Reliability Engineering & System Safety最新文献_第9页

A new approach for product reliability prediction by considering the production factory lifecycle information

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2025-02-15 DOI: 10.1016/j.ress.2025.110915

Shashi Bhushan Gunjan, D.S. Srinivasu, Ramesh Babu N

Product reliability prediction is essential for OEMs to plan product maintenance and design improvement. Traditional approaches rely on 'product' lifecycle data for reliability prediction, often not capturing the uncertainties in OEMs' decision-making. To address this, the present work focuses on 'factory' lifecycle information in reliability prediction by introducing the concept of 'factory age,' i.e., cumulative interval for the observed factory lifecycle. The failure times for each product, in each factory age interval, were used to estimate the Weibull parameters, creating temporal data. A combination of grey- and support vector machine (SVM)-models, which complement each other in capturing global and local trends, and handling uncertainty from limited temporal data, was proposed to forecast the Weibull parameters accurately in the future factory age interval. The proposed approach was validated on two failure modes in a factory-producing turning centers, using data from the first 11 factory age intervals for model development. Reliability predictions for the last three intervals achieved root mean square errors (RMSEs) of 0.67 % and 1.48 % for failure modes I and II. Comparatively, individual grey (4.37 %, 5.11 %) and SVM (8.03 %, 10.60 %) models yielded higher RMSEs, while other reported models in literature showed in the range of 1.63 %–34.07 %, demonstrating the proposed approach's efficacy.

{"title":"A new approach for product reliability prediction by considering the production factory lifecycle information","authors":"Shashi Bhushan Gunjan, D.S. Srinivasu, Ramesh Babu N","doi":"10.1016/j.ress.2025.110915","DOIUrl":"10.1016/j.ress.2025.110915","url":null,"abstract":"<div><div>Product reliability prediction is essential for OEMs to plan product maintenance and design improvement. Traditional approaches rely on '<em>product</em>' lifecycle data for reliability prediction, often not capturing the uncertainties in OEMs' decision-making. To address this, the present work focuses on '<em>factory' lifecycle information</em> in reliability prediction by introducing the concept of '<em>factory age,' i.e.,</em> cumulative interval for the observed <em>factory lifecycle</em>. The failure times for each product<em>,</em> in each <em>factory age interval,</em> were used to estimate the Weibull parameters, creating temporal data. A combination of grey- and support vector machine (SVM)-models, which complement each other in capturing global and local trends, and handling uncertainty from limited temporal data, was proposed to forecast the Weibull parameters accurately in the future <em>factory age interval</em>. The proposed approach was validated on two failure modes in a factory-producing turning centers, using data from the first 11 factory age intervals for model development. Reliability predictions for the last three intervals achieved root mean square errors (RMSEs) of 0.67 % and 1.48 % for failure modes I and II. Comparatively, individual grey (4.37 %, 5.11 %) and SVM (8.03 %, 10.60 %) models yielded higher RMSEs, while other reported models in literature showed in the range of 1.63 %–34.07 %, demonstrating the proposed approach's efficacy.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"258 ","pages":"Article 110915"},"PeriodicalIF":9.4,"publicationDate":"2025-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143437029","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

AGFCN:A bearing fault diagnosis method for high-speed train bogie under complex working conditions

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2025-02-15 DOI: 10.1016/j.ress.2025.110907

Deqiang He , Jinxin Wu , Zhenzhen Jin , ChengGeng Huang , Zexian Wei , Cai Yi

The operating conditions of high-speed train bogie (HSTB) bearings are sophisticated and changeable, making the nonlinear characteristics of bearing vibration signals more prominent and the noise in the signals more significant. To fully obtain the characteristic information in the vibration signal and improve the accuracy of HSTB bearing fault diagnosis, this paper fully considers the working conditions of HSTB bearing with intense noise and variable load. A fault diagnosis framework of adaptive graph framelet convolutional network (AGFCN) is proposed. Firstly, the vibration signal is constructed into a graph to obtain the characteristic information between the sample topologies. To better adapt to the complex and changeable working conditions of HSTB bearings, a neural network with learnable weight vectors is proposed to achieve a dynamic learning graph structure. Then, considering the practical factors of harrowing fault feature extraction in an intense noise background, a graph convolution based on framelet transform is designed. The framelet transform technology is used to reduce the signal interference and increase the model's feature learning capability. Finally, the actual data of the HSTB bearing test bench verify the reliability of AGFCN, which has significant advantages compared with six advanced models.

