Distributed process monitoring gains popularity recently to perform system health management for large-scale industrial processes and support the decision-making for system maintenance. However, process monitoring for complex large -scale systems using distributed approaches is often challenging due to significant nexus among variables. Therefore, this article proposed a novel distributed process monitoring method to achieve efficient monitoring with a reasonable and interpretable division scheme which is only given by the spatial distribution of each variable and the results of Granger causality analysis. At each subblock, a local canonical variate analysis model with Wasserstein-distance-based indices can be built to monitor each local system. With the help of a Bayesian inference strategy, all the local monitoring results are fused into a global one. Then, from both block-level and variable-level, the proposed hierarchical fault isolation method can sort out candidates for the rooting causality analysis of the detected fault, respectively. Depending on the causal analysis, the rooting cause can be identified from the intersection of two candidate sets, thereby virtualizing the propagation path of a fault. Lastly, the presented methodology of distributed process monitoring is verified by a numeral case study and the Tennessee Eastman (TE) benchmarking platform, respectively. The conclusions show that the presented methodology can perform more accurately and efficiently than traditional approaches. In particular, the proposed method can detect simulated faults in a mathematical case and the fault 15 in the TE process with nearly 100 % and 94.72 %, respectively, in terms of fault detection rates, which is barely achieved by reported methods
{"title":"Distributed process monitoring of the large-scale system using spatio-temporal-causality and Wasserstein-distance-based canonical variate analysis","authors":"Chong Xu , Daoping Huang , Guangping Yu , Yiqi Liu","doi":"10.1016/j.jprocont.2024.103367","DOIUrl":"10.1016/j.jprocont.2024.103367","url":null,"abstract":"<div><div>Distributed process monitoring gains popularity recently to perform system health management for large-scale industrial processes and support the decision-making for system maintenance. However, process monitoring for complex large -scale systems using distributed approaches is often challenging due to significant nexus among variables. Therefore, this article proposed a novel distributed process monitoring method to achieve efficient monitoring with a reasonable and interpretable division scheme which is only given by the spatial distribution of each variable and the results of Granger causality analysis. At each subblock, a local canonical variate analysis model with Wasserstein-distance-based indices can be built to monitor each local system. With the help of a Bayesian inference strategy, all the local monitoring results are fused into a global one. Then, from both block-level and variable-level, the proposed hierarchical fault isolation method can sort out candidates for the rooting causality analysis of the detected fault, respectively. Depending on the causal analysis, the rooting cause can be identified from the intersection of two candidate sets, thereby virtualizing the propagation path of a fault. Lastly, the presented methodology of distributed process monitoring is verified by a numeral case study and the Tennessee Eastman (TE) benchmarking platform, respectively. The conclusions show that the presented methodology can perform more accurately and efficiently than traditional approaches. In particular, the proposed method can detect simulated faults in a mathematical case and the fault 15 in the TE process with nearly 100 % and 94.72 %, respectively, in terms of fault detection rates, which is barely achieved by reported methods</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"146 ","pages":"Article 103367"},"PeriodicalIF":3.3,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143174925","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-01DOI: 10.1016/j.jprocont.2024.103372
Yan Chen, Xiaoyu Zhang, Dazi Li, Jinglin Zhou
Nonlinear, non-Gaussian, and dynamic features pose a great challenge for complex fault detection and fault diagnosis (FDD). Focusing on fault detection, independent component analysis (ICA) and adversarial autoencoder (AAE) are fused to form a new method for nonlinear non-Gaussian latent variable extraction: ICA–AAE. In addition, a strategy for establishing more accurate fault detection thresholds using tail distribution features is presented. Furthermore, a new class of fault diagnosis frameworks to fully exploit the information obtained from normal samples is developed. Fault data are first re-represented using the established ICA–AAE model. Then, the low-dimensional spatial distribution features with their inherited high-dimensional temporal dependencies are synthesized into image information using an image-based approach, and a spatio-temporal fusion fault diagnosis method is implemented using a convolutional neural network (CNN). Tennessee Eastman (TE) process results show that the proposed methods can achieve more accurate fault detection and diagnosis.
