Pub Date : 2023-06-05DOI: 10.1109/icphm57936.2023.10194087
{"title":"Toc","authors":"","doi":"10.1109/icphm57936.2023.10194087","DOIUrl":"https://doi.org/10.1109/icphm57936.2023.10194087","url":null,"abstract":"","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116441682","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-05DOI: 10.1109/ICPHM57936.2023.10193933
Cheng Chen, Meng Mei, Haidong Shao, Pei Liang
Turnouts play a crucial role in the safety and efficiency of trains. Traditionally, the railway turnout fault diagnosis method relied on vectorized data from time series monitoring. However, such data format fails to fully capture the signals' spatial structure and profile information, which are crucial for inspectors to analyze and make judgments. In this study, a novel fault diagnosis method for the railway is developed with the hyperdisk-based supervised tensor machine (HDSTM) and monitoring signal images, which solves the limitations of the existing method. Besides, a novel tensor-form multi-class classifier called HDSTM is proposed to address the limitation of the convex-hull-based support tensor machine (CHSTM) in the underestimation problem. First, the time series monitoring signals are preprocessed and transformed into two-dimensional images. Next, CANDECOMP/PARAFAC decomposition is used to calculate the feature tensor. Then, the HDSTM model is built with the extracted feature tensor to implement the fault diagnosis. The proposed method's performance is evaluated using real-world operational current and power datasets. Experiment results indicate that the proposed method achieved higher average accuracy and precision than existing methods.
{"title":"A support tensor machine-based fault diagnosis method for railway turnout","authors":"Cheng Chen, Meng Mei, Haidong Shao, Pei Liang","doi":"10.1109/ICPHM57936.2023.10193933","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10193933","url":null,"abstract":"Turnouts play a crucial role in the safety and efficiency of trains. Traditionally, the railway turnout fault diagnosis method relied on vectorized data from time series monitoring. However, such data format fails to fully capture the signals' spatial structure and profile information, which are crucial for inspectors to analyze and make judgments. In this study, a novel fault diagnosis method for the railway is developed with the hyperdisk-based supervised tensor machine (HDSTM) and monitoring signal images, which solves the limitations of the existing method. Besides, a novel tensor-form multi-class classifier called HDSTM is proposed to address the limitation of the convex-hull-based support tensor machine (CHSTM) in the underestimation problem. First, the time series monitoring signals are preprocessed and transformed into two-dimensional images. Next, CANDECOMP/PARAFAC decomposition is used to calculate the feature tensor. Then, the HDSTM model is built with the extracted feature tensor to implement the fault diagnosis. The proposed method's performance is evaluated using real-world operational current and power datasets. Experiment results indicate that the proposed method achieved higher average accuracy and precision than existing methods.","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126915808","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-05DOI: 10.1109/ICPHM57936.2023.10194231
Zhong Ren, X. Qin, Qing Zhang, Yuantao Sun
Fatigue damage and subsequent failure account for the majority of crane structural failure. In this paper, the fatigue damage evolution behavior of low carbon alloy steel Q355B under multi-axial proportional and non-proportional loading is studied. Using the strain field nephogram obtained by digital image correlation, and the indirect damage variable characterized by the unloading stiffness recorded by the tension-torsion composite extensometer, the fatigue damage evolution process is analyzed qualitatively. The fatigue failure process under different loading conditions is uniformly divided into three stages: meso-crack initiation stage, meso-crack propagation (macro-crack initiation) stage, and macro crack propagation stage. According to the damage mechanics, the corresponding turning points at different stages are quantitatively analyzed, and it is found that the damage threshold values corresponding to different loading conditions are different, that is, under the same strain control, the damage threshold values required for non-proportional loading are less than that for proportional loading, and there are differences under different strain control. Based on the damage evolution model and the shear damage model, combined with the test results, it is proved that the damage evolution mode caused by multi-axial non-proportional loading is different from that caused by proportional loading.
