrial process data. Firstly, in order to characterise the similarities of samples in both temporal and spatial dimensions, a new measurement, based on Finite Markov theory, is defined to replace the Euclidean distance in traditional HLLE. Secondly, by mining the relationships between process variables and the key performance indicator, the KPI oriented feature extraction method is developed. On this basis, the monitoring statistics is constructed and the corresponding control limit is determined for the real-time fault detection. After that, a new operating performance assessment approach based on sliding window Kullback–Leibler divergence is put forward to facilitate maintenance or adjustments. Finally, the proposed method is applied to the hot strip mill process, and the results show the effectiveness.