{"title":"AGFCN:A bearing fault diagnosis method for high-speed train bogie under complex working conditions","authors":"Deqiang He , Jinxin Wu , Zhenzhen Jin , ChengGeng Huang , Zexian Wei , Cai Yi","doi":"10.1016/j.ress.2025.110907","DOIUrl":"10.1016/j.ress.2025.110907","url":null,"abstract":"<div><div>The operating conditions of high-speed train bogie (HSTB) bearings are sophisticated and changeable, making the nonlinear characteristics of bearing vibration signals more prominent and the noise in the signals more significant. To fully obtain the characteristic information in the vibration signal and improve the accuracy of HSTB bearing fault diagnosis, this paper fully considers the working conditions of HSTB bearing with intense noise and variable load. A fault diagnosis framework of adaptive graph framelet convolutional network (AGFCN) is proposed. Firstly, the vibration signal is constructed into a graph to obtain the characteristic information between the sample topologies. To better adapt to the complex and changeable working conditions of HSTB bearings, a neural network with learnable weight vectors is proposed to achieve a dynamic learning graph structure. Then, considering the practical factors of harrowing fault feature extraction in an intense noise background, a graph convolution based on framelet transform is designed. The framelet transform technology is used to reduce the signal interference and increase the model's feature learning capability. Finally, the actual data of the HSTB bearing test bench verify the reliability of AGFCN, which has significant advantages compared with six advanced models.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"258 ","pages":"Article 110907"},"PeriodicalIF":9.4,"publicationDate":"2025-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143429643","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Reconstruction of 3-D pipeline defect profile based on MFL signals and hybrid neural networks

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2025-02-13 DOI: 10.1016/j.ress.2025.110890

Yinuo Chen , Zhigang Tian , Haotian Wei , Shaohua Dong

The pipelines' in-line inspection (ILI) is critical within the integrity management framework in the oil and gas industry. Furthermore, the reconstruction of defects' three-dimensional (3-D) profile using the magnetic flux leakage (MFL) signals acquired has great significance. However, most existing methods only focus on estimating defect sizes or shape parameters instead of the defect's 3-D profile. This study proposes an innovative approach for reconstructing the defect profile using a novel hybrid neural network to accurately and efficiently map three-axial MFL signals to the defects' 3-D profile. This paper utilizes the neural ordinary differential equation (ODE) as a module within the neural network architecture. The neural ODE is used to map the processed MFL signals to the spatial position of each point on the defective concave surface. Additionally, the model incorporates the Fourier integration kernel (FIK) to enhance computational efficiency. The proposed model is trained using finite element (FE) simulation data and then transferred to an experimental dataset, which addresses the challenge of limited availability of experimental data while maintaining accuracy. Furthermore, the proposed method also exhibits a high degree of accuracy in reconstructing the rotational angles of the defects. Therefore, the proposed method helps visualize defects in underground pipes via the analysis of MFL signals, facilitating operators in undertaking subsequent maintenance measures and providing a foundation for pipeline digital integrity management.

{"title":"Reconstruction of 3-D pipeline defect profile based on MFL signals and hybrid neural networks","authors":"Yinuo Chen , Zhigang Tian , Haotian Wei , Shaohua Dong","doi":"10.1016/j.ress.2025.110890","DOIUrl":"10.1016/j.ress.2025.110890","url":null,"abstract":"<div><div>The pipelines' in-line inspection (ILI) is critical within the integrity management framework in the oil and gas industry. Furthermore, the reconstruction of defects' three-dimensional (3-D) profile using the magnetic flux leakage (MFL) signals acquired has great significance. However, most existing methods only focus on estimating defect sizes or shape parameters instead of the defect's 3-D profile. This study proposes an innovative approach for reconstructing the defect profile using a novel hybrid neural network to accurately and efficiently map three-axial MFL signals to the defects' 3-D profile. This paper utilizes the neural ordinary differential equation (ODE) as a module within the neural network architecture. The neural ODE is used to map the processed MFL signals to the spatial position of each point on the defective concave surface. Additionally, the model incorporates the Fourier integration kernel (FIK) to enhance computational efficiency. The proposed model is trained using finite element (FE) simulation data and then transferred to an experimental dataset, which addresses the challenge of limited availability of experimental data while maintaining accuracy. Furthermore, the proposed method also exhibits a high degree of accuracy in reconstructing the rotational angles of the defects. Therefore, the proposed method helps visualize defects in underground pipes via the analysis of MFL signals, facilitating operators in undertaking subsequent maintenance measures and providing a foundation for pipeline digital integrity management.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"258 ","pages":"Article 110890"},"PeriodicalIF":9.4,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143421254","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

