Pub Date : 2022-06-29DOI: 10.36001/phme.2022.v7i1.3359
Martin Hervé de Beaulieu, Mayank Shekhar Jha, H. Garnier, Farid Cerbah
Prediction of the Remaining Useful Life (RUL) for industrial systems has been facilitated by the acquisition of large amounts of real-time data and the use of deep learning methods. However, the vast majority of these methods rely on the availability of extensive RUL-labeled data, which is not the case for most of real industrial applications. The goal of this paper is to show how unsupervised learning can provide alternative ways to address this issue. The proposed method is essentially made of two steps. First, a Virtual Health Index (VHI) is extracted in an unsupervised manner from the raw sensor data using a Deep Convolutional Neural Network (CNN) autoencoder. Secondly, an Long-Short Term Memory (LSTM) Encoder-Decoder predicts the future values of the VHI, until an End-of-Life (EOL) pattern is recognized (using a sliding window DTW algorithm). The suggested method is tested on the C-MAPSS dataset and offers promising results with a great potential to be applicable on real-life use cases.
{"title":"Unsupervised Prognostics based on Deep Virtual Health Index Prediction","authors":"Martin Hervé de Beaulieu, Mayank Shekhar Jha, H. Garnier, Farid Cerbah","doi":"10.36001/phme.2022.v7i1.3359","DOIUrl":"https://doi.org/10.36001/phme.2022.v7i1.3359","url":null,"abstract":"Prediction of the Remaining Useful Life (RUL) for industrial systems has been facilitated by the acquisition of large amounts of real-time data and the use of deep learning methods. However, the vast majority of these methods rely on the availability of extensive RUL-labeled data, which is not the case for most of real industrial applications. The goal of this paper is to show how unsupervised learning can provide alternative ways to address this issue. The proposed method is essentially made of two steps. First, a Virtual Health Index (VHI) is extracted in an unsupervised manner from the raw sensor data using a Deep Convolutional Neural Network (CNN) autoencoder. Secondly, an Long-Short Term Memory (LSTM) Encoder-Decoder predicts the future values of the VHI, until an End-of-Life (EOL) pattern is recognized (using a sliding window DTW algorithm). The suggested method is tested on the C-MAPSS dataset and offers promising results with a great potential to be applicable on real-life use cases.","PeriodicalId":422825,"journal":{"name":"PHM Society European Conference","volume":"4 6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123679065","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 : 2022-06-29DOI: 10.36001/phme.2022.v7i1.3370
A. Gaffet, Nathalie Barbosa Roa, P. Ribot, E. Chanthery, C. Merle
This paper presents XGBoost classifier-based methods to solve three tasks proposed by the European Prognostics and Health Management Society (PHME) 2022 conference. These tasks are based on real data from a Surface Mount Technologies line. Each of these tasks aims to improve the efficiency of the Printed Circuit Board (PCB) manufacturing process, facilitate the operator’s work and minimize the cases of manual intervention. Due to the structured nature of the problems proposed for each task, an XGBoost method based on encoding and feature engineering is proposed. The proposed methods utilise the fusion of test values and system characteristics extracted from two different testing equipment of the Surface Mount Technologies lines. This work also explores the problems of generalising prediction at the system level using information from the subsystem data. For this particular industrial case: the challenges with the changes in the number of subsystems. For Industry 4.0, the need for interpretability is very important. This is why the results of the models are analysed using Shapley values. With the proposed method, our team took the first place, capable of successfully detecting at an early stage the defective components for tasks 2 and 3.
