To see the other types of publications on this topic, follow the link: Predictive maintenance.
Dissertations / Theses on the topic 'Predictive maintenance'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 dissertations / theses for your research on the topic 'Predictive maintenance.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
1
Li, Jiawei M. Eng Massachusetts Institute of Technology. "A case model for predictive maintenance." Thesis, Massachusetts Institute of Technology, 2007. http://hdl.handle.net/1721.1/43139.
Full textAbstract:
Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Mechanical Engineering, February 2008.Includes bibliographical references (leaves 59-60).
This project is to respond to a need by Varian Semiconductor Equipment Associates, Inc. (VSEA) to help predict failure of ion implanters. Predictive maintenance would help to reduce the unscheduled downtime of ion implanters, whose throughput and uptime is highly important to customers. Statistical analysis is performed on historical data to extract metadata that can reflect the machine health, and statistical process control (SPC) is applied to detect deviations from normal or in-control behavior. Methods for failure prevention are also investigated. Challenging points in this project are the noise in raw signal data and the difference in data signals of different robots. To address these challenges, we apply signal filtering to extract cycle motions from raw data, and develop different generic as well as specific metadata extraction techniques for different robots. We test the extraction approaches and results using healthy data of ten machines, and find that the metadata on which we chose to perform SPC is suitable and can serve as a consistent indicator of a machine's health. We further develop an application using Visual Basic based on our study, and provide a user guide on how to generate the analysis reports on new data using our application.
by Jiawei Li.
M.Eng.
2
Tyagi, Prakhar. "Chassis predictive maintenance and service solutions." Thesis, KTH, Fordonsdynamik, 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-265587.
Full textAbstract:
Predictive Maintenance (PdM) accumulates data from multiple sensors developing a statistical model which identifies the key failures even before they take place. The main focus of this thesis work has been the proposal of a machine learning based system designed for predicting the failure of mechanical parts that require replacement. The main investigation explores the possibilities of implementing machine learning algorithm for predicting the parts that require replacement and which is found from the electronic errors that the vehicle exhibits. A strong association between the parts that cause faults and electronic error codes helps in yielding a powerful diagnostics tool. The study has considered three error components namely; broken damper, noisy wheel hub and the reference value for the validation purpose. The model vehicle used for the study is Volvo V90. To acquire variance in this study data, diverse tracks with different speeds were used. The machine learning algorithm that was developed can classify and detect mechanical failures using an Support Vector Machine (SVM) algorithm based on various statistical learning methods. The study carried out an fast Fourier transform (FFT) analysis in association with the data acquired from front left wheel. The main area of interest is the FFT domain of 5-20hz. The study outcome indicated that the used model is capable of predicting the hysteretic responses associated with the faulty components like broken damper and noisy wheel hub. The designed model can be used for analysing the system’s response and for designing and controlling the faulty components in the car. However, the results of this thesis work can be used to implement the time-based prediction of mechanical component decay.Prediktivt Underhåll (PdM) är en statistisk modell som samlar data från flera olika sensorer och som identifierar fel innan de äger rum. Huvudfokus för detta examensarbete har varit förslaget till ett maskininlärningsbaserat system som är utformat för att förutsäga fel i mekaniska delar som kräver utbyte. Examensarbetet undersöker möjligheterna att implementera en maskininlärningsalgoritm för att förutsäga de mekaniska delar som kräver utbyte och som framgår av de elektroniska fel som fordonet uppvisar. En stark koppling mellan de delar som orsakar fel och elektroniska felkoder hjälper till att ge ett kraftfullt diagnostiskt verktyg. Studien har beaktat tre felkomponenter nämligen; trasig dämpare, missljud från hjulnav och referensvärdet för valideringsändamål. Modellfordonet som används för studien är Volvo V90. För att få varians i informationen för detta arbete användes olika provbanor med olika vägförhållanden med olika hastigheter. Maskininlärningsalgoritmen som utvecklades kan klassificera och upptäcka mekaniska fel med hjälp av en SVM-algoritm (Support Vector Machine) baserad på olika statistiska inlärningsmetoder. Studien genomförde en snabb Fourier-transform (FFT) analys i samband med de data som förvärvades från det främre vänstra hjulet. Huvudintresseområdet är FFT-domänen 5-20 Hz. Studiens resultat visade att den använda modellen kan: Identifiera och klassificera data som är förknippade med de felaktiga komponenterna som trasig dämpare och missljud i hjulnav. Modellen kan användas för vidare prediktera och ge förslag när ett mekaniskt fel på dämpare eller hjulnav håller på att ske. Det här examensarbetet täcker inte tidsbunden prediktion utan snarare identifierar när nedbrytningen av mekaniska komponenter har skett. Resultaten från detta examensarbete kan emellertid användas för att implementera en tidsbaserad prediktion för mekaniska komponentfel.
3
Korvesis, Panagiotis. "Machine Learning for Predictive Maintenance in Aviation." Thesis, Université Paris-Saclay (ComUE), 2017. http://www.theses.fr/2017SACLX093/document.
Full textAbstract:
L'augmentation des données disponibles dans presque tous les domaines soulève la nécessité d'utiliser des algorithmes pour l'analyse automatisée des données. Cette nécessité est mise en évidence dans la maintenance prédictive, où l'objectif est de prédire les pannes des systèmes en observant continuellement leur état, afin de planifier les actions de maintenance à l'avance. Ces observations sont générées par des systèmes de surveillance habituellement sous la forme de séries temporelles et de journaux d'événements et couvrent la durée de vie des composants correspondants. Le principal défi de la maintenance prédictive est l'analyse de l'historique d'observation afin de développer des modèles prédictifs.Dans ce sens, l'apprentissage automatique est devenu omniprésent puisqu'il fournit les moyens d'extraire les connaissances d'une grande variété de sources de données avec une intervention humaine minimale. L'objectif de cette thèse est d'étudier et de résoudre les problèmes dans l'aviation liés à la prévision des pannes de composants à bord. La quantité de données liées à l'exploitation des avions est énorme et, par conséquent, l'évolutivité est une condition essentielle dans chaque approche proposée.Cette thèse est divisée en trois parties qui correspondent aux différentes sources de données que nous avons rencontrées au cours de notre travail. Dans la première partie, nous avons ciblé le problème de la prédiction des pannes des systèmes, compte tenu de l'historique des Post Flight Reports. Nous avons proposé une approche statistique basée sur la régression précédée d'une formulation méticuleuse et d'un prétraitement / transformation de données. Notre méthode estime le risque d'échec avec une solution évolutive, déployée dans un environnement de cluster en apprentissage et en déploiement. À notre connaissance, il n'y a pas de méthode disponible pour résoudre ce problème jusqu'au moment où cette thèse a été écrite.La deuxième partie consiste à analyser les données du livre de bord, qui consistent en un texte décrivant les problèmes d'avions et les actions de maintenance correspondantes. Le livre de bord contient des informations qui ne sont pas présentes dans les Post Flight Reports bien qu'elles soient essentielles dans plusieurs applications, comