Pub Date : 2023-06-05DOI: 10.1109/ICPHM57936.2023.10194126
Jiahui Wang, Lin Ma, Ankang Chen, Qiannan Liu, M. Ma
The level and speed of performance degradation after maintenance will be affected by the maintenance effect. Aiming at the degradation process of repairable equipment, a fault state prediction model considering maintenance effect was established to obtain the state transfer process of repaired equipment within $n$ detection cycles. Firstly, according to the maintenance effect of the equipment, the regression model of degradation degree and the regression rate model are proposed. Secondly, considering the unobservability of equipment performance degradation state, based on hidden Markov model, and on the basis of state division of equipment degradation quantity, the required state space and observation space of the model are constructed, and finally the fault state prediction model considering maintenance effect is established. The case takes the temperature state of the repaired circuit breaker as the observable variable of the HMM model. The calculated results of the model are closer to the real situation, which shows the feasibility of the model and can be applied in the field of maintenance optimization.
{"title":"Fault State Prediction Model of Repaired Equipment Considering Maintenance Effect","authors":"Jiahui Wang, Lin Ma, Ankang Chen, Qiannan Liu, M. Ma","doi":"10.1109/ICPHM57936.2023.10194126","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194126","url":null,"abstract":"The level and speed of performance degradation after maintenance will be affected by the maintenance effect. Aiming at the degradation process of repairable equipment, a fault state prediction model considering maintenance effect was established to obtain the state transfer process of repaired equipment within $n$ detection cycles. Firstly, according to the maintenance effect of the equipment, the regression model of degradation degree and the regression rate model are proposed. Secondly, considering the unobservability of equipment performance degradation state, based on hidden Markov model, and on the basis of state division of equipment degradation quantity, the required state space and observation space of the model are constructed, and finally the fault state prediction model considering maintenance effect is established. The case takes the temperature state of the repaired circuit breaker as the observable variable of the HMM model. The calculated results of the model are closer to the real situation, which shows the feasibility of the model and can be applied in the field of maintenance optimization.","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":"115494377","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/ICPHM57936.2023.10194168
Ricardo Ludeke, P. S. Heyns
Industrial operational environments are stochastic and can have complex system dynamics which introduce multiple levels of uncertainty. This uncertainty can lead to sub-optimal decision making and resource allocation. Digitalization and automation of production equipment and the maintenance environment enable predictive maintenance, which means that equipment can be stopped for maintenance at the optimal time instant. Resource constraints in maintenance capacity could however result in further undesired downtime if maintenance cannot be performed when scheduled. In this paper the use of a multi-agent deep reinforcement learning based approach for decision making is investigated to determine the optimal maintenance scheduling policy for a fleet of assets where there are maintenance resource constraints. By considering the underlying system dynamics of maintenance capacity, as well as the health state of individual assets, a near-optimal decision making policy is found that increases equipment availability while also maximizing maintenance capacity. The proposed solution is compared to a run-to-failure corrective maintenance strategy, a constant interval preventive maintenance strategy and a condition based predictive maintenance strategy. The proposed approach outperformed traditional maintenance strategies across several asset and operational maintenance performance metrics. It is concluded that deep reinforcement learning based decision making for asset health management and resource allocation is more effective than human based decision making.