本文提出了基于XGBoost分类器的方法来解决欧洲预后和健康管理学会(PHME) 2022年会议提出的三个任务。这些任务都是基于来自Surface Mount Technologies生产线的真实数据。这些任务中的每一项都旨在提高印刷电路板(PCB)制造过程的效率，方便操作员的工作，并最大限度地减少人工干预的情况。由于每个任务所提出问题的结构化性质，提出了一种基于编码和特征工程的XGBoost方法。所提出的方法利用了从表面贴装技术生产线的两个不同测试设备中提取的测试值和系统特性的融合。这项工作还探讨了利用子系统数据中的信息在系统级推广预测的问题。对于这个特殊的工业案例:子系统数量变化带来的挑战。对于工业4.0，对可解释性的需求非常重要。这就是为什么使用Shapley值来分析模型的结果。通过提出的方法，我们的团队获得了第一名，能够在任务2和3的早期阶段成功地检测到有缺陷的组件。
{"title":"A Hierarchical XGBoost Early Detection Method for Quality and Productivity Improvement of Electronics Manufacturing Systems","authors":"A. Gaffet, Nathalie Barbosa Roa, P. Ribot, E. Chanthery, C. Merle","doi":"10.36001/phme.2022.v7i1.3370","DOIUrl":"https://doi.org/10.36001/phme.2022.v7i1.3370","url":null,"abstract":"This paper presents XGBoost classifier-based methods to solve three tasks proposed by the European Prognostics and Health Management Society (PHME) 2022 conference. These tasks are based on real data from a Surface Mount Technologies line. Each of these tasks aims to improve the efficiency of the Printed Circuit Board (PCB) manufacturing process, facilitate the operator’s work and minimize the cases of manual intervention. Due to the structured nature of the problems proposed for each task, an XGBoost method based on encoding and feature engineering is proposed. The proposed methods utilise the fusion of test values and system characteristics extracted from two different testing equipment of the Surface Mount Technologies lines. This work also explores the problems of generalising prediction at the system level using information from the subsystem data. For this particular industrial case: the challenges with the changes in the number of subsystems. For Industry 4.0, the need for interpretability is very important. This is why the results of the models are analysed using Shapley values. With the proposed method, our team took the first place, capable of successfully detecting at an early stage the defective components for tasks 2 and 3.","PeriodicalId":422825,"journal":{"name":"PHM Society European Conference","volume":"353 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122755569","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 : 2022-06-29DOI: 10.36001/phme.2022.v7i1.3353
Andrea De Martin, G. Jacazio, Vincenzo Parisi, M. Sorli
The evolution towards “more electric” aircrafts has seen a decisive push in the last decade, due to the growing environmental concerns and the development of new market segments (flying taxis). Such push interested both the propulsion components and the aircraft systems, with the latter seeing a progressive trend in replacing the traditional solutions based on hydraulic power with electrical or electromechanical devices. Although more attention is usually devised towards the flight control actuation, an interesting and fast-developing application field for electro-mechanical systems is that of the aeronautical brakes. Electro-mechanical brakes, or E-Brakes hereby onwards, would present several advantages over their hydraulic counterparts, mainly related to the avoidance of leakage issues and the simplification of the system architecture. The more difficult heat dissipation, associated with the thermal issues that usually constitute one of the most significant sizing constraints for electromechanical actuators, limits so far, their application (or proposal of application) to light-weight vehicles. Within this context, the development of PHM solutions would align with the need for an on-line monitoring of a relatively unproven component. This paper deals with the preliminary stages of the development of such PHM system for an E-Brake to be employed on a future executive class aircraft, where the brake is actuated through four electro-mechanical actuators. Since literature on fault diagnosis and prognosis for electrical motors is fairly extensive, we focused this preliminary analysis on the development of PHM techniques suitable to monitor and prognose the evolution of the brake pads wear instead. The paper opens detailing the system architecture and continues presenting the high-fidelity dynamic model used to build synthetic data-sets representative of the possible operating conditions faced by the E-Brake within realistic operative scenarios. Such data are then used to foster a preliminary feature selection process, where physics-based indexes are compared and evaluated. Simulated degradation histories are then used to test the application of data-driven fault detection algorithm and the possible application of particle-filtering routines for prognosis.