SIGTN: A novel structural Infomax Graph Transfer Networks for rotating machinery fault diagnosis in cross-condition and cross-equipment scenarios

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2025-02-13 DOI: 10.1016/j.ress.2025.110898

Kongliang Zhang , Hongkun Li , Shunxin Cao , Chen Yang , Wei Xiang

Graph-based networks have proven effective in node classification for diagnosing faults in rotating machinery. However, current graph neural networks often prioritize local information over global influences, hindering cross-graph transfer diagnosis in unlabeled graphs. To address these challenges, we propose SIGTN (Structure Infomax Graph Transfer Network), a novel algorithm for cross-graph diagnosis. Initially, raw and corrupted graph data is individually fed into the feature extractor, enhancing learned node representations to capture global structural properties by maximizing local-global mutual information. The node classifier then predicts labels based on these representations. During training process, both the feature extractor and node classifiers are trained concurrently to minimize cross-entropy loss for labeled nodes. Additionally, a conditional domain adversarial network alleviates distributional disparities between source and target domain graphs. Finally, experimental validation across various datasets demonstrates SIGTN's effectiveness in handling cross-graph transfer across different rotation speeds, loads, and equipment.

{"title":"SIGTN: A novel structural Infomax Graph Transfer Networks for rotating machinery fault diagnosis in cross-condition and cross-equipment scenarios","authors":"Kongliang Zhang , Hongkun Li , Shunxin Cao , Chen Yang , Wei Xiang","doi":"10.1016/j.ress.2025.110898","DOIUrl":"10.1016/j.ress.2025.110898","url":null,"abstract":"<div><div>Graph-based networks have proven effective in node classification for diagnosing faults in rotating machinery. However, current graph neural networks often prioritize local information over global influences, hindering cross-graph transfer diagnosis in unlabeled graphs. To address these challenges, we propose SIGTN (Structure Infomax Graph Transfer Network), a novel algorithm for cross-graph diagnosis. Initially, raw and corrupted graph data is individually fed into the feature extractor, enhancing learned node representations to capture global structural properties by maximizing local-global mutual information. The node classifier then predicts labels based on these representations. During training process, both the feature extractor and node classifiers are trained concurrently to minimize cross-entropy loss for labeled nodes. Additionally, a conditional domain adversarial network alleviates distributional disparities between source and target domain graphs. Finally, experimental validation across various datasets demonstrates SIGTN's effectiveness in handling cross-graph transfer across different rotation speeds, loads, and equipment.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"258 ","pages":"Article 110898"},"PeriodicalIF":9.4,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143429646","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Advancing stochastic modeling for nonlinear problems: Leveraging the transformation law of probability density

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2025-02-13 DOI: 10.1016/j.ress.2025.110895

Qais Saifi , Huapeng Wu , William Brace

In engineering, uncertainties pervade product lifecycles, presenting significant challenges to design reliability and safety, particularly in safety-sensitive industries such as nuclear. Stochastic simulations, leveraging Monte Carlo Sampling, machine learning, and parallel computing, are indispensable for addressing these uncertainties. However, they often overlook the direct influence of prediction models on predicted probability distributions, compromising both efficiency and accuracy. This paper thoroughly investigates the impact of prediction models on predicted probability distributions, presenting a novel mathematical framework to establish the transformation law of probability density. Additionally, we develop the Finite Cell Weight Variation method based on this transformation law. The proposed method seamlessly integrates prediction models into state probability predictions, enhancing reliability assessments while preserving high levels of accuracy and computational efficiency. We illustrate the method's effectiveness with practical examples and validation using Latin Hypercube Sampling (LHC), where several input variables are statistically determined. Our estimation of the probability of the predicted state closely aligns with results obtained using LHC. Furthermore, we explore the implications of our findings and outline future directions in stochastic simulations aimed at strengthening reliability assessments.