{"title":"Towards a Deep Reinforcement Learning based approach for real time decision making and resource allocation for Prognostics and Health Management applications","authors":"Ricardo Ludeke, P. S. Heyns","doi":"10.1109/ICPHM57936.2023.10194168","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194168","url":null,"abstract":"Industrial operational environments are stochastic and can have complex system dynamics which introduce multiple levels of uncertainty. This uncertainty can lead to sub-optimal decision making and resource allocation. Digitalization and automation of production equipment and the maintenance environment enable predictive maintenance, which means that equipment can be stopped for maintenance at the optimal time instant. Resource constraints in maintenance capacity could however result in further undesired downtime if maintenance cannot be performed when scheduled. In this paper the use of a multi-agent deep reinforcement learning based approach for decision making is investigated to determine the optimal maintenance scheduling policy for a fleet of assets where there are maintenance resource constraints. By considering the underlying system dynamics of maintenance capacity, as well as the health state of individual assets, a near-optimal decision making policy is found that increases equipment availability while also maximizing maintenance capacity. The proposed solution is compared to a run-to-failure corrective maintenance strategy, a constant interval preventive maintenance strategy and a condition based predictive maintenance strategy. The proposed approach outperformed traditional maintenance strategies across several asset and operational maintenance performance metrics. It is concluded that deep reinforcement learning based decision making for asset health management and resource allocation is more effective than human based decision making.","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"45 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":"122746957","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/ICPHM57936.2023.10193908
Mengqi Miao, Jianbo Yu, Pu Yang, Shang Yue, Ruixu Zhou
Lithium-ion battery health monitoring is crucial in ensuring the reliability of the power system. Due to complex and dynamic battery operating conditions (e.g., ambient temperature and discharge current), domain shift is an ineluctable issue in battery health monitoring. In this study, a novel transfer learning (TL) method, i.e., selective domain adaptation network (SDANet) is developed for solving the problem of domain shift and performing battery health monitoring. Firstly, an unsupervised domain selection mechanism is established to select the optimal source domain, so as to minimize negative transfer in TL. Then, an adaptive feature transmission mechanism (AFTM) is proposed to improve gradient propagation and the performance of feature learning. Thirdly, the selective domain adaptation method is carried out according to channel similarity, which effectively solves the problem of domain shift and improves the performance of battery health estimation. The experiment results demonstrate that SDANet has excellent battery health monitoring performance under various working conditions.
{"title":"Selective Domain Adaptation Network for Lithium-ion Battery Health Monitoring","authors":"Mengqi Miao, Jianbo Yu, Pu Yang, Shang Yue, Ruixu Zhou","doi":"10.1109/ICPHM57936.2023.10193908","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10193908","url":null,"abstract":"Lithium-ion battery health monitoring is crucial in ensuring the reliability of the power system. Due to complex and dynamic battery operating conditions (e.g., ambient temperature and discharge current), domain shift is an ineluctable issue in battery health monitoring. In this study, a novel transfer learning (TL) method, i.e., selective domain adaptation network (SDANet) is developed for solving the problem of domain shift and performing battery health monitoring. Firstly, an unsupervised domain selection mechanism is established to select the optimal source domain, so as to minimize negative transfer in TL. Then, an adaptive feature transmission mechanism (AFTM) is proposed to improve gradient propagation and the performance of feature learning. Thirdly, the selective domain adaptation method is carried out according to channel similarity, which effectively solves the problem of domain shift and improves the performance of battery health estimation. The experiment results demonstrate that SDANet has excellent battery health monitoring performance under various working conditions.","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":"129570576","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/ICPHM57936.2023.10194151
Lucas Dimitri, J. Liscouët
The use of unmanned aerial vehicles (UAVs) is rapidly expanding across numerous industries, with a diverse range of applications. Ensuring reliable operation is crucial for safety, costs, and customer satisfaction, especially in the aviation sector. This paper presents a novel approach to optimizing flight control by incorporating a reliability-based control allocation system with physics-based reliability models. More specifically, the control allocation is based on physical estimations of reliability parameters. The reliability model incorporates a Weibull distribution reformulated to express reliability as a function of cumulated damage instead of time. The failure mechanisms of the rotor components are modeled based on physics, allowing for the calculation of cumulated damages as a function of the UAV's operation. The parameterization of the reliability and failure mechanism models is entirely based on publicly available manufacturer catalog data to ensure that the models are readily applicable to new designs with off-the-shelf components. Additionally, this approach facilitates the verification and validation of the models. The developed integrated control strategy and physics-based models have been implemented in Matlab-Simulink and applied to the case study of a coaxial quadrotor UAV for validation. When applied to the case study, the controller efficiently redistributes the control duties of rotors with a high probability of failure while maintaining the desired system response, thus increasing the operation's reliability.
{"title":"Optimizing Flight Control of Unmanned Aerial Vehicles with Physics-Based Reliability Models","authors":"Lucas Dimitri, J. Liscouët","doi":"10.1109/ICPHM57936.2023.10194151","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194151","url":null,"abstract":"The use of unmanned aerial vehicles (UAVs) is rapidly expanding across numerous industries, with a diverse range of applications. Ensuring reliable operation is crucial for safety, costs, and customer satisfaction, especially in the aviation sector. This paper presents a novel approach to optimizing flight control by incorporating a reliability-based control allocation system with physics-based reliability models. More specifically, the control allocation is based on physical estimations of reliability parameters. The reliability model incorporates a Weibull distribution reformulated to express reliability as a function of cumulated damage instead of time. The failure mechanisms of the rotor components are modeled based on physics, allowing for the calculation of cumulated damages as a function of the UAV's operation. The parameterization of the reliability and failure mechanism models is entirely based on publicly available manufacturer catalog data to ensure that the models are readily applicable to new designs with off-the-shelf components. Additionally, this approach facilitates the verification and validation of the models. The developed integrated control strategy and physics-based models have been implemented in Matlab-Simulink and applied to the case study of a coaxial quadrotor UAV for validation. When applied to the case study, the controller efficiently redistributes the control duties of rotors with a high probability of failure while maintaining the desired system response, thus increasing the operation's reliability.","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":"114347499","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/ICPHM57936.2023.10194181
Na Zhang, Li-xiang Duan, Xiaofeng Li, Xiangwu Liu
In the case of sufficient and balanced training data, the intelligent diagnosis models can accurately determine the state of the planetary gearbox and play a significant role in ensuring its healthy operation. However, the planetary gearbox operates normally for much longer than the moment of failure in practical engineering, which makes the sample size of fault state extremely small and the training data imbalanced. As a result, the model fail to detect the extremely small samples and thus serious fault missed diagnosis. In order to improve the performance of imbalanced diagnosis of planetary gearboxes with containing extremely small samples, this paper proposed an imbalanced fault diagnosis method for planetary gearboxes based on noise enhancement and threshold adaptive Siamese decoupled network. Firstly, the extremely-samples are enhanced into small samples by adding noise appropriately, and a set of metrics are proposed to evaluate the quality of the enhanced samples. Then, the Siamese network is constructed, and the special input requirements of the Siamese network are used to expand the training data again, which solves the problems of poor generalization and missed diagnosis caused by small samples and imbalance. Finally, a threshold adaptive and multi-scale decoupled convolution is proposed to improve the Siamese network and further improve the diagnostic performance. It is verified by imbalanced planetary gearbox data. On the imbalanced training data with small samples, the diagnostic accuracy of the proposed method was up to 98.33 %. In the extreme cases of high imbalance (fault / total < 5%) and small sample size of fault (only 3 samples per class), the diagnostic accuracy still reached 71.11 %. It shows that the proposed method has great advantages and potential in imbalanced fault diagnosis with small samples.
{"title":"Imbalanced fault diagnosis of planetary gearboxes based on noise enhancement and threshold adaptive Siamese decoupled network","authors":"Na Zhang, Li-xiang Duan, Xiaofeng Li, Xiangwu Liu","doi":"10.1109/ICPHM57936.2023.10194181","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194181","url":null,"abstract":"In the case of sufficient and balanced training data, the intelligent diagnosis models can accurately determine the state of the planetary gearbox and play a significant role in ensuring its healthy operation. However, the planetary gearbox operates normally for much longer than the moment of failure in practical engineering, which makes the sample size of fault state extremely small and the training data imbalanced. As a result, the model fail to detect the extremely small samples and thus serious fault missed diagnosis. In order to improve the performance of imbalanced diagnosis of planetary gearboxes with containing extremely small samples, this paper proposed an imbalanced fault diagnosis method for planetary gearboxes based on noise enhancement and threshold adaptive Siamese decoupled network. Firstly, the extremely-samples are enhanced into small samples by adding noise appropriately, and a set of metrics are proposed to evaluate the quality of the enhanced samples. Then, the Siamese network is constructed, and the special input requirements of the Siamese network are used to expand the training data again, which solves the problems of poor generalization and missed diagnosis caused by small samples and imbalance. Finally, a threshold adaptive and multi-scale decoupled convolution is proposed to improve the Siamese network and further improve the diagnostic performance. It is verified by imbalanced planetary gearbox data. On the imbalanced training data with small samples, the diagnostic accuracy of the proposed method was up to 98.33 %. In the extreme cases of high imbalance (fault / total < 5%) and small sample size of fault (only 3 samples per class), the diagnostic accuracy still reached 71.11 %. It shows that the proposed method has great advantages and potential in imbalanced fault diagnosis with small samples.","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"73 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":"124638472","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/ICPHM57936.2023.10194172