{"title":"A new class of fault detection and diagnosis methods by fusion of spatially distributed and time-dependent features","authors":"Yan Chen, Xiaoyu Zhang, Dazi Li, Jinglin Zhou","doi":"10.1016/j.jprocont.2024.103372","DOIUrl":"10.1016/j.jprocont.2024.103372","url":null,"abstract":"<div><div>Nonlinear, non-Gaussian, and dynamic features pose a great challenge for complex fault detection and fault diagnosis (FDD). Focusing on fault detection, independent component analysis (ICA) and adversarial autoencoder (AAE) are fused to form a new method for nonlinear non-Gaussian latent variable extraction: ICA–AAE. In addition, a strategy for establishing more accurate fault detection thresholds using tail distribution features is presented. Furthermore, a new class of fault diagnosis frameworks to fully exploit the information obtained from normal samples is developed. Fault data are first re-represented using the established ICA–AAE model. Then, the low-dimensional spatial distribution features with their inherited high-dimensional temporal dependencies are synthesized into image information using an image-based approach, and a spatio-temporal fusion fault diagnosis method is implemented using a convolutional neural network (CNN). Tennessee Eastman (TE) process results show that the proposed methods can achieve more accurate fault detection and diagnosis.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"146 ","pages":"Article 103372"},"PeriodicalIF":3.3,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143174923","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-01DOI: 10.1016/j.jprocont.2024.103370
Wenli Chen , Xiaojian Li
This paper investigates the adaptive data-driven design issue for fault-tolerant control systems with unknown dynamics. Initially, the fault-tolerant control problem is transformed into a stabilization problem for switched systems, where both the switching signal and system dynamics are unknown due to the uncertainties in fault occurrence instants and faulty modes. While extensive research has been conducted on switched systems, the strategies for addressing unknown switching signals remain comparatively scarce, especially when system dynamics are also unknown. To tackle this issue, a Lyapunov function-based monitoring scheme is provided to determine the time instants of switching in system dynamics during operation. Subsequently, a data-driven adaptive learning control mechanism is introduced to update feedback gains. Considering the asynchronous issue between the switching of the controller and system dynamics due to the learning process, sufficient conditions concerning the switching frequency of the system dynamics are provided. Thereby, a data-driven adaptive learning fault-tolerant control algorithm is proposed. Under the frequency constraint on the switching of system dynamics, it is shown that the offered control scheme maintains the closed-loop system’s stability. Finally, two simulation examples are provided to show the effectiveness of the proposed approach.
{"title":"Adaptive data-driven design of fault-tolerant control systems with unknown dynamics","authors":"Wenli Chen , Xiaojian Li","doi":"10.1016/j.jprocont.2024.103370","DOIUrl":"10.1016/j.jprocont.2024.103370","url":null,"abstract":"<div><div>This paper investigates the adaptive data-driven design issue for fault-tolerant control systems with unknown dynamics. Initially, the fault-tolerant control problem is transformed into a stabilization problem for switched systems, where both the switching signal and system dynamics are unknown due to the uncertainties in fault occurrence instants and faulty modes. While extensive research has been conducted on switched systems, the strategies for addressing unknown switching signals remain comparatively scarce, especially when system dynamics are also unknown. To tackle this issue, a Lyapunov function-based monitoring scheme is provided to determine the time instants of switching in system dynamics during operation. Subsequently, a data-driven adaptive learning control mechanism is introduced to update feedback gains. Considering the asynchronous issue between the switching of the controller and system dynamics due to the learning process, sufficient conditions concerning the switching frequency of the system dynamics are provided. Thereby, a data-driven adaptive learning fault-tolerant control algorithm is proposed. Under the frequency constraint on the switching of system dynamics, it is shown that the offered control scheme maintains the closed-loop system’s stability. Finally, two simulation examples are provided to show the effectiveness of the proposed approach.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"146 ","pages":"Article 103370"},"PeriodicalIF":3.3,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143174924","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-01DOI: 10.1016/j.jprocont.2024.103357
Alireza Arastou , Ye Wang , Erik Weyer
Control of large-scale networks can be challenging due to difficulties in implementation of high-order control systems, data collection, and actuation. Distributed and decentralized control systems are therefore commonly used. This paper proposes an optimization-based partitioning approach for use in decentralized and distributed control. It factors in both computational and communication costs, while also taking into consideration the controllability of the subsystems. An efficient algorithm for solving the optimization problem is also provided. The proposed approach is demonstrated on case studies from water distribution systems.