{"title":"Damage Evolution Characterization of Low Carbon Alloy Steel Based on Multiaxial Fatigue Test and DIC","authors":"Zhong Ren, X. Qin, Qing Zhang, Yuantao Sun","doi":"10.1109/ICPHM57936.2023.10194231","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194231","url":null,"abstract":"Fatigue damage and subsequent failure account for the majority of crane structural failure. In this paper, the fatigue damage evolution behavior of low carbon alloy steel Q355B under multi-axial proportional and non-proportional loading is studied. Using the strain field nephogram obtained by digital image correlation, and the indirect damage variable characterized by the unloading stiffness recorded by the tension-torsion composite extensometer, the fatigue damage evolution process is analyzed qualitatively. The fatigue failure process under different loading conditions is uniformly divided into three stages: meso-crack initiation stage, meso-crack propagation (macro-crack initiation) stage, and macro crack propagation stage. According to the damage mechanics, the corresponding turning points at different stages are quantitatively analyzed, and it is found that the damage threshold values corresponding to different loading conditions are different, that is, under the same strain control, the damage threshold values required for non-proportional loading are less than that for proportional loading, and there are differences under different strain control. Based on the damage evolution model and the shear damage model, combined with the test results, it is proved that the damage evolution mode caused by multi-axial non-proportional loading is different from that caused by proportional loading.","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126829263","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-05DOI: 10.1109/ICPHM57936.2023.10193920
Peng Liu
This manuscript documents our approach to addressing the data challenge posted by ICPHM23 conference [1]. The task is a time series classification problem. We see two general toolsets can be used to complete the task and produce promising high accuracy for such a large data set. One is deep neural networks, and the other is gradient boosting. We choose gradient boosting. During feature preparation, we developed a customized C++ parallel computing software to extract all desired features. The manuscript includes our thought process and final cross validation results.
{"title":"Vibration Time Series Classification using Parallel Computing and XGBoost","authors":"Peng Liu","doi":"10.1109/ICPHM57936.2023.10193920","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10193920","url":null,"abstract":"This manuscript documents our approach to addressing the data challenge posted by ICPHM23 conference [1]. The task is a time series classification problem. We see two general toolsets can be used to complete the task and produce promising high accuracy for such a large data set. One is deep neural networks, and the other is gradient boosting. We choose gradient boosting. During feature preparation, we developed a customized C++ parallel computing software to extract all desired features. The manuscript includes our thought process and final cross validation results.","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123858378","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-05DOI: 10.1109/ICPHM57936.2023.10194222
Parvathy Sobha, Midhun Xavier, Praneeth Chandran
Gearbox faults can lead to significant damage and downtime in industrial machinery, resulting in substantial losses for manufacturers. Detection of faults in gears in the incipient state is essential to ensure safe and reliable operation of industrial machineries. In recent years, there has been an increasing interest in using machine learning algorithms to automate gearbox fault detection. This paper proposes a machine learning approach for identifying different categories of faults in a gearbox based on vibration signals. The proposed method was evaluated on a dataset of vibration signals collected from a two-stage gearbox under different operational conditions. The research is focused on developing a sequential neural network-based method for detecting multiple gear faults simultaneously. The results showed that the developed method achieved high training and validation accuracies and relatively low training and validation losses, indicating the model's ability to accurately detect and classify faults in gearboxes. The testing accuracies were also high, demonstrating the model's ability to generalize well to new data. The practical implications of the research are significant for improving the reliability and maintenance of gearboxes in various industrial applications. The developed method has the potential to reduce downtime, maintenance costs, and improve safety and efficiency.
{"title":"A Comprehensive Approach for Gearbox Fault Detection and Diagnosis Using Sequential Neural Networks","authors":"Parvathy Sobha, Midhun Xavier, Praneeth Chandran","doi":"10.1109/ICPHM57936.2023.10194222","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194222","url":null,"abstract":"Gearbox faults can lead to significant damage and downtime in industrial machinery, resulting in substantial losses for manufacturers. Detection of faults in gears in the incipient state is essential to ensure safe and reliable operation of industrial machineries. In recent years, there has been an increasing interest in using machine learning algorithms to automate gearbox fault detection. This paper proposes a machine learning approach for identifying different categories of faults in a gearbox based on vibration signals. The proposed method was evaluated on a dataset of vibration signals collected from a two-stage gearbox under different operational conditions. The research is focused on developing a sequential neural network-based method for detecting multiple gear faults simultaneously. The results showed that the developed method achieved high training and validation accuracies and relatively low training and validation losses, indicating the model's ability to accurately detect and classify faults in gearboxes. The testing accuracies were also high, demonstrating the model's ability to generalize well to new data. The practical implications of the research are significant for improving the reliability and maintenance of gearboxes in various industrial applications. The developed method has the potential to reduce downtime, maintenance costs, and improve safety and efficiency.","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"104 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114532695","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-05DOI: 10.1109/ICPHM57936.2023.10194079
Nobal B. Niraula, Hai Nguyen, Jennifer Kansal, Sean Hafner, Logan M. Branscum, Eric Brown, Ricardo Garcia
Service Difficulty Reports (SDRs) are reports submitted by aircraft operators and certified repair stations after they discover or experience a failure, malfunction, or defect while operating, or performing maintenance on an aircraft. The SDR records are rich in information pertaining to aviation safety. However, most of that data is not easily accessible as the problems are described in free form text. The text records often describe critical safety events such as depressurization, onboard fire, and runway excursion. Extracting critical information like the safety events in millions of records is labor intensive and infeasible without automated methods. In this study, we describe a machine learning approach to automatically discover depressurization safety events in SDR records. We are able to achieve the F1 score up to 95% to discover the depressurization events.
{"title":"Discovering Depressurization Events in Service Difficulty Reports using Machine Learning","authors":"Nobal B. Niraula, Hai Nguyen, Jennifer Kansal, Sean Hafner, Logan M. Branscum, Eric Brown, Ricardo Garcia","doi":"10.1109/ICPHM57936.2023.10194079","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194079","url":null,"abstract":"Service Difficulty Reports (SDRs) are reports submitted by aircraft operators and certified repair stations after they discover or experience a failure, malfunction, or defect while operating, or performing maintenance on an aircraft. The SDR records are rich in information pertaining to aviation safety. However, most of that data is not easily accessible as the problems are described in free form text. The text records often describe critical safety events such as depressurization, onboard fire, and runway excursion. Extracting critical information like the safety events in millions of records is labor intensive and infeasible without automated methods. In this study, we describe a machine learning approach to automatically discover depressurization safety events in SDR records. We are able to achieve the F1 score up to 95% to discover the depressurization events.","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130535230","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-05DOI: 10.1109/ICPHM57936.2023.10194033
Yan Li, Navid Zaman, J. Stecki, C. Stecki
Mechanical systems across most industries are mortal instruments, they will fail due to use or otherwise. Staying ahead of such catastrophes are crucial, especially in mission critical scenarios where loss of life is a very real danger. In such cases, corrective maintenance is too risky and scheduled maintenance is often costly; thus the collective shift towards predictive maintenance. Until recent advancements in artificial intelligence and sensor networks, such a strategy would not be so achievable. The ‘predictive’ aspect of thus type of maintenance implies that anomalies and failures are expected to be forecast ahead of occurrence - this can be accomplished with well placed sensors and sufficiently trained correlation methods. However, aspects such as shifting operating modes and varying sensor ranges make it difficult to make predictions solely using raw sensor data. This paper will outline methods and technologies to estimate healthy states of the monitored system to aid in the detection of failures before they affect function. Mechanical systems across most industries are mortal instruments, they will fail due to use or otherwise. Staying ahead of such catastrophes are crucial, especially in mission critical scenarios where loss of life is a very real danger. In such cases, corrective maintenance is too risky and scheduled maintenance is often costly; thus the collective shift towards predictive maintenance. Until recent advancements in artificial intelligence and sensor networks, such a strategy would not be so achievable. The ‘predictive’ aspect of thus type of maintenance implies that anomalies and failures are expected to be forecast ahead of occurrence - this can be accomplished with well placed sensors and sufficiently trained correlation methods. However, aspects such as shifting operating modes and varying sensor ranges make it difficult to make predictions solely using raw sensor data. This paper will outline methods and technologies to estimate healthy states of the monitored system to aid in the detection of failures before they affect function.M
