Predictive Maintenance (PdM)

Top PDF Predictive Maintenance (PdM):

A new dynamic predictive maintenance framework using deep learning for failure prognostics

A new dynamic predictive maintenance framework using deep learning for failure prognostics

1. Introduction Due to the increasing requirement of reliability, availability, maintainability and safety of systems, the traditional maintenance strategies are becoming less effective and obsolete. Beside, the revolu- tion of Industry 4.0 provides more convenient supports for the wide development of the predictive maintenance (PdM) in practice. For ex- ample, the use of intelligent sensors provides a reliable solution for system monitoring in real time. Having this information, the manager can plan the maintenance activities more effectively to reduce machine downtimes and improve the production flow.
En savoir plus

13 En savoir plus

Towards Model Transformation from a CBM Model to CEP Rules to Support Predictive Maintenance

Towards Model Transformation from a CBM Model to CEP Rules to Support Predictive Maintenance

Keywords: Maintenance, Knowledge Base, Model Transformation. Abstract: Over the past decades, the development of predictive maintenance strategies, like Prognostics and Health Management (PHM), have brought new opportunities to the maintenance domain. However, implementing such systems addresses several challenges. First, all information related to the system description and failure definition must be collected and processed. In this regard, using an expert system (ES) seems interesting. The second challenge, when monitoring complex systems, is to deal with the high volume and velocity of the input data. To reduce them, Complex Event Processing (CEP) can be used to identify relevant events, based on predefined rules. These rules can be extracted from the ES knowledge base using model transformation. This process consists in transforming some concepts from a source to a target model using transformation rules. In this paper, we propose to transform a part of the knowledge from a condition-based maintenance (CBM) model into CEP rules. After further explaining the motivations behind this work and defining the principles behind model-driven architecture and model transformation, the transformation from a CBM model to a “generic rules” model will be proposed. This model will then be transformed into an Event Processing Language (EPL) model. Examples will be given as illustrations for each transformation.
En savoir plus

12 En savoir plus

Towards a hierarchical modelling approach for planning aircraft tail assignment and predictive maintenance

Towards a hierarchical modelling approach for planning aircraft tail assignment and predictive maintenance

ilyass.haloui@airbus.com - caroline.chanel@isae-supaero.fr - alain.hait@isae-supaero.fr Abstract Aircraft equipment health monitoring system plays a promis- ing role for airlines operation cost reduction, as it can be ex- ploited to perform predictive maintenance. In this vein, a hi- erarchical sequential decision making model is proposed to plan predictive maintenance. It combines linear optimization for routing assignment and MDP planning to handle mainte- nance actions based on the stochastic evolution of health in- dicators. This entangled model should reduce planning time while ensuring a cost-efficient policy.
En savoir plus

3 En savoir plus

Towards multi-model approaches to predictive maintenance: A systematic literature survey on diagnostics and prognostics

Towards multi-model approaches to predictive maintenance: A systematic literature survey on diagnostics and prognostics

• The lack of a systematic approach to design and develop predictive maintenance systems [ 108 , 170 ]. There exist standards, norms and generic architectures to develop new predictive maintenance sys- tems, such as OSA-CBM [ 15 ]. However, they only focus on the basic functional components of the system and do not cover important aspects regarding performance indicators or context constraints of the technical system. Besides, they do not offer yet a consistent explanation on which models to use depending on the initial needs of the predictive maintenance system. The lack of a systematic ap- proach limits the implementation of predictive maintenance systems on real scale industrial applications. When developing a new pre- dictive maintenance system the number of potential models to solve the problem is too high. For engineers the simple fact of choosing the right model or a reduced set of models remains a challenging task. It turns out to be very di fficult to perform an objective selec- tion of models as there are not enough comparative studies of the use of different models on the same tasks for predictive maintenance systems. None of the consulted publications in this survey gives extensive explanations for the selection of the proposed method and the architecture methodologies to create a concept of the system varies from one study to another. Besides, many studies do not present detailed design parameters for their proposed models, or the case study data is not available. All these aspects make it difficult to reproduce results and even more difficult to retrieve models from previous studies for use in new predictive maintenance systems. There are no clear guidelines for selecting the right model or models for a specific task given the operational modes and available data to perform diagnostics and prognostics.
En savoir plus

20 En savoir plus

Predictive maintenance from event logs using wavelet-based features: an industrial application

Predictive maintenance from event logs using wavelet-based features: an industrial application