{"title":"Prognosis of Wear Progression in Electrical Brakes for Aeronautical Applications","authors":"Andrea De Martin, G. Jacazio, Vincenzo Parisi, M. Sorli","doi":"10.36001/phme.2022.v7i1.3353","DOIUrl":"https://doi.org/10.36001/phme.2022.v7i1.3353","url":null,"abstract":"The evolution towards “more electric” aircrafts has seen a decisive push in the last decade, due to the growing environmental concerns and the development of new market segments (flying taxis). Such push interested both the propulsion components and the aircraft systems, with the latter seeing a progressive trend in replacing the traditional solutions based on hydraulic power with electrical or electromechanical devices. Although more attention is usually devised towards the flight control actuation, an interesting and fast-developing application field for electro-mechanical systems is that of the aeronautical brakes. Electro-mechanical brakes, or E-Brakes hereby onwards, would present several advantages over their hydraulic counterparts, mainly related to the avoidance of leakage issues and the simplification of the system architecture. The more difficult heat dissipation, associated with the thermal issues that usually constitute one of the most significant sizing constraints for electromechanical actuators, limits so far, their application (or proposal of application) to light-weight vehicles. Within this context, the development of PHM solutions would align with the need for an on-line monitoring of a relatively unproven component. This paper deals with the preliminary stages of the development of such PHM system for an E-Brake to be employed on a future executive class aircraft, where the brake is actuated through four electro-mechanical actuators. Since literature on fault diagnosis and prognosis for electrical motors is fairly extensive, we focused this preliminary analysis on the development of PHM techniques suitable to monitor and prognose the evolution of the brake pads wear instead. The paper opens detailing the system architecture and continues presenting the high-fidelity dynamic model used to build synthetic data-sets representative of the possible operating conditions faced by the E-Brake within realistic operative scenarios. Such data are then used to foster a preliminary feature selection process, where physics-based indexes are compared and evaluated. Simulated degradation histories are then used to test the application of data-driven fault detection algorithm and the possible application of particle-filtering routines for prognosis.","PeriodicalId":422825,"journal":{"name":"PHM Society European Conference","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122038725","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 : 2022-06-29DOI: 10.36001/phme.2022.v7i1.3333
F. Federici, C. Tonelli, M. Le Cam, Marcello Torchio, David Larsen
In recent years, Prognostic & Health Management (PHM) has become a topic of strong interest in the aerospace domain. Health assessment and remaining useful life estimation for on-board systems provide several advantages, mainly related to the increased analysis capabilities and the reduction of maintenance interventions (and, consequently, of operating costs). For this reason, it is of interest for the aerospace industry to identify and define efficient strategies both for the introduction of native PHM capabilities in new generation on-board systems and for the retrofit of existing ones. This paper proposes a strategy for the scalable deployment of PHM techniques for on-board systems, with particular focus on edge computing capabilities. Different reference scenarios (ranging from cloud-based processing to local-only processing) are presented, and an edge-focused PHM architecture is discussed in detail, with the relative challenges addressed. The design and validation of proposed edge-based solution is described, with specific reference to its support for an existing data analytics framework. The solution is then assessed against a reference aerospace use case involving a representative aircraft braking system, focusing on computational aspects to highlight the compatibility of the proposed deployment strategy with efficient on-board computations.
{"title":"Design and validation of scalable PHM solutions for aerospace onboard systems","authors":"F. Federici, C. Tonelli, M. Le Cam, Marcello Torchio, David Larsen","doi":"10.36001/phme.2022.v7i1.3333","DOIUrl":"https://doi.org/10.36001/phme.2022.v7i1.3333","url":null,"abstract":"In recent years, Prognostic & Health Management (PHM) has become a topic of strong interest in the aerospace domain. Health assessment and remaining useful life estimation for on-board systems provide several advantages, mainly related to the increased analysis capabilities and the reduction of maintenance interventions (and, consequently, of operating costs). For this reason, it is of interest for the aerospace industry to identify and define efficient strategies both for the introduction of native PHM capabilities in new generation on-board systems and for the retrofit of existing ones. This paper proposes a strategy for the scalable deployment of PHM techniques for on-board systems, with particular focus on edge computing capabilities. Different reference scenarios (ranging from cloud-based processing to local-only processing) are presented, and an edge-focused PHM architecture is discussed in detail, with the relative challenges addressed. The design and validation of proposed edge-based solution is described, with specific reference to its support for an existing data analytics framework. The solution is then assessed against a reference aerospace use case involving a representative aircraft braking system, focusing on computational aspects to highlight the compatibility of the proposed deployment strategy with efficient on-board computations.","PeriodicalId":422825,"journal":{"name":"PHM Society European Conference","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129885428","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 : 2022-06-29DOI: 10.36001/phme.2022.v7i1.3338
Khoury Boutrous, I. Bessa, V. Puig, F. Nejjari, R. Palhares
The increasing application of power converter systems based on semiconductor devices such as Insulated-Gate Bipolar Transistors (IGBTs) has motivated the investigation of strategies for their prognostics and health management. However, physicsbased degradation modelling for semiconductors is usually complex and depends on uncertain parameters, which motivates the use of data-driven approaches. This paper addresses the problem of data-driven prognostics of IGBTs based on evolving fuzzy models learned from degradation data streams. The model depends on two classes of degradation features: one group of features that are very sensitive to the degradation stages is used as a premise variable of the fuzzy model, and another group that provides good trendability and monotonicity is used for the auto-regressive consequent of the fuzzy model for degradation prediction. This strategy allows obtaining interpretable degradation models, which are improved when more degradation data is obtained from the Unit Under Test (UUT) in real time. Furthermore, the fuzzy-based Remaining Useful Life (RUL) prediction is equipped with an uncertainty quantification mechanism to better aid decisionmakers. The proposed approach is then used for the RUL prediction considering an accelerated aging IGBT dataset from the NASA Ames Research Center.