{"title":"State Reconstruction: Generating a Reference for Improved Diagnostics","authors":"Yan Li, Navid Zaman, J. Stecki, C. Stecki","doi":"10.1109/ICPHM57936.2023.10194033","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194033","url":null,"abstract":"Mechanical systems across most industries are mortal instruments, they will fail due to use or otherwise. Staying ahead of such catastrophes are crucial, especially in mission critical scenarios where loss of life is a very real danger. In such cases, corrective maintenance is too risky and scheduled maintenance is often costly; thus the collective shift towards predictive maintenance. Until recent advancements in artificial intelligence and sensor networks, such a strategy would not be so achievable. The ‘predictive’ aspect of thus type of maintenance implies that anomalies and failures are expected to be forecast ahead of occurrence - this can be accomplished with well placed sensors and sufficiently trained correlation methods. However, aspects such as shifting operating modes and varying sensor ranges make it difficult to make predictions solely using raw sensor data. This paper will outline methods and technologies to estimate healthy states of the monitored system to aid in the detection of failures before they affect function. Mechanical systems across most industries are mortal instruments, they will fail due to use or otherwise. Staying ahead of such catastrophes are crucial, especially in mission critical scenarios where loss of life is a very real danger. In such cases, corrective maintenance is too risky and scheduled maintenance is often costly; thus the collective shift towards predictive maintenance. Until recent advancements in artificial intelligence and sensor networks, such a strategy would not be so achievable. The ‘predictive’ aspect of thus type of maintenance implies that anomalies and failures are expected to be forecast ahead of occurrence - this can be accomplished with well placed sensors and sufficiently trained correlation methods. However, aspects such as shifting operating modes and varying sensor ranges make it difficult to make predictions solely using raw sensor data. This paper will outline methods and technologies to estimate healthy states of the monitored system to aid in the detection of failures before they affect function.M","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129761633","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Rolling bearing of rotating machinery's key component will inevitably fail due to the complex and changeable operating environment such as variable speed, large disturbance, high and low temperature. It is quite challenging to obtain abundant labeled bearing fault samples because the rotating machinery is typically in a healthy and operational state. For addressing the issue, an intelligent fault diagnosis method based on a deep transfer learning network is proposed. First, a bidirectional gated recurrent unit (Bi-GRU) network is utilized to mine the latent relationship between labeled source domain samples and few labeled target domain samples, the parameters of Bi-GRU are trained to obtain the instance transfer bidirectional gated recurrent unit model (ITBi-GRU), and auxiliary samples are generated based on the ITBi-GRU. Second, as a feature transfer learning method, joint distribution adaptation is used to simultaneously decrease the distribution discrepancies between the generated auxiliary samples and the unlabeled target domain samples. Finally, extensive experiments are employed to evaluate the effectiveness of the proposed method in the case of scarce labeled samples.
{"title":"Intelligent fault diagnosis of rolling bearing based on a deep transfer learning network","authors":"Zhenghong Wu, Hongkai Jiang, Sicheng Zhang, Xin Wang, Haidong Shao, Haoxuan Dou","doi":"10.1109/ICPHM57936.2023.10194043","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194043","url":null,"abstract":"Rolling bearing of rotating machinery's key component will inevitably fail due to the complex and changeable operating environment such as variable speed, large disturbance, high and low temperature. It is quite challenging to obtain abundant labeled bearing fault samples because the rotating machinery is typically in a healthy and operational state. For addressing the issue, an intelligent fault diagnosis method based on a deep transfer learning network is proposed. First, a bidirectional gated recurrent unit (Bi-GRU) network is utilized to mine the latent relationship between labeled source domain samples and few labeled target domain samples, the parameters of Bi-GRU are trained to obtain the instance transfer bidirectional gated recurrent unit model (ITBi-GRU), and auxiliary samples are generated based on the ITBi-GRU. Second, as a feature transfer learning method, joint distribution adaptation is used to simultaneously decrease the distribution discrepancies between the generated auxiliary samples and the unlabeled target domain samples. Finally, extensive experiments are employed to evaluate the effectiveness of the proposed method in the case of scarce labeled samples.","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133978855","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-05DOI: 10.1109/ICPHM57936.2023.10194224
Rocco Cassandro, Quing Li, Zhaojun Li
The primary task in data mining is to find potential patterns or to discover hidden and useful knowledge from given data sets. However, with the increasing data quantity and exploding complexity, the capabilities of dealing with massive data becomes very crucial. The data preprocessing module is an integral part of data mining procedure, which aims to optimize input data usability for subsequent tasks such as classification, clustering, association analysis as well as other data mining algorithms. In general, data preprocessing procedures can effectively reduce the computational complexity while as possible to ensure accuracy and efficiency of prediction or classification, but meanwhile it even can assist to extract some unknown knowledge before applying more advanced data mining algorithms. This research proposes a three-patterns feature variables technique and an entropy-based data reduction (EBDR) algorithm for data preprocessing based on information theory. The goal is to explore high-purity subsets in which the values of an attribute are directly linked to specific class labels. The results of experiments demonstrate the efficacy of EBDR algorithm on datasets of varying sizes.