{"title":"Optimization-based network partitioning for distributed and decentralized control","authors":"Alireza Arastou , Ye Wang , Erik Weyer","doi":"10.1016/j.jprocont.2024.103357","DOIUrl":"10.1016/j.jprocont.2024.103357","url":null,"abstract":"<div><div>Control of large-scale networks can be challenging due to difficulties in implementation of high-order control systems, data collection, and actuation. Distributed and decentralized control systems are therefore commonly used. This paper proposes an optimization-based partitioning approach for use in decentralized and distributed control. It factors in both computational and communication costs, while also taking into consideration the controllability of the subsystems. An efficient algorithm for solving the optimization problem is also provided. The proposed approach is demonstrated on case studies from water distribution systems.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"146 ","pages":"Article 103357"},"PeriodicalIF":3.3,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143174883","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-01DOI: 10.1016/j.jprocont.2024.103371
Zhangming Lan , Yun Wang , Yuchen He , Lijuan Qian
Incipient fault detection has been considered as one of the most efficient approaches to reduce the risks of systematic failures. However, incipient fault signals are often obscured by nonstationary characteristics, such as trend features and periodic features. In this paper, a hybrid-supervised trend-period variational autoencoder (HSTPVAE) is proposed to achieve fault detection for incipient faults in nonstationary processes. The features of trend, period and residual are extracted from a novel trend-period variational autoencoder (TPVAE). Then, these features are optimized by a hybrid supervised strategy, which includes fault trend semi-supervised module and trend-period self-supervised module. The former enhances the distinctiveness between normal and fault trend features, while the latter prevents the overfitting issues. Finally, the effectiveness of the HSTPVAE is demonstrated on a numerical simulation process and real boiler combustion process of thermal power generation. The comparison with state-of-the-art (SOTA) methods proves that the proposed HSTPVAE method can fully utilize the trend and period features of nonstationary process and outperform other comparison methods in incipient fault detection.
{"title":"Nonstationary incipient fault detection based on hybrid supervised trend-period variational autoencoder and its application in thermal power generation","authors":"Zhangming Lan , Yun Wang , Yuchen He , Lijuan Qian","doi":"10.1016/j.jprocont.2024.103371","DOIUrl":"10.1016/j.jprocont.2024.103371","url":null,"abstract":"<div><div>Incipient fault detection has been considered as one of the most efficient approaches to reduce the risks of systematic failures. However, incipient fault signals are often obscured by nonstationary characteristics, such as trend features and periodic features. In this paper, a hybrid-supervised trend-period variational autoencoder (HSTPVAE) is proposed to achieve fault detection for incipient faults in nonstationary processes. The features of trend, period and residual are extracted from a novel trend-period variational autoencoder (TPVAE). Then, these features are optimized by a hybrid supervised strategy, which includes fault trend semi-supervised module and trend-period self-supervised module. The former enhances the distinctiveness between normal and fault trend features, while the latter prevents the overfitting issues. Finally, the effectiveness of the HSTPVAE is demonstrated on a numerical simulation process and real boiler combustion process of thermal power generation. The comparison with state-of-the-art (SOTA) methods proves that the proposed HSTPVAE method can fully utilize the trend and period features of nonstationary process and outperform other comparison methods in incipient fault detection.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"146 ","pages":"Article 103371"},"PeriodicalIF":3.3,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143174877","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-28DOI: 10.1016/j.jprocont.2025.103376
Nasreldin Ibrahim , Na Dong , Modawy Adam Ali Abdalla
This study investigates a dual input and dual output Model-Free Adaptive Iterative Learning Control (MFAILC)-based energy-saving control of the refrigeration system to maintain a minimum stable superheat and a constant evaporation temperature. Traditional PID control for superheat control is often unstable due to the complex and high-order nature of the refrigeration systems. Additionally, the presence of nonlinearities and time variations complicates the design of smart controllers. To get around these problems, an advanced control technique MFAILC algorithm was first designed for single input and single output. Subsequently, the proposed MFAILC algorithm was extended to dual-input and dual-output energy-saving control of refrigeration systems. To test the performance of this innovative methodology, a qualitative and quantitative comparisons, as well as a statistical ANOVA test, have been conducted between the proposed method and the Model-Free Adaptive Control (MFAC) algorithm to evaluate the performance. Step signals have been utilized as the given signals for comprehensive performance testing. As a result, the proposed approach demonstrates higher tracking stability and fast response speed, with an average tracking accuracy of 98.10% for superheat and 91.72% for evaporator temperature, among the simulation experiments.