{"title":"Advancing stochastic modeling for nonlinear problems: Leveraging the transformation law of probability density","authors":"Qais Saifi , Huapeng Wu , William Brace","doi":"10.1016/j.ress.2025.110895","DOIUrl":"10.1016/j.ress.2025.110895","url":null,"abstract":"<div><div>In engineering, uncertainties pervade product lifecycles, presenting significant challenges to design reliability and safety, particularly in safety-sensitive industries such as nuclear. Stochastic simulations, leveraging Monte Carlo Sampling, machine learning, and parallel computing, are indispensable for addressing these uncertainties. However, they often overlook the direct influence of prediction models on predicted probability distributions, compromising both efficiency and accuracy. This paper thoroughly investigates the impact of prediction models on predicted probability distributions, presenting a novel mathematical framework to establish the transformation law of probability density. Additionally, we develop the Finite Cell Weight Variation method based on this transformation law. The proposed method seamlessly integrates prediction models into state probability predictions, enhancing reliability assessments while preserving high levels of accuracy and computational efficiency. We illustrate the method's effectiveness with practical examples and validation using Latin Hypercube Sampling (LHC), where several input variables are statistically determined. Our estimation of the probability of the predicted state closely aligns with results obtained using LHC. Furthermore, we explore the implications of our findings and outline future directions in stochastic simulations aimed at strengthening reliability assessments.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"258 ","pages":"Article 110895"},"PeriodicalIF":9.4,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143421327","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Physics-informed neural network supported wiener process for degradation modeling and reliability prediction

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2025-02-12 DOI: 10.1016/j.ress.2025.110906

Zhongze He , Shaoping Wang , Jian Shi , Di Liu , Xiaochuan Duan , Yaoxing Shang

Due to strong data-processing capabilities, machine learning haves been widely applied and combined with stochastic processes to quantify the inherent uncertainty in degradation modeling. These approaches typically first extract health index using machine learning methods, then model them using stochastic processes. While, the machine learning models and stochastic processes are independent of each other, making it difficult to ensure their mutual compatibility. Furthermore, actual available data is often limited, which restricts the accuracy of extracting health indexes through machine learning methods. Hence, this paper proposes a prediction method based on physics-informed neural network supported Wiener process, which includes offline modeling and online prediction stages. In the offline modeling phase, degradation path is fitted using a deep network framework, and degradation mechanics-related prior physical knowledge is embedded into the network along with the Wiener process through parametric expression. Accordingly, a compound loss function is designed to simultaneously train network parameters and process parameters. In the online prediction phase, real-time data is integrated using Bayesian inference methods to update the process parameters, ensuring the robustness of the model. The effectiveness of this method is confirmed using actual datasets, highlighting that the accuracy can be guaranteed even without path information and/or sufficient data.

{"title":"Physics-informed neural network supported wiener process for degradation modeling and reliability prediction","authors":"Zhongze He , Shaoping Wang , Jian Shi , Di Liu , Xiaochuan Duan , Yaoxing Shang","doi":"10.1016/j.ress.2025.110906","DOIUrl":"10.1016/j.ress.2025.110906","url":null,"abstract":"<div><div>Due to strong data-processing capabilities, machine learning haves been widely applied and combined with stochastic processes to quantify the inherent uncertainty in degradation modeling. These approaches typically first extract health index using machine learning methods, then model them using stochastic processes. While, the machine learning models and stochastic processes are independent of each other, making it difficult to ensure their mutual compatibility. Furthermore, actual available data is often limited, which restricts the accuracy of extracting health indexes through machine learning methods. Hence, this paper proposes a prediction method based on physics-informed neural network supported Wiener process, which includes offline modeling and online prediction stages. In the offline modeling phase, degradation path is fitted using a deep network framework, and degradation mechanics-related prior physical knowledge is embedded into the network along with the Wiener process through parametric expression. Accordingly, a compound loss function is designed to simultaneously train network parameters and process parameters. In the online prediction phase, real-time data is integrated using Bayesian inference methods to update the process parameters, ensuring the robustness of the model. The effectiveness of this method is confirmed using actual datasets, highlighting that the accuracy can be guaranteed even without path information and/or sufficient data.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"258 ","pages":"Article 110906"},"PeriodicalIF":9.4,"publicationDate":"2025-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143421255","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Remaining useful life prediction based on graph feature attention networks with missing multi-sensor features