Fabian Fingerhut, Sarah Klein, Mathias Verbeke, Sreeraj Rajendran, E. Tsiporkova
Nowadays, most industrial assets are equipped with a multitude of different sensors continuously examining the asset's status and health. For a reliable estimation of an asset's performance it is crucial though to consider that most assets are exposed to different and typically varying contexts during their operations. These contexts are defined by both internal and external factors and complicate the task of asset condition monitoring and profiling. In this article, an unsupervised approach for asset performance profiling is proposed based on multi-view representation and matrix decomposition. It enables one to derive specific fingerprints characterising asset performance behaviour in a context-sensitive fashion. The data is processed in two separate data views: 1) the process view, in which variables related to the asset's internal working are processed and partitioned such that each measurement point is associated with a specific label representing the context; and 2) the vibration view, where vibration profiles are extracted via non-negative matrix decomposition. Subsequently, the two views are linked together allowing to derive characteristic fingerprints using a suitable contextual representation and performance-related indicators. The proposed methodology is validated on a real-world industrial data set, consisting of vibration and operational sensor measurements of feedwater pumps. The obtained results illustrate that the profiling methodology is able to deliver a meaningful risk assessment estimation associated to different operating contexts.
{"title":"Multi-view contextual performance profiling in rotating machinery","authors":"Fabian Fingerhut, Sarah Klein, Mathias Verbeke, Sreeraj Rajendran, E. Tsiporkova","doi":"10.1109/ICPHM57936.2023.10194172","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194172","url":null,"abstract":"Nowadays, most industrial assets are equipped with a multitude of different sensors continuously examining the asset's status and health. For a reliable estimation of an asset's performance it is crucial though to consider that most assets are exposed to different and typically varying contexts during their operations. These contexts are defined by both internal and external factors and complicate the task of asset condition monitoring and profiling. In this article, an unsupervised approach for asset performance profiling is proposed based on multi-view representation and matrix decomposition. It enables one to derive specific fingerprints characterising asset performance behaviour in a context-sensitive fashion. The data is processed in two separate data views: 1) the process view, in which variables related to the asset's internal working are processed and partitioned such that each measurement point is associated with a specific label representing the context; and 2) the vibration view, where vibration profiles are extracted via non-negative matrix decomposition. Subsequently, the two views are linked together allowing to derive characteristic fingerprints using a suitable contextual representation and performance-related indicators. The proposed methodology is validated on a real-world industrial data set, consisting of vibration and operational sensor measurements of feedwater pumps. The obtained results illustrate that the profiling methodology is able to deliver a meaningful risk assessment estimation associated to different operating contexts.","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":"121017923","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/ICPHM57936.2023.10194132
Aungshula Chowdhury, M. Lipsett
As Uncrewed Aerial Vehicle systems (UAVs) become more common and useful in public airspaces, this technology and its operation must be highly reliable to reduce risk to the general public. The objective of the present work is to improve the chances of mission success by analyzing and controlling the risk of UAV missions during different operational phases. Given the lack of reliability models for UAVs, we employ a systems reliability modeling methodology based on task decomposition and conditional risk analysis of each activity during a mission. The various risks involved in a specific mission activity are identified using Hazop techniques and Failure Modes and Effects Analysis (FMEA), along with the stopping conditions necessary to limit risks to an acceptable level. Different parts of a mission have different risk priorities, and the internal and external causes of failures of each activity are identified, described, and ranked according to their impact and uncertainties. This work constitutes the first phase of a broader research project. The risks of the UAV mission are modeled, after which it is verified by subject matter experts prior to implementing controls in an industrial case study.