{"title":"On dual-loop model-free adaptive iterative learning control and its application","authors":"Nasreldin Ibrahim , Na Dong , Modawy Adam Ali Abdalla","doi":"10.1016/j.jprocont.2025.103376","DOIUrl":"10.1016/j.jprocont.2025.103376","url":null,"abstract":"<div><div>This study investigates a dual input and dual output Model-Free Adaptive Iterative Learning Control (MFAILC)-based energy-saving control of the refrigeration system to maintain a minimum stable superheat and a constant evaporation temperature. Traditional PID control for superheat control is often unstable due to the complex and high-order nature of the refrigeration systems. Additionally, the presence of nonlinearities and time variations complicates the design of smart controllers. To get around these problems, an advanced control technique MFAILC algorithm was first designed for single input and single output. Subsequently, the proposed MFAILC algorithm was extended to dual-input and dual-output energy-saving control of refrigeration systems. To test the performance of this innovative methodology, a qualitative and quantitative comparisons, as well as a statistical ANOVA test, have been conducted between the proposed method and the Model-Free Adaptive Control (MFAC) algorithm to evaluate the performance. Step signals have been utilized as the given signals for comprehensive performance testing. As a result, the proposed approach demonstrates higher tracking stability and fast response speed, with an average tracking accuracy of 98.10% for superheat and 91.72% for evaporator temperature, among the simulation experiments.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"147 ","pages":"Article 103376"},"PeriodicalIF":3.3,"publicationDate":"2025-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143182259","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-28DOI: 10.1016/j.jprocont.2025.103375
Ali Moradvandi , Sjoerd Heegstra , Pamela Ceron-Chafla , Bart De Schutter , Edo Abraham , Ralph E.F. Lindeboom
Temperature plays a critical role in performance and stability of anaerobic digestion processes, subject to frequent meteorological fluctuations. However, state-of-the-art modeling and process control approaches for anaerobic digestion often do not consider the temporal dynamics of the temperature, which can influence microbial communities, kinetics, and chemical equilibrium, and consequently, biogas production efficiency. Therefore, to account for anaerobic digesters operating under fluctuating meteorological conditions, the Anaerobic Digestion Model no. 1 (ADM1) is mechanistically extended in this paper to incorporate temporal changes into temperature-dependent parameters by defining inhibition functions for microbial activities using the cardinal temperature model, and accounting for the lag in microbial adaptation to temperature fluctuations using a time-lag adaptation function. Thereafter, given that temperature fluctuations are a significant disturbance, a control framework based on Model Predictive Control (MPC) is developed to regulate the feeding flow rate and to ensure stable production rates despite temperature disturbances without relying on direct temperature control. An adaptive MPC approach is formulated based on a linear input–output model, where the parameters of the linear model are updated online to capture the nonlinear dynamics of the process and frequent changes in the dynamics accurately. In addition, a fuzzy logic system is employed to assign a reference trajectory for the production rate based on the temperature and its rate of change. Integrating this fuzzy logic system with the MPC controller enhances the production rate on warm days and avoids the operational failure in production on cold days. Additionally, to enhance biogas production rates, the feasibility of utilizing a portion of the produced biogas for external heating purposes is also investigated. It is demonstrated that by utilizing the proposed MPC approach, the additional amount of feed for the digester to produce methane required for a self-consumption biogas-fueled heating system can be calculated according to the meteorological variations. This enhances the process performance and stability. Finally, a thermally optimized dome digester semi-buried in the ground, operating under climate conditions of The Netherlands is considered as a case study to validate the extended model in agreement with biological and physicochemical behaviors of real-world applications, and to demonstrate the effectiveness of the proposed control system in handling temperature changes and enhancing performance.