{"title":"Data-driven Prognostics based on Evolving Fuzzy Degradation Models for Power Semiconductor Devices","authors":"Khoury Boutrous, I. Bessa, V. Puig, F. Nejjari, R. Palhares","doi":"10.36001/phme.2022.v7i1.3338","DOIUrl":"https://doi.org/10.36001/phme.2022.v7i1.3338","url":null,"abstract":"The increasing application of power converter systems based on semiconductor devices such as Insulated-Gate Bipolar Transistors (IGBTs) has motivated the investigation of strategies for their prognostics and health management. However, physicsbased degradation modelling for semiconductors is usually complex and depends on uncertain parameters, which motivates the use of data-driven approaches. This paper addresses the problem of data-driven prognostics of IGBTs based on evolving fuzzy models learned from degradation data streams. The model depends on two classes of degradation features: one group of features that are very sensitive to the degradation stages is used as a premise variable of the fuzzy model, and another group that provides good trendability and monotonicity is used for the auto-regressive consequent of the fuzzy model for degradation prediction. This strategy allows obtaining interpretable degradation models, which are improved when more degradation data is obtained from the Unit Under Test (UUT) in real time. Furthermore, the fuzzy-based Remaining Useful Life (RUL) prediction is equipped with an uncertainty quantification mechanism to better aid decisionmakers. The proposed approach is then used for the RUL prediction considering an accelerated aging IGBT dataset from the NASA Ames Research Center.","PeriodicalId":422825,"journal":{"name":"PHM Society European Conference","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127891023","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 : 2022-06-29DOI: 10.36001/phme.2022.v7i1.3295
Washington Mhangami, S. King, Dave Barry
One area the aviation industry is grappling with is the quantification of the probability of occurrence of safety incidents. Currently, aviation professionals involved in safety risk management mostly rely on collective experience to determine probability of incident occurrences and apply it to the International Civil Aviation Organisation (ICAO) matrix or equivalent to evaluate the risk. A number of limitations linked to the use of risk matrices will be explored in this paper. It is the aim of this paper to explore statistical methods that can be used to determine the probability of safety occurrences and come up with an algorithm that can be used by airlines using available safety data. The novelty of this research is that it combines the exploration of use of statistical techniques to quantitatively assess risk using Flight Data Monitoring (FDM) and other data, with acceptability of Safety Risk Management (SRM) data analytics by operational personnel. The paper also explores the contributory factors leading to the reluctance of operational personnel to use data analytics to inform their risk assessments despite the increasing availability of operational data and advancement in technology.