Predictive maintenance from event logs using wavelet-based features: an industrial application Stéphane Bonnevay 1 Jairo Cugliari 1 Victoria Granger 2 1 ERIC EA3083, Université de Lyon, 5 av. Pierre Mendès France, 69676 Bron Cedex, France 2 ENEDIS, 124 boulevard Marius Vivier Merle, Lyon, France

14 En savoir plus

A Parametric Predictive Maintenance Decision-Making Framework Considering Improved System Health Prognosis Precision

A Parametric Predictive Maintenance Decision-Making Framework Considering Improved System Health Prognosis Precision

by a scalar random variable X t . In the absence of maintenance operation, X t evolves according to an increasing stochastic process {X t } t≥0 with X 0 = 0 (i.e., system new at t = 0). We also assume that the deterioration increment between times t and s (t ≤ s), X s − X t , is s-independent of deterioration levels before t. Under these assumptions, any monotone stochastic process belonging to L´evy family [48] can model the system deterioration. Hereinafter, a univariate homogeneous Gamma process with shape parameter α and scale parameter β is used. This choice is due to the three following reasons. Firstly, the Gamma deterioration process has been justified by diverse practical applications (e.g., fatigue crack growth [65], carbon-film resistors deterioration [66], corrosion damage mechanism [67], SiC MOSFET threshold voltage deterioration [68], actuator performance loss [69]) and considered appropriate by experts [70]. Secondly, using the homogeneous Gamma process can make the mathematical formulation feasible. And finally, we will see in the following that relying on such an univariate process allows a fair comparison on the performance and robustness of the two parametric predictive maintenance frameworks considered in this paper. Thus, for t ≤ s, the probability density function
En savoir plus

27 En savoir plus

Monitoring wear with integrated ultrasonic transducers for predictive maintenance in mining

Monitoring wear with integrated ultrasonic transducers for predictive maintenance in mining

ABSTRACT Maintenance and downtime costs can take a significant part of the cost of mining operation. Additionally, unplanned downtime could cost several times as much as scheduled downtime. Real time measurement of wear and corrosion can provide crucial information needed to determine an optimal schedule for maintenance to maximise equipment availability while ensuring its reliability. It also leads to savings from optimized spare parts handling and helps minimize costs associated with unnecessary preventative maintenance and negative impact on safety and the environment caused by unexpected equipment failures. This paper presents a technology of ȃpainted-onȄ ultrasonic transducers that can be integrated into structures to be monitored and accurately measure wear or corrosion induced structure material losses through ultrasonic thickness measurement. These transducers have a small footprint, performance comparable to other commercially available ultrasonic transducers, and can sustain temperatures as high as 400 ºC. They are part of the next generation condition-based maintenance that embraces new advanced sensors connected to wireless network and advanced algorithms to provide a powerful tool for maintenance scheduling and consequently an enormous potential for cost savings. The application of this technology to mining equipment and its synergy with other sensors are discussed. A vision of next generation assets management that includes predictive maintenance enabled by wide use of innovative, low cost and wireless sensors is also discussed.
En savoir plus

2 En savoir plus

Fault Prognostics for the Predictive Maintenance of Wind Turbines: State of the Art

Fault Prognostics for the Predictive Maintenance of Wind Turbines: State of the Art

2 CEA LIST, DM2I, LADIS, France koceila.abid@cea.fr laurence.cornez@cea.fr Abstract. Reliability and availability of wind turbines are crucial due to several reasons. On the one hand, the number and size of wind tur- bines are growing exponentially. On the other hand, installation of these farms at remote locations, such as offshore sites where the environment conditions are favorable, makes maintenance a more tedious task. For this purpose, predictive maintenance is a very attractive strategy in or- der to reduce unscheduled downtime and maintenance cost. Prognostic is an online technique that can provide valuable information for proactive actions such as the current health state and the Remaining Useful Life (RUL). Several fault prognostic works have been published in the litera- ture. This paper provides an overview of the different prognostic phases, including: health indicator construction, degradation detection, and RUL estimation. Different prognostic approaches are presented and compared according to their requirements and performance. Finally, this paper dis- cusses the suitable prognostic approaches for the proactive maintenance of wind turbines, allowing to address the latter challenges.
En savoir plus

14 En savoir plus

Towards a hierarchical modelling approach for planning aircraft tail assignment and predictive maintenance