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2025-02-12 DOI: 10.1016/j.ress.2025.110902

Yu Wang , Shangjing Peng , Hong Wang , Mingquan Zhang , Hongrui Cao , Liwei Ma

Prognostics and health management (PHM) is important to ensure the reliable operation of industrial equipment, where monitoring the degradation process of machinery through multi-source sensors for remaining useful life (RUL) prediction is one of the key tasks. In recent years, deep learning-based time series forecasting methods have been proposed to predict RUL as they have strong capability on temporal correlation modeling for time series gathered by sensors. However, these methods usually operate under the assumption of a static, fixed-dimensional feature set. The proliferation of sensors inevitably escalates the probability of missing and anomalous features within measurement data, thereby causing the dimensions of input features to dynamically fluctuate over time. Therefore, this paper proposes a Graph Feature-Gated Graph Attention Network (GF-GGAT), which is capable of fusing multi-sensor data with partially missing sensor data and performing RUL prediction. First, the problem of spatio-temporal map construction when some sensor data are missing is solved by introducing dynamic time regularization. Second, the feature-deficient multi-sensor data are inductively learned through graph feature transformation and stepwise graph convolution. Finally, spatio-temporal features are extracted by a gated graph attention network (GGAT) to accomplish RUL prediction. Two case studies demonstrate the superiority of the proposed method over state-of-the-art RUL prediction methods.

{"title":"Remaining useful life prediction based on graph feature attention networks with missing multi-sensor features","authors":"Yu Wang , Shangjing Peng , Hong Wang , Mingquan Zhang , Hongrui Cao , Liwei Ma","doi":"10.1016/j.ress.2025.110902","DOIUrl":"10.1016/j.ress.2025.110902","url":null,"abstract":"<div><div>Prognostics and health management (PHM) is important to ensure the reliable operation of industrial equipment, where monitoring the degradation process of machinery through multi-source sensors for remaining useful life (RUL) prediction is one of the key tasks. In recent years, deep learning-based time series forecasting methods have been proposed to predict RUL as they have strong capability on temporal correlation modeling for time series gathered by sensors. However, these methods usually operate under the assumption of a static, fixed-dimensional feature set. The proliferation of sensors inevitably escalates the probability of missing and anomalous features within measurement data, thereby causing the dimensions of input features to dynamically fluctuate over time. Therefore, this paper proposes a Graph Feature-Gated Graph Attention Network (GF-GGAT), which is capable of fusing multi-sensor data with partially missing sensor data and performing RUL prediction. First, the problem of spatio-temporal map construction when some sensor data are missing is solved by introducing dynamic time regularization. Second, the feature-deficient multi-sensor data are inductively learned through graph feature transformation and stepwise graph convolution. Finally, spatio-temporal features are extracted by a gated graph attention network (GGAT) to accomplish RUL prediction. Two case studies demonstrate the superiority of the proposed method over state-of-the-art RUL prediction methods.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"258 ","pages":"Article 110902"},"PeriodicalIF":9.4,"publicationDate":"2025-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143421326","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Physics-informed Gaussian process probabilistic modeling with multi-source data for prognostics of degradation processes

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2025-02-12 DOI: 10.1016/j.ress.2025.110893

Chen Jiang , Teng Zhong , Hyunhee Choi , Byeng D. Youn

The integration of physics-based and data-driven methods in prognostics has become increasingly important in understanding the underlying degradation process using available knowledge. In this paper, we propose a physics-informed Gaussian process modeling method with multi-source data (PIGP-MD) that improves the predictive power of standard Gaussian process model by incorporating the prior knowledge from both physical models and historical data, and the current knowledge from limited observation data. We first propose a PIGP modeling method that incorporates the prior knowledge solely from physical models. In PIGP, a standard GP directly trained with current observation data is used to truncate the random realizations generated from physical models within the confidence bounds of the GP prediction. The truncated random realizations are used to derive the PIGP's nonstationary priors including the mean and autocovariance functions, as the unknown stochastic degradation process is considered as a nonstationary Gaussian process. Built upon PIGP, we propose PIGP-MD to filter credible knowledge from historical data. Some random realizations are generated from the credible knowledge from historical data and combined with the truncated random realizations generated from physical models. The combined random realizations are used to derive the nonstationary priors of PIGP-MD in the same way. With the physics model-informed and historical data-driven nonstationary priors as well as the current observation data, we can efficiently obtain the posterior future prediction without any parameter optimization. We demonstrate the applicability and efficacy of our proposed method for fatigue damage prognostics.