{"title":"Model predictive control of feed rate for stabilizing and enhancing biogas production in anaerobic digestion under meteorological fluctuations","authors":"Ali Moradvandi , Sjoerd Heegstra , Pamela Ceron-Chafla , Bart De Schutter , Edo Abraham , Ralph E.F. Lindeboom","doi":"10.1016/j.jprocont.2025.103375","DOIUrl":"10.1016/j.jprocont.2025.103375","url":null,"abstract":"<div><div>Temperature plays a critical role in performance and stability of anaerobic digestion processes, subject to frequent meteorological fluctuations. However, state-of-the-art modeling and process control approaches for anaerobic digestion often do not consider the temporal dynamics of the temperature, which can influence microbial communities, kinetics, and chemical equilibrium, and consequently, biogas production efficiency. Therefore, to account for anaerobic digesters operating under fluctuating meteorological conditions, the Anaerobic Digestion Model no. 1 (ADM1) is mechanistically extended in this paper to incorporate temporal changes into temperature-dependent parameters by defining inhibition functions for microbial activities using the cardinal temperature model, and accounting for the lag in microbial adaptation to temperature fluctuations using a time-lag adaptation function. Thereafter, given that temperature fluctuations are a significant disturbance, a control framework based on Model Predictive Control (MPC) is developed to regulate the feeding flow rate and to ensure stable production rates despite temperature disturbances without relying on direct temperature control. An adaptive MPC approach is formulated based on a linear input–output model, where the parameters of the linear model are updated online to capture the nonlinear dynamics of the process and frequent changes in the dynamics accurately. In addition, a fuzzy logic system is employed to assign a reference trajectory for the production rate based on the temperature and its rate of change. Integrating this fuzzy logic system with the MPC controller enhances the production rate on warm days and avoids the operational failure in production on cold days. Additionally, to enhance biogas production rates, the feasibility of utilizing a portion of the produced biogas for external heating purposes is also investigated. It is demonstrated that by utilizing the proposed MPC approach, the additional amount of feed for the digester to produce methane required for a self-consumption biogas-fueled heating system can be calculated according to the meteorological variations. This enhances the process performance and stability. Finally, a thermally optimized dome digester semi-buried in the ground, operating under climate conditions of The Netherlands is considered as a case study to validate the extended model in agreement with biological and physicochemical behaviors of real-world applications, and to demonstrate the effectiveness of the proposed control system in handling temperature changes and enhancing performance.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"147 ","pages":"Article 103375"},"PeriodicalIF":3.3,"publicationDate":"2025-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143182258","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-27DOI: 10.1016/j.jprocont.2025.103377
Shumei Zhang , Weifeng Mao , Feng Dong , Sijia Wang
In industrial processes, sensor aging and harsh field environments often introduce uncertainties into process data. These uncertainties obscure fault symptoms and undermine the precision of fault diagnosis. To address this challenge, this paper proposes a robust fault diagnosis model with interval distribution analysis for abnormal recognition under data uncertainties in complex industrial settings. Specifically, this research first transforms uncertain data collected from complex industrial sites into interval-valued data, which can globally capture the internal structural characteristics of data objects and effectively represent the uncertainty inherent in the single-valued data. Subsequently, a complete information principal component analysis (CIPCA)-based dimensionality reduction model is constructed to exploit the distribution information within the interval and extract interval fault features. Finally, an interval radial basis function neural network (IRBFNN) is developed to handle the interval upper and lower bound matrices through subtractive clustering algorithm, facilitating fault prediction and diagnosis in industrial processes contaminated by uncertainties. The key to discriminate the proposed method from many well-established fault diagnosis methods is its ability to cluster the interval fault features from uncertain data with embedded interval distribution analysis. The superiority of the proposed fault diagnosis model is validated by the Tennessee Eastman process (TEP).