Towards a hierarchical modelling approach for planning aircraft tail assignment and predictive maintenance

ilyass.haloui@airbus.com - caroline.chanel@isae-supaero.fr - alain.hait@isae-supaero.fr Abstract Aircraft equipment health monitoring system plays a promis- ing role for airlines operation cost reduction, as it can be ex- ploited to perform predictive maintenance. In this vein, a hi- erarchical sequential decision making model is proposed to plan predictive maintenance. It combines linear optimization for routing assignment and MDP planning to handle mainte- nance actions based on the stochastic evolution of health in- dicators. This entangled model should reduce planning time while ensuring a cost-efficient policy.
En savoir plus

4 En savoir plus

Assessing the impact of historical operational data from complex assets on predictive maintenance models

Assessing the impact of historical operational data from complex assets on predictive maintenance models

alongside is companion engine it is very clear that EGT margin is deteriorating at a faster rate. EGT margin for Engine 1 continues to decrease until just after the halfway point of the operational period where it rises suddenly. As a similar jump is not realized in Engine 2, changes in environmental conditions are unlikely, and the jump in EGT margin is likely the result of on-wing maintenance. Just before the third quarter of the operational period, the EGT margin of Engine 2 drops faster compared to Engine 1 and falls below Engine 1. This is an example of the signals that predictive maintenance seeks to find. Both engines see additional, smaller jumps in EGT margin until they are removed for their first SVs. Engine 1 was removed first, and Engine 2 operated for an additional short period of time as shown by the longer red line on the far right of the figure. Both engines were removed before their EGT margin approached zero.
En savoir plus

128 En savoir plus

A cost driven predictive maintenance policy for structural airframe maintenance

A cost driven predictive maintenance policy for structural airframe maintenance

6. Conclusions A cost driven predictive maintenance policy (CDPM) that ensures safety is proposed for structural airframe maintenance. The SHM system is assumed to be employed to track the fati- gue crack in the fuselage panel continuously and to trigger unscheduled maintenance according to the fuselage health state. The CDPM leverages the benefit from both the sched- uled and unscheduled maintenance. On one hand, it skips some unnecessary scheduled maintenance stops. On the other hand, it guarantees the aircraft safety by querying the health state of the fuselage frequently and triggering unscheduled maintenance whenever needed. For each aircraft panel, a model-based prognostics method is developed to estimate the current crack size and to forecast the future reliability of the panel. The proposed maintenance policy is developed at air- craft level. Based on the predicted reliability of all panels, it selects a group of panels which are to be repaired at a sched- uled maintenance stop so as to minimize the cost. The CDPM is applied to the example of a short range commercial aircraft. The simulation results are compared with the traditional scheduled maintenance and the threshold-based maintenance in terms of the average number of maintenance stops, the aver- age number of repaired panels and the average cost per aircraft under same operational conditions. The results show a signif- icant cost reduction achieved by employing the CDPM. By comparing the cost difference between the CDPM and the scheduled maintenance, one can make the decision concerning the implementation of the SHM system on aircraft. More specifically, if the cost incurred by installing and operating an SHM system is lower than the cost saved by employing SHM, then it is worth to install the SHM system on the air- craft. Furthermore the proposed approach allows to assure the cost optimality of the maintenance policy without having to tune any additional parameters. The cost optimality then allows to squeeze out the last few percent of cost savings from prediction based maintenance.
En savoir plus

17 En savoir plus

Towards multi-model approaches to predictive maintenance: A systematic literature survey on diagnostics and prognostics

Towards multi-model approaches to predictive maintenance: A systematic literature survey on diagnostics and prognostics

• The lack of a systematic approach to design and develop predictive maintenance systems [ 108 , 170 ]. There exist standards, norms and generic architectures to develop new predictive maintenance sys- tems, such as OSA-CBM [ 15 ]. However, they only focus on the basic functional components of the system and do not cover important aspects regarding performance indicators or context constraints of the technical system. Besides, they do not offer yet a consistent explanation on which models to use depending on the initial needs of the predictive maintenance system. The lack of a systematic ap- proach limits the implementation of predictive maintenance systems on real scale industrial applications. When developing a new pre- dictive maintenance system the number of potential models to solve the problem is too high. For engineers the simple fact of choosing the right model or a reduced set of models remains a challenging task. It turns out to be very di fficult to perform an objective selec- tion of models as there are not enough comparative studies of the use of different models on the same tasks for predictive maintenance systems. None of the consulted publications in this survey gives extensive explanations for the selection of the proposed method and the architecture methodologies to create a concept of the system varies from one study to another. Besides, many studies do not present detailed design parameters for their proposed models, or the case study data is not available. All these aspects make it difficult to reproduce results and even more difficult to retrieve models from previous studies for use in new predictive maintenance systems. There are no clear guidelines for selecting the right model or models for a specific task given the operational modes and available data to perform diagnostics and prognostics.
En savoir plus