{"title":"Modeling Operational Risk to Improve Reliability of Unmanned Aerial Vehicles","authors":"Aungshula Chowdhury, M. Lipsett","doi":"10.1109/ICPHM57936.2023.10194132","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194132","url":null,"abstract":"As Uncrewed Aerial Vehicle systems (UAVs) become more common and useful in public airspaces, this technology and its operation must be highly reliable to reduce risk to the general public. The objective of the present work is to improve the chances of mission success by analyzing and controlling the risk of UAV missions during different operational phases. Given the lack of reliability models for UAVs, we employ a systems reliability modeling methodology based on task decomposition and conditional risk analysis of each activity during a mission. The various risks involved in a specific mission activity are identified using Hazop techniques and Failure Modes and Effects Analysis (FMEA), along with the stopping conditions necessary to limit risks to an acceptable level. Different parts of a mission have different risk priorities, and the internal and external causes of failures of each activity are identified, described, and ranked according to their impact and uncertainties. This work constitutes the first phase of a broader research project. The risks of the UAV mission are modeled, after which it is verified by subject matter experts prior to implementing controls in an industrial case study.","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"19 7","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"113975150","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/ICPHM57936.2023.10193965
T. Ravichandran, Yuan Liu, Amar Kumar, A. Srivastava
In this work, a data-driven and deep learning-based predictive modeling framework has been developed for generating accurate prediction models intended for gas turbine engine performance analysis. This paper focuses on the application of Convolutional Neural Networks (CNNs) along with tabular data to image conversion techniques to predict exhaust gas temperature (EGT) and power outputs of Gas Turbine Engines (GTE). Using one such tabular data to image conversion method called Image Generator for Tabular Data (IGTD), several CNN model architectures were explored, and their predictive capabilities were compared. The effectiveness of the proposed predictive modeling framework which combines CNNs and the IGTD algorithm has been demonstrated for EGT and power prediction using GTE operational data collected over a period of three years. The CNN models using images converted from tabular data exhibit superior predictive capabilities for both EGT and power, with a more significant improvement observed for EGT prediction. To the best of our knowledge, this is the first attempt to apply IGTD based CNNs for developing GTE models for EGT and power prediction.
在这项工作中，开发了一个基于数据驱动和深度学习的预测建模框架，用于生成用于燃气涡轮发动机性能分析的准确预测模型。本文主要研究了卷积神经网络(cnn)与表格数据在图像转换技术中的应用，以预测燃气涡轮发动机(GTE)的排气温度(EGT)和功率输出。利用一种名为image Generator for tabular data (IGTD)的表格数据到图像的转换方法，探讨了几种CNN模型架构，并比较了它们的预测能力。结合cnn和IGTD算法所提出的预测建模框架的有效性已被证明用于使用收集超过三年的GTE运行数据进行EGT和功率预测。使用从表格数据转换而来的图像的CNN模型对EGT和功率都表现出优越的预测能力，对EGT的预测有更显著的改进。据我们所知，这是第一次尝试应用基于IGTD的cnn来开发用于EGT和功率预测的GTE模型。
{"title":"Convolutional Neural Networks for Gas Turbine Exhaust Gas Temperature and Power Predictions","authors":"T. Ravichandran, Yuan Liu, Amar Kumar, A. Srivastava","doi":"10.1109/ICPHM57936.2023.10193965","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10193965","url":null,"abstract":"In this work, a data-driven and deep learning-based predictive modeling framework has been developed for generating accurate prediction models intended for gas turbine engine performance analysis. This paper focuses on the application of Convolutional Neural Networks (CNNs) along with tabular data to image conversion techniques to predict exhaust gas temperature (EGT) and power outputs of Gas Turbine Engines (GTE). Using one such tabular data to image conversion method called Image Generator for Tabular Data (IGTD), several CNN model architectures were explored, and their predictive capabilities were compared. The effectiveness of the proposed predictive modeling framework which combines CNNs and the IGTD algorithm has been demonstrated for EGT and power prediction using GTE operational data collected over a period of three years. The CNN models using images converted from tabular data exhibit superior predictive capabilities for both EGT and power, with a more significant improvement observed for EGT prediction. To the best of our knowledge, this is the first attempt to apply IGTD based CNNs for developing GTE models for EGT and power prediction.","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"19 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":"131256066","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/ICPHM57936.2023.10194114
Y. Koch, M. Fett, E. Kirchner
This study examines the angle-specific tooth signature of a gear wheel and describes the possible use for monitoring the gearbox condition. The tests were conducted on a gearbox test stand with two 1-stage gearboxes equipped with a magnetoresistive sensor, which measures the angle using the gear wheel as a material measure. This new usage of a gear wheel and its specific properties are described. The gearboxes were driven by 30 kW asynchronous motors and tested in four quadrants with varying speed and torque conditions. The raw signals were processed using Matlab®. The sensor concept generates a sine-like signal for each tooth of the gear wheel. The high and low peaks of the sine-like wave were extracted and compared to analyze the reproducibility of the angle-specific tooth signature. The results show that the high and low peaks at one tooth of one sensor have a standard deviation of 0.0014 V over 10 revolutions and at different operating conditions, demonstrating the reproducibility of the angle-specific tooth signature. To further utilize this signature, a procedure is presented to identify an absolute reference with these signals, and the potential usage of the angle-specific tooth signature for absolute position detection is described.