{"title":"A robust fault diagnosis model with interval distribution analysis for industrial processes with data uncertainties","authors":"Shumei Zhang , Weifeng Mao , Feng Dong , Sijia Wang","doi":"10.1016/j.jprocont.2025.103377","DOIUrl":"10.1016/j.jprocont.2025.103377","url":null,"abstract":"<div><div>In industrial processes, sensor aging and harsh field environments often introduce uncertainties into process data. These uncertainties obscure fault symptoms and undermine the precision of fault diagnosis. To address this challenge, this paper proposes a robust fault diagnosis model with interval distribution analysis for abnormal recognition under data uncertainties in complex industrial settings. Specifically, this research first transforms uncertain data collected from complex industrial sites into interval-valued data, which can globally capture the internal structural characteristics of data objects and effectively represent the uncertainty inherent in the single-valued data. Subsequently, a complete information principal component analysis (CIPCA)-based dimensionality reduction model is constructed to exploit the distribution information within the interval and extract interval fault features. Finally, an interval radial basis function neural network (IRBFNN) is developed to handle the interval upper and lower bound matrices through subtractive clustering algorithm, facilitating fault prediction and diagnosis in industrial processes contaminated by uncertainties. The key to discriminate the proposed method from many well-established fault diagnosis methods is its ability to cluster the interval fault features from uncertain data with embedded interval distribution analysis. The superiority of the proposed fault diagnosis model is validated by the Tennessee Eastman process (TEP).</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"147 ","pages":"Article 103377"},"PeriodicalIF":3.3,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143182260","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-22DOI: 10.1016/j.jprocont.2024.103355
Bingbing Shen , Xiaoyu Jiang , Le Yao , Jiusun Zeng
Most industrial process data is time series data and contains multi-mode characteristics, which poses difficulties and challenges in the establishment of soft sensing models. To address these issues, this paper proposes a Gaussian mixture based time series decomposition model. This model innovatively introduces Gaussian mixture distributions into the latent space and utilizes a time series decomposition module in the decoder to decompose complex distributions. On one hand, the latent variables of the Gaussian mixture distribution can better extract complex features from time series inputs. On the other hand, the time series decomposition module can break down and extract disentangled features from the time series perspective. Furthermore, to tackle the problem of poor fitting in peak or extreme data due to information imbalance, it generates virtual time series data. The generated virtual time series can complement the information of poorly fitted data, supplementing the original data, and contribute to a better soft sensing model. Finally, to validate the effectiveness of the proposed methods, the soft sensors based on the proposed model are applied to two real industrial cases. The experimental results show that the proposed models have superior predictive performance compared to other state-of-the-art methods.
{"title":"Gaussian mixture TimeVAE for industrial soft sensing with deep time series decomposition and generation","authors":"Bingbing Shen , Xiaoyu Jiang , Le Yao , Jiusun Zeng","doi":"10.1016/j.jprocont.2024.103355","DOIUrl":"10.1016/j.jprocont.2024.103355","url":null,"abstract":"<div><div>Most industrial process data is time series data and contains multi-mode characteristics, which poses difficulties and challenges in the establishment of soft sensing models. To address these issues, this paper proposes a Gaussian mixture based time series decomposition model. This model innovatively introduces Gaussian mixture distributions into the latent space and utilizes a time series decomposition module in the decoder to decompose complex distributions. On one hand, the latent variables of the Gaussian mixture distribution can better extract complex features from time series inputs. On the other hand, the time series decomposition module can break down and extract disentangled features from the time series perspective. Furthermore, to tackle the problem of poor fitting in peak or extreme data due to information imbalance, it generates virtual time series data. The generated virtual time series can complement the information of poorly fitted data, supplementing the original data, and contribute to a better soft sensing model. Finally, to validate the effectiveness of the proposed methods, the soft sensors based on the proposed model are applied to two real industrial cases. The experimental results show that the proposed models have superior predictive performance compared to other state-of-the-art methods.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"147 ","pages":"Article 103355"},"PeriodicalIF":3.3,"publicationDate":"2025-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143182257","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In the realm of industrial applications for machine learning, multiple challenges are frequently encountered, such as concept drift (CD) and the prohibitive costs associated with data labeling. CD refers to the scenario where the underlying data distribution of the model shifts over time, potentially deteriorating model performance. Addressing these challenges, this paper proposes an innovative semi-supervised CD detection method, specifically designed to tackle both CD and the high costs of data labeling in regression tasks. Initially, considering the high expense of acquiring labeled data in industrial application scenarios, a semi-supervised learning strategy based on self-training is utilized. In this strategy, prediction intervals generated by Conformal Prediction (CP) are used to select high-reliability pseudo-labels. Furthermore, to effectively address CD in real-world industrial settings, the Conformal Martingale (CM) is employed for real-time detection. This framework detects changes by identifying increases in martingale values when CD occurs. Upon detection, the model is promptly retrained using the most recent data following the drift. Finally, the proposed method is validated through experiments conducted on three datasets: the UCI dataset, the alumina evaporation process dataset, and the blast furnace ironmaking dataset. Experimental results demonstrate that the proposed semi-supervised method significantly enhances the performance of the original training model. The detection method accurately identifies CD and notably reduces test errors through model retraining, thereby improving the effectiveness of the model in real-world industrial applications.