{"title":"Application, Utility and Acceptability of Data Analytics in Safety Risk Management of Airline Operations","authors":"Washington Mhangami, S. King, Dave Barry","doi":"10.36001/phme.2022.v7i1.3295","DOIUrl":"https://doi.org/10.36001/phme.2022.v7i1.3295","url":null,"abstract":"\u0000One area the aviation industry is grappling with is the quantification of the probability of occurrence of safety incidents. Currently, aviation professionals involved in safety risk management mostly rely on collective experience to determine probability of incident occurrences and apply it to the International Civil Aviation Organisation (ICAO) matrix or equivalent to evaluate the risk. A number of limitations linked to the use of risk matrices will be explored in this paper. It is the aim of this paper to explore statistical methods that can be used to determine the probability of safety occurrences and come up with an algorithm that can be used by airlines using available safety data. The novelty of this research is that it combines the exploration of use of statistical techniques to quantitatively assess risk using Flight Data Monitoring (FDM) and other data, with acceptability of Safety Risk Management (SRM) data analytics by operational personnel. The paper also explores the contributory factors leading to the reluctance of operational personnel to use data analytics to inform their risk assessments despite the increasing availability of operational data and advancement in technology. \u0000","PeriodicalId":422825,"journal":{"name":"PHM Society European Conference","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133096748","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 : 2022-06-29DOI: 10.36001/phme.2022.v7i1.3343
Vishal Jadhav, A. Deodhar, Ashit Gupta, V. Runkana
Air Preheater (APH) is a regenerative heat exchanger employed in thermal power plants to save fuel by improving their thermal efficiency. Monitoring the health of APH vis-a-vis its fouling is critical because fouling often results in forced outages of the power plant, incurring huge revenue losses. APH fouling is a complex thermo-chemical phenomenon governed by flue gas composition, operating temperatures, fuel type and ambient conditions. Absence of sensors within the APH make it difficult to estimate the level of fouling and its progression even for an experienced operator. Attempts to estimate APH fouling in real-time via modeling are scarce. Here we present a physics-informed neural network (PINN) that tracks the health of an APH by real-time estimation of fouling conditions within the APH as a function of real-time sensor measurements. To account for multi-fluid operation in a multi-sector design of APH, the domain is decomposed into several sub-domains. PINN is applied to each sub-domain and the overall solution is ensured by applying continuity conditions at the sub-domain interfaces. The model predicts the interior temperatures and fouling zones within the APH using external sensor measurements such as air temperature and gas composition. The model predictions are consistent with physics and yet computationally efficient in run-time. The model does not need sensor data but can be improved further by accommodating available sensor data. The real-time predictions by the model improve operator’s visibility in fouling. The predictions can be used further for estimating the remaining useful cycle life of the APH, thereby avoiding forced outages. The model can easily be integrated with the digital twin of an APH for its predictive maintenance.
{"title":"Physics Informed Neural Network for Health Monitoring of an Air Preheater","authors":"Vishal Jadhav, A. Deodhar, Ashit Gupta, V. Runkana","doi":"10.36001/phme.2022.v7i1.3343","DOIUrl":"https://doi.org/10.36001/phme.2022.v7i1.3343","url":null,"abstract":"Air Preheater (APH) is a regenerative heat exchanger employed in thermal power plants to save fuel by improving their thermal efficiency. Monitoring the health of APH vis-a-vis its fouling is critical because fouling often results in forced outages of the power plant, incurring huge revenue losses. APH fouling is a complex thermo-chemical phenomenon governed by flue gas composition, operating temperatures, fuel type and ambient conditions. Absence of sensors within the APH make it difficult to estimate the level of fouling and its progression even for an experienced operator. Attempts to estimate APH fouling in real-time via modeling are scarce. Here we present a physics-informed neural network (PINN) that tracks the health of an APH by real-time estimation of fouling conditions within the APH as a function of real-time sensor measurements. To account for multi-fluid operation in a multi-sector design of APH, the domain is decomposed into several sub-domains. PINN is applied to each sub-domain and the overall solution is ensured by applying continuity conditions at the sub-domain interfaces. The model predicts the interior temperatures and fouling zones within the APH using external sensor measurements such as air temperature and gas composition. The model predictions are consistent with physics and yet computationally efficient in run-time. The model does not need sensor data but can be improved further by accommodating available sensor data. The real-time predictions by the model improve operator’s visibility in fouling. The predictions can be used further for estimating the remaining useful cycle life of the APH, thereby avoiding forced outages. The model can easily be integrated with the digital twin of an APH for its predictive maintenance.","PeriodicalId":422825,"journal":{"name":"PHM Society European Conference","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117322805","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 : 2022-06-29DOI: 10.36001/phme.2022.v7i1.3373
J. Taco, Prayag Gore, T. Minami, Pradeep Kundu, J. Lee
The demand for Printed circuit boards (PCBs) has increased due to the rapid change in technology in recent years. Consequently, PCBs health assessment and fault detection play an important role in improving productivity. This study proposed a novel method which focused on feature engineering for health assessment in PCBs. The performance of the proposed method has been validated using data obtained from PHM Europe 2022 data challenge. In this data challenge, PCBs health assessment needs to be performed with data from the Solder Paste Inspection (SPI) and the Automated Optical Inspection (AOI) machine. The challenge has three tasks: 1) Predict the labels of the AOI machine using the SPI data. 2) Using both the SPI and AOI machine data, predict the operator's verification that the AOI machine correctly detected a defect. 3) With the SPI and AOI data, predict the classification of the defective PCBs as either repairable or unrepairable. The component level features are extracted from the original SPI and AOI data which contain the pin level features to solve these tasks. Two machine learning-based classification models, i.e., Light Gradient Boosting Machine (LightGBM) and eXtreme Gradient Boosting (XGBoost), have been used for classification purposes. Training data given by the organizer was divided into 70% training and 30% validation. Based on the validation data, the highest F1-score was observed with LightGBM in Tasks 1 and 2, whereas, in Task 3, the highest F1-score was observed with the XGBoost model. Hence, the LightGBM model has been used in Tasks 1 and 2, and the XGBoost model was developed for Task 3.