{"title":"An Entropy-based Data Reduction Method for Data Preprocessing","authors":"Rocco Cassandro, Quing Li, Zhaojun Li","doi":"10.1109/ICPHM57936.2023.10194224","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194224","url":null,"abstract":"The primary task in data mining is to find potential patterns or to discover hidden and useful knowledge from given data sets. However, with the increasing data quantity and exploding complexity, the capabilities of dealing with massive data becomes very crucial. The data preprocessing module is an integral part of data mining procedure, which aims to optimize input data usability for subsequent tasks such as classification, clustering, association analysis as well as other data mining algorithms. In general, data preprocessing procedures can effectively reduce the computational complexity while as possible to ensure accuracy and efficiency of prediction or classification, but meanwhile it even can assist to extract some unknown knowledge before applying more advanced data mining algorithms. This research proposes a three-patterns feature variables technique and an entropy-based data reduction (EBDR) algorithm for data preprocessing based on information theory. The goal is to explore high-purity subsets in which the values of an attribute are directly linked to specific class labels. The results of experiments demonstrate the efficacy of EBDR algorithm on datasets of varying sizes.","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133997941","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-05DOI: 10.1109/ICPHM57936.2023.10194057
H. Rasay, Fariba Azizi, Mehrnaz Salmani, F. Naderkhani
This paper focuses on development of joint optimal maintenance and production policy for a specific type of production system that allows for adjustable production rates. The rate of deterioration of the system is directly related to the production rate, with higher production rates resulting in greater expected deterioration. The system's deterioration can be controlled through two main actions: (1) scheduling and conducting maintenance actions referred to as maintenance policy; and (2) adjusting the production rate referred to as production policy. To determine the optimal actions given the system's state, a Markov decision process (MDP) is developed and a reinforcement learning algorithm, specifically a Q-learning algorithm, is utilized. The algorithm's hyper parameters are tuned using a value-iteration algorithm of dynamic programming. The goal is to minimize expected costs for the system over a finite planning horizon.
{"title":"A Reinforcement Learning Algorithm for Optimal Dynamic Policies of Joint Condition-based Maintenance and Condition-based Production","authors":"H. Rasay, Fariba Azizi, Mehrnaz Salmani, F. Naderkhani","doi":"10.1109/ICPHM57936.2023.10194057","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194057","url":null,"abstract":"This paper focuses on development of joint optimal maintenance and production policy for a specific type of production system that allows for adjustable production rates. The rate of deterioration of the system is directly related to the production rate, with higher production rates resulting in greater expected deterioration. The system's deterioration can be controlled through two main actions: (1) scheduling and conducting maintenance actions referred to as maintenance policy; and (2) adjusting the production rate referred to as production policy. To determine the optimal actions given the system's state, a Markov decision process (MDP) is developed and a reinforcement learning algorithm, specifically a Q-learning algorithm, is utilized. The algorithm's hyper parameters are tuned using a value-iteration algorithm of dynamic programming. The goal is to minimize expected costs for the system over a finite planning horizon.","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116632725","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：