21 En savoir plus

Predictive Maintenance of Technical Faults in Aircraft

Predictive Maintenance of Technical Faults in Aircraft

HEC Liège Management School - University of Liège, QuantOM e-mail: m.schyns@uliege.be A key issue for handlers in the air cargo industry is arrival delays due to air- craft maintenance [1]. This maintenance can be planned or unplanned depending on the underlying cause : it can either be a routine check, which usually causes very little delay due to its periodicity and predictable nature, or it can also be an undetected technical failure which manifests itself during the pre-ight check. The latter is formally known as technical delays. With the recent resurgence of articial intelligence techniques in decision making, especially in cases such as predictive maintenance [2], the approach followed in this work is to gather delay and maintenance data from a cargo handler in order to train a machine learning classier that can predict if a ight will suer from an unexpected technical de- lay or not. In this study, a few typical machine learning techniques are explored and a novel one is also proposed. A new version of the proposed decision tree extension by Hoait & Schyns [3], which will be referred to as condence trees, is presented.
En savoir plus

2 En savoir plus

Machine Learning for Predictive Maintenance in Aviation

Machine Learning for Predictive Maintenance in Aviation

4 FA I L U R E P R E D I C T I O N I N P O S T F L I G H T R E P O R T S In this chapter we present an approach to tackle the problem of event prediction for the purpose of performing predictive maintenance in aviation. Given a collection of recorded events that correspond to equipment failures, our method predicts the next occurrence of one or more events of interest (target events or critical failures). Our objective is to develop an alerting sys- tem that would notify aviation engineers well in advance for upcoming aircraft failures, providing enough time to prepare the corresponding maintenance actions. We formulate a regression problem in order to approximate the risk of occurrence of a target event, given the past occurrences of other events. In order to achieve the best results we employed a multiple instance learning scheme (multiple instance regression) along with extensive data preprocessing. We applied our method on data coming from a fleet of aircraft and our predictions involve failures of compo- nents onboard, specifically components that are related to the landing gear. The event logs correspond to post flight reports retrieved from multiple aircraft during several years of operation. To the best of our knowledge, this is the first attempt on aircraft failure prediction using post flight report data and finally, our findings show high potential impact on the aviation industry.
En savoir plus

161 En savoir plus

A SAW wireless sensor network platform for industrial predictive maintenance

A SAW wireless sensor network platform for industrial predictive maintenance

Abstract Predictive maintenance predicts the system health, based on the current condition, and defines the needed main- tenance activities accordingly. This way, the system is only taken out of service if direct evidence exists that deterio- ration has actually taken place. This increases maintenance efficiency and productivity on one hand, and decreases main- tenance support costs and logistics footprints on the other. We propose a system based on wireless sensor network to moni- tor industrial systems in order to prevent faults and damages. The sensors use the surface acoustic wave technology with an architecture composed of an electronic interrogation device and a passive sensor (without energy at the transducer) which is powered by the radio frequency transmitted by the inter- rogation unit. The radio frequency link transfers energy to the sensor to perform its measurement and to transmit the result to the interrogation unit—or in a description closer to the implemented, characterize the cooperative target cross section characteristics to recover the physical quantity defin- ing the transducer material properties. We use this sensing architecture to measure the temperature of industrial machine components and we evaluate the robustness of the method. This technology can be applied to other physical parame- ters to be monitored. Captured information is transmitted to the base station through multi-hop communications. We also B Violeta Felea
En savoir plus

17 En savoir plus

Predictive maintenance policy for a gradually deteriorating system subject to stress

Predictive maintenance policy for a gradually deteriorating system subject to stress

Abstract This paper deals with a predictive maintenance policy for a continuously deteri- orating system subject to stress. We consider a system with two failure mecha- nisms which are respectively due to an excessive deterioration level and a shock. To optimize the maintenance policy of the system, an approach combining Statisti- cal Process Control (SPC) and Condition-Based Maintenance (CBM) is proposed. CBM policy is used to inspect and replace the system according to the observed deterioration level. SPC is used to monitor the stress covariate. In order to assess the performance of the proposed maintenance policy and to minimize the long-run expected maintenance cost per unit of time, a mathematical model for the main- tained system cost is derived. Analysis based on numerical results are conducted to highlight the properties of the proposed maintenance policy in respect to the different maintenance parameters.
En savoir plus