{"title":"Novel Methodology for Health Assessment in Printed Circuit Boards","authors":"J. Taco, Prayag Gore, T. Minami, Pradeep Kundu, J. Lee","doi":"10.36001/phme.2022.v7i1.3373","DOIUrl":"https://doi.org/10.36001/phme.2022.v7i1.3373","url":null,"abstract":"The demand for Printed circuit boards (PCBs) has increased due to the rapid change in technology in recent years. Consequently, PCBs health assessment and fault detection play an important role in improving productivity. This study proposed a novel method which focused on feature engineering for health assessment in PCBs. The performance of the proposed method has been validated using data obtained from PHM Europe 2022 data challenge. In this data challenge, PCBs health assessment needs to be performed with data from the Solder Paste Inspection (SPI) and the Automated Optical Inspection (AOI) machine. The challenge has three tasks: 1) Predict the labels of the AOI machine using the SPI data. 2) Using both the SPI and AOI machine data, predict the operator's verification that the AOI machine correctly detected a defect. 3) With the SPI and AOI data, predict the classification of the defective PCBs as either repairable or unrepairable. The component level features are extracted from the original SPI and AOI data which contain the pin level features to solve these tasks. Two machine learning-based classification models, i.e., Light Gradient Boosting Machine (LightGBM) and eXtreme Gradient Boosting (XGBoost), have been used for classification purposes. Training data given by the organizer was divided into 70% training and 30% validation. Based on the validation data, the highest F1-score was observed with LightGBM in Tasks 1 and 2, whereas, in Task 3, the highest F1-score was observed with the XGBoost model. Hence, the LightGBM model has been used in Tasks 1 and 2, and the XGBoost model was developed for Task 3.","PeriodicalId":422825,"journal":{"name":"PHM Society European Conference","volume":"69 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116272645","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 : 2022-06-29DOI: 10.36001/phme.2022.v7i1.3352
Simon Hagmeyer, P. Zeiler, Marco F. Huber
In Prognostics and Health Management, there are three main approaches for implementing diagnostic and prognostic applications. These approaches are data-driven methods, physical model-based methods, and combinations of them, in the form of hybrid methods. Each of them has specific advantages but also limitations for their purposeful implementation. In the case of data-driven methods, one of the main limitations is the availability of sufficient training data that adequately cover the relevant state space. For model-based methods, on the other hand, it is often the case that the degradation process of the considered technical system is of significant complexity. In such a scenario physics-based modeling requires great effort or is not possible at all. Combinations of data-driven and model-based approaches in form of hybrid approaches offer the possibility to partially mitigate the shortcomings of the other two approaches, however, require a sufficiently detailed data-driven and physics-based model. This paper addresses the transitional field between data-driven and hybrid approaches. Despite the issues of formulating a physics-based model that provides a representation of the degradation process, basic knowledge of the considered system and of the laws governing its degradation process is usually available. Integration of such knowledge into a machine learning process is part of a research field that is either called theory-guided data science, (physics) informed machine learning, physics-based learning or physics guided machine learning. First, the state of research in Prognostics and Health Management on methods of this field is presented and existing research gaps are outlined. Then, a concept is introduced for incorporating fundamental knowledge, such as monotonicity constraints, into data-driven diagnostic and prognostic applications using approaches from theory-guided data science. A special aspect of this concept is its cross-application usability through the consideration of knowledge that repeatedly occurs in diagnostics and prognostics. This is, for example, knowledge about physically justified boundaries whose compliance makes a prediction of the data-driven model plausible in the first place.