在机器学习的工业应用领域，经常会遇到多种挑战，例如概念漂移（CD）和与数据标注相关的高昂成本。概念漂移指的是模型的基础数据分布随着时间的推移而发生变化，从而可能导致模型性能下降。为了应对这些挑战，本文提出了一种创新的半监督 CD 检测方法，专门用于解决回归任务中的 CD 和数据标注的高成本问题。首先，考虑到在工业应用场景中获取标记数据的高昂成本，本文采用了基于自我训练的半监督学习策略。在这一策略中，利用共形预测（CP）生成的预测区间来选择高可靠性的伪标签。此外，为了有效解决实际工业环境中的 CD 问题，还采用了共形马丁格尔（CM）进行实时检测。当 CD 发生时，该框架通过识别马氏值的增加来检测变化。一经检测到，就会立即使用漂移后的最新数据对模型进行重新训练。最后，通过在三个数据集（UCI 数据集、氧化铝蒸发过程数据集和高炉炼铁数据集）上进行实验，对所提出的方法进行了验证。实验结果表明，所提出的半监督方法显著提高了原始训练模型的性能。该检测方法能准确识别 CD，并通过模型再训练显著减少了测试误差，从而提高了模型在实际工业应用中的有效性。
{"title":"Semi-supervised concept drift detection and adaptation based on conformal martingale framework","authors":"Yu Zhang, Ping Zhou, Ruiyao Zhang, Shaowen Lu, Tianyou Chai","doi":"10.1016/j.jprocont.2025.103374","DOIUrl":"10.1016/j.jprocont.2025.103374","url":null,"abstract":"<div><div>In the realm of industrial applications for machine learning, multiple challenges are frequently encountered, such as concept drift (CD) and the prohibitive costs associated with data labeling. CD refers to the scenario where the underlying data distribution of the model shifts over time, potentially deteriorating model performance. Addressing these challenges, this paper proposes an innovative semi-supervised CD detection method, specifically designed to tackle both CD and the high costs of data labeling in regression tasks. Initially, considering the high expense of acquiring labeled data in industrial application scenarios, a semi-supervised learning strategy based on self-training is utilized. In this strategy, prediction intervals generated by Conformal Prediction (CP) are used to select high-reliability pseudo-labels. Furthermore, to effectively address CD in real-world industrial settings, the Conformal Martingale (CM) is employed for real-time detection. This framework detects changes by identifying increases in martingale values when CD occurs. Upon detection, the model is promptly retrained using the most recent data following the drift. Finally, the proposed method is validated through experiments conducted on three datasets: the UCI dataset, the alumina evaporation process dataset, and the blast furnace ironmaking dataset. Experimental results demonstrate that the proposed semi-supervised method significantly enhances the performance of the original training model. The detection method accurately identifies CD and notably reduces test errors through model retraining, thereby improving the effectiveness of the model in real-world industrial applications.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"147 ","pages":"Article 103374"},"PeriodicalIF":3.3,"publicationDate":"2025-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143182256","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号