{"title":"Angular measurement with a gear wheel as a material measure - Extension as absolute sensor","authors":"Y. Koch, M. Fett, E. Kirchner","doi":"10.1109/ICPHM57936.2023.10194114","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194114","url":null,"abstract":"This study examines the angle-specific tooth signature of a gear wheel and describes the possible use for monitoring the gearbox condition. The tests were conducted on a gearbox test stand with two 1-stage gearboxes equipped with a magnetoresistive sensor, which measures the angle using the gear wheel as a material measure. This new usage of a gear wheel and its specific properties are described. The gearboxes were driven by 30 kW asynchronous motors and tested in four quadrants with varying speed and torque conditions. The raw signals were processed using Matlab®. The sensor concept generates a sine-like signal for each tooth of the gear wheel. The high and low peaks of the sine-like wave were extracted and compared to analyze the reproducibility of the angle-specific tooth signature. The results show that the high and low peaks at one tooth of one sensor have a standard deviation of 0.0014 V over 10 revolutions and at different operating conditions, demonstrating the reproducibility of the angle-specific tooth signature. To further utilize this signature, a procedure is presented to identify an absolute reference with these signals, and the potential usage of the angle-specific tooth signature for absolute position detection is described.","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"165 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":"122896783","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/ICPHM57936.2023.10194194
P. Kakosimos
The electrification of powertrains is rising as the objective for a more viable future is intensified. To ensure continuous and reliable operation without undesirable malfunctions, it is essential to monitor the internal temperatures of machines and keep them within safe operating limits. Conventional modeling methods can be complex and usually require expert knowledge. With the amount of data collected these days, it is possible to use information models to assess thermal behaviors. This paper investigates artificial intelligence techniques for monitoring the cooling efficiency of induction machines. Experimental data was collected under specific operating conditions, and three machine-learning models have been developed. The optimal configuration for each approach was determined through rigorous hyperparameter searches, and the models were evaluated using a variety of metrics. The three solutions performed well in monitoring the condition of the machine even under transient operation, highlighting the potential of data-driven methods in improving the thermal management.
{"title":"Reliable Thermal Monitoring of Electric Machines through Machine Learning","authors":"P. Kakosimos","doi":"10.1109/ICPHM57936.2023.10194194","DOIUrl":"https://doi.org/10.1109/ICPHM57936.2023.10194194","url":null,"abstract":"The electrification of powertrains is rising as the objective for a more viable future is intensified. To ensure continuous and reliable operation without undesirable malfunctions, it is essential to monitor the internal temperatures of machines and keep them within safe operating limits. Conventional modeling methods can be complex and usually require expert knowledge. With the amount of data collected these days, it is possible to use information models to assess thermal behaviors. This paper investigates artificial intelligence techniques for monitoring the cooling efficiency of induction machines. Experimental data was collected under specific operating conditions, and three machine-learning models have been developed. The optimal configuration for each approach was determined through rigorous hyperparameter searches, and the models were evaluated using a variety of metrics. The three solutions performed well in monitoring the condition of the machine even under transient operation, highlighting the potential of data-driven methods in improving the thermal management.","PeriodicalId":169274,"journal":{"name":"2023 IEEE International Conference on Prognostics and Health Management (ICPHM)","volume":"16 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":"121929002","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：