{"title":"Physics-informed Gaussian process probabilistic modeling with multi-source data for prognostics of degradation processes","authors":"Chen Jiang , Teng Zhong , Hyunhee Choi , Byeng D. Youn","doi":"10.1016/j.ress.2025.110893","DOIUrl":"10.1016/j.ress.2025.110893","url":null,"abstract":"<div><div>The integration of physics-based and data-driven methods in prognostics has become increasingly important in understanding the underlying degradation process using available knowledge. In this paper, we propose a physics-informed Gaussian process modeling method with multi-source data (PIGP-MD) that improves the predictive power of standard Gaussian process model by incorporating the prior knowledge from both physical models and historical data, and the current knowledge from limited observation data. We first propose a PIGP modeling method that incorporates the prior knowledge solely from physical models. In PIGP, a standard GP directly trained with current observation data is used to truncate the random realizations generated from physical models within the confidence bounds of the GP prediction. The truncated random realizations are used to derive the PIGP's nonstationary priors including the mean and autocovariance functions, as the unknown stochastic degradation process is considered as a nonstationary Gaussian process. Built upon PIGP, we propose PIGP-MD to filter credible knowledge from historical data. Some random realizations are generated from the credible knowledge from historical data and combined with the truncated random realizations generated from physical models. The combined random realizations are used to derive the nonstationary priors of PIGP-MD in the same way. With the physics model-informed and historical data-driven nonstationary priors as well as the current observation data, we can efficiently obtain the posterior future prediction without any parameter optimization. We demonstrate the applicability and efficacy of our proposed method for fatigue damage prognostics.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"258 ","pages":"Article 110893"},"PeriodicalIF":9.4,"publicationDate":"2025-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143429644","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Bayesian inference-assisted reliability analysis framework for robotic motion systems in future factories

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2025-02-11 DOI: 10.1016/j.ress.2025.110894

Shuoshuo Shen , Jin Cheng , Zhenyu Liu , Jianrong Tan , Dequan Zhang

Reliability assessment of robotic motion systems subject to complex dynamic properties and multi-source uncertainties in open environments registers an important yet challenging task. To tackle this task, this study proposes a new reliability analysis framework for robotic motion systems, which incorporates the moment-based method and Bayesian inference-guided probabilistic model updating strategy. To start with, the fractional exponential moments calculated by the sparse grid method are adopted to quantify the uncertainty of performance indexes for robotic motion systems. Subsequently, a versatile mixture probability distribution model is established to evaluate the reliability of the performance indexes, facilitating the probability distribution modeling of various features. To capture sufficient uncertainty information of the system performance, two solution strategies for probabilistic model parameters are developed by incorporating the direct and sequential Bayesian updating methods. With fractional exponential moments, the proposed probability model is calibrated to reconstruct the probability distribution and calculate the failure probability for robotic motion systems. The effectiveness of the proposed framework is validated by three numerical examples, wherein Monte Carlo simulation and other prevailing methods are performed for comparison. The case studies indicate that the proposed framework is viable to assess the performance reliability of robotic motion systems with satisfactory computational accuracy and efficiency.