33 En savoir plus

A case model for predictive maintenance

A case model for predictive maintenance

The result in Chapter 2 tells that, for the orienter, std/mean and peak-to-peak ratio can serve as good metadata for both DAC and position data; for the tilt, max value is a go[r]

68 En savoir plus

Predictive airframe maintenance strategies using model-based prognostics

Predictive airframe maintenance strategies using model-based prognostics

called EKF-FOP method that couples the Extended Kalman filter (EKF) and First-Order Perturbation (FOP), developed in our previous work. 20 EKF-FOP allows to make the repair decision taking into account the future reliability of each individual panel rather than a fixed threshold for all the panels. The EKF-FOP method has two functions: filtering the measurement noise to give a better estimate of damage level (achieved by EKF) and predicting the damage distribution in future time (achieved by FOP). Once the damage distribution of a panel is predicted, the reliability of the panel in future time is calculated. This “predicted reliability information” is used to form the repair policy, which is the core of the predictive maintenance presented in this paper. Similar to Pattabhiraman, we proposed two strategies: the purely predictive maintenance called PdM without considering the aircraft scheduled maintenance and the one called PdM- skip, who takes into account the maintenance schedule. The performance of PdM and PdM-skip is assessed through a cost model by comparing with Pattabhiraman’s two CBM strategies as well as with the traditional scheduled maintenance.
En savoir plus

40 En savoir plus

Development of predictive structural maintenance strategies for aircraft using model-based prognostics

Development of predictive structural maintenance strategies for aircraft using model-based prognostics

CHAPTER 2 LITERATURE REVIEW processing techniques, the research of CBM grows fast since 2000 especially in the last 5 years. Although CBM takes advantage of the known state of the system, the threshold-based maintenance is not always an optimal solution. Determining the thresholds that guarantees safety under different kinds of uncertainties can be a tough work. It requires the experience of reliability experts as well as some robust and accurate algorithm, which might be computational costly. In addition, the pure CBM, i.e., planning the maintenance only based on the actual health state of the system and trigging maintenance anytime when needed without taking into account the preventative maintenance, leads to unscheduled maintenance, which is costlier due to less notification in advance. Recently, a lot of attention moves to predictive maintenance (PdM). In many literatures, CBM refers to as PdM. Indeed, both CBM and PdM rely on the condition monitoring to the system, thus rely on sensor technique, data acquisition and storage technique etc. to know the damage state of the system. However, after being aware of the health state, the policy of planning maintenance scheme in CBM and PdM is different. Zhou [20] presented the concept of condition-based predictive maintenance that integrates the prediction tools into CBM to provide the assessment and prediction of the system hazard rate based on the collected information through continuous monitoring, with the aim of determining the required maintenance action prior to any predicted failure. This concept is very close to current idea of PdM and could be seen as a transition from CBM to PdM. According to [21, 22], the big difference between CBM and PdM is that CBM only uses current system state information to make the maintenance policy while by contrast, PdM makes use, in addition to current degradation information, of predictive information in the form of remaining useful life or the predicted damage index distribution to optimally schedule maintenance actions. With PdM, it is possible to predict the future degradation trajectory of the system thus to predict the possible time when the monitored damage index will reach or exceed a threshold. In that case, the staff could plan the maintenance actions in advance. There are a lot of literatures that study the predictive maintenance from different aspects. The following paragraphs will review the research paper of PdM from policy level, condition-monitoring level and system level.
En savoir plus

142 En savoir plus

Classification des différentes architectures en maintenance.

Classification des différentes architectures en maintenance.

La diversité et la nature des différents systèmes informatiques existants et leur évolution dans le domaine de la maintenance industrielle nous amène à étudier l’architecture logicielle de ceux-ci sous un certain niveau d’abstraction. Nous nous intéressons particulièrement au type d’informations échangées, et à la complexité des relations liant les différents systèmes et applications intégrés dans ces architectures. La section 2 sera consacrée à définir ces caractéristiques qui seront à la base de la définition des différentes architectures de maintenance qui seront abordés à la section 4. Auparavant, on abordera l’historique des systèmes d’information en maintenance qui seront repris à la section suivante sous forme d’architecture : architecture de maintenance, télémaintenance, e-maintenance. Afin de pallier aux manques proposés par les web services issus d’une architecture d’e-maintenance, nous proposons à la section 6 une architecture de s-maintenance basée sur le Web sémantique et adaptée à l’intégration des différents systèmes et applications en maintenance. Ce concept ouvre la possibilité d’utiliser des techniques de gestion des connaissances, de retour d’expérience, etc.
En savoir plus

13 En savoir plus

Show all 1014 documents...