{"title":"On the Integration of Fundamental Knowledge about Degradation Processes into Data-Driven Diagnostics and Prognostics Using Theory-Guided Data Science","authors":"Simon Hagmeyer, P. Zeiler, Marco F. Huber","doi":"10.36001/phme.2022.v7i1.3352","DOIUrl":"https://doi.org/10.36001/phme.2022.v7i1.3352","url":null,"abstract":"In Prognostics and Health Management, there are three main approaches for implementing diagnostic and prognostic applications. These approaches are data-driven methods, physical model-based methods, and combinations of them, in the form of hybrid methods. Each of them has specific advantages but also limitations for their purposeful implementation. In the case of data-driven methods, one of the main limitations is the availability of sufficient training data that adequately cover the relevant state space. For model-based methods, on the other hand, it is often the case that the degradation process of the considered technical system is of significant complexity. In such a scenario physics-based modeling requires great effort or is not possible at all. Combinations of data-driven and model-based approaches in form of hybrid approaches offer the possibility to partially mitigate the shortcomings of the other two approaches, however, require a sufficiently detailed data-driven and physics-based model.\u0000This paper addresses the transitional field between data-driven and hybrid approaches. Despite the issues of formulating a physics-based model that provides a representation of the degradation process, basic knowledge of the considered system and of the laws governing its degradation process is usually available. Integration of such knowledge into a machine learning process is part of a research field that is either called theory-guided data science, (physics) informed machine learning, physics-based learning or physics guided machine learning. First, the state of research in Prognostics and Health Management on methods of this field is presented and existing research gaps are outlined. Then, a concept is introduced for incorporating fundamental knowledge, such as monotonicity constraints, into data-driven diagnostic and prognostic applications using approaches from theory-guided data science. A special aspect of this concept is its cross-application usability through the consideration of knowledge that repeatedly occurs in diagnostics and prognostics. This is, for example, knowledge about physically justified boundaries whose compliance makes a prediction of the data-driven model plausible in the first place.","PeriodicalId":422825,"journal":{"name":"PHM Society European Conference","volume":"244 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116390782","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 : 2022-06-29DOI: 10.36001/phme.2022.v7i1.3334
Zahra Taghiyarrenani, A. Berenji
Noise will be unavoidably present in the data collected from physical environments, regardless of how sophisticated the measurement equipment is. Furthermore, collecting enough faulty data is a challenge since operating industrial machines in faulty modes not only has severe consequences to the machine health, but also may affect collateral machinery critically, from health state point of view. In this paper, we propose a method of denoising with limited data for the purpose of fault identification. In addition, our method is capable of removing multiple levels of noise simultaneously. For this purpose, inspired by unsupervised contrastive learning, we first augment the data with multiple levels of noise. Later, we construct a new feature representation using Contrastive Loss. The last step is building a classifier on top of the learned representation; this classifier can detect various faults in noisy environments. The experiments on the SOUTHEAST UNIVERSITY (SEU) dataset of bearings confirm that our method can simultaneously remove multiple noise levels.
{"title":"Noise-Robust Representation for Fault Identification with Limited Data via Data Augmentation","authors":"Zahra Taghiyarrenani, A. Berenji","doi":"10.36001/phme.2022.v7i1.3334","DOIUrl":"https://doi.org/10.36001/phme.2022.v7i1.3334","url":null,"abstract":"Noise will be unavoidably present in the data collected from physical environments, regardless of how sophisticated the measurement equipment is. Furthermore, collecting enough faulty data is a challenge since operating industrial machines in faulty modes not only has severe consequences to the machine health, but also may affect collateral machinery critically, from health state point of view. In this paper, we propose a method of denoising with limited data for the purpose of fault identification. In addition, our method is capable of removing multiple levels of noise simultaneously. For this purpose, inspired by unsupervised contrastive learning, we first augment the data with multiple levels of noise. Later, we construct a new feature representation using Contrastive Loss. The last step is building a classifier on top of the learned representation; this classifier can detect various faults in noisy environments. The experiments on the SOUTHEAST UNIVERSITY (SEU) dataset of bearings confirm that our method can simultaneously remove multiple noise levels.","PeriodicalId":422825,"journal":{"name":"PHM Society European Conference","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124405172","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：