{"title":"Bayesian inference-assisted reliability analysis framework for robotic motion systems in future factories","authors":"Shuoshuo Shen , Jin Cheng , Zhenyu Liu , Jianrong Tan , Dequan Zhang","doi":"10.1016/j.ress.2025.110894","DOIUrl":"10.1016/j.ress.2025.110894","url":null,"abstract":"<div><div>Reliability assessment of robotic motion systems subject to complex dynamic properties and multi-source uncertainties in open environments registers an important yet challenging task. To tackle this task, this study proposes a new reliability analysis framework for robotic motion systems, which incorporates the moment-based method and Bayesian inference-guided probabilistic model updating strategy. To start with, the fractional exponential moments calculated by the sparse grid method are adopted to quantify the uncertainty of performance indexes for robotic motion systems. Subsequently, a versatile mixture probability distribution model is established to evaluate the reliability of the performance indexes, facilitating the probability distribution modeling of various features. To capture sufficient uncertainty information of the system performance, two solution strategies for probabilistic model parameters are developed by incorporating the direct and sequential Bayesian updating methods. With fractional exponential moments, the proposed probability model is calibrated to reconstruct the probability distribution and calculate the failure probability for robotic motion systems. The effectiveness of the proposed framework is validated by three numerical examples, wherein Monte Carlo simulation and other prevailing methods are performed for comparison. The case studies indicate that the proposed framework is viable to assess the performance reliability of robotic motion systems with satisfactory computational accuracy and efficiency.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"258 ","pages":"Article 110894"},"PeriodicalIF":9.4,"publicationDate":"2025-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143437026","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Dynamic model-driven dictionary learning-inspired domain adaptation strategy for cross-domain bearing fault diagnosis

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2025-02-11 DOI: 10.1016/j.ress.2025.110905

Zhengyu Du , Dongdong Liu , Lingli Cui

Cross-domain fault diagnosis methods have been extensively investigated to improve practical engineering implications for data-driven models. However, the annotated data in practical applications is often insufficient, which makes it difficult to train the model effectively. Additionally, existing methods typically transfer knowledge learned from one device to another, where collected data from different devices exhibit different distribution representations. To address the above issues, a dynamic model-driven dictionary learning-inspired domain adaptation strategy is proposed. First, a novel dynamic model that quantitatively considers the effects of slip and lubrication is established to generate a mass of labeled data. Second, a novel deep discriminative transfer dictionary neural network (DDTDNN) is developed, in which a new multi-layer deep dictionary learning module (MDDL) and an adaptive bandwidth maximum mean discrepancy (ABMMD) metric are designed. MDDL leverages iterative soft thresholding and gradient descent processes to extract domain invariant representation within sparse representation space, while ABMMD is incorporated into the loss function and works alongside the classification loss to jointly influence the model. This new metric can dynamically set kernel widths by a median heuristic method, which helps the model to adapt the scale of the data and align feature distributions more effectively. The effectiveness of DDTDNN is validated on two cross-domain datasets. Experiment results show that DDTDNN achieves classification accuracies of 99.1 %, and 98.5 %, respectively, which outperforms several state-of-the-art methods.

{"title":"Dynamic model-driven dictionary learning-inspired domain adaptation strategy for cross-domain bearing fault diagnosis","authors":"Zhengyu Du , Dongdong Liu , Lingli Cui","doi":"10.1016/j.ress.2025.110905","DOIUrl":"10.1016/j.ress.2025.110905","url":null,"abstract":"<div><div>Cross-domain fault diagnosis methods have been extensively investigated to improve practical engineering implications for data-driven models. However, the annotated data in practical applications is often insufficient, which makes it difficult to train the model effectively. Additionally, existing methods typically transfer knowledge learned from one device to another, where collected data from different devices exhibit different distribution representations. To address the above issues, a dynamic model-driven dictionary learning-inspired domain adaptation strategy is proposed. First, a novel dynamic model that quantitatively considers the effects of slip and lubrication is established to generate a mass of labeled data. Second, a novel deep discriminative transfer dictionary neural network (DDTDNN) is developed, in which a new multi-layer deep dictionary learning module (MDDL) and an adaptive bandwidth maximum mean discrepancy (ABMMD) metric are designed. MDDL leverages iterative soft thresholding and gradient descent processes to extract domain invariant representation within sparse representation space, while ABMMD is incorporated into the loss function and works alongside the classification loss to jointly influence the model. This new metric can dynamically set kernel widths by a median heuristic method, which helps the model to adapt the scale of the data and align feature distributions more effectively. The effectiveness of DDTDNN is validated on two cross-domain datasets. Experiment results show that DDTDNN achieves classification accuracies of 99.1 %, and 98.5 %, respectively, which outperforms several state-of-the-art methods.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"258 ","pages":"Article 110905"},"PeriodicalIF":9.4,"publicationDate":"2025-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143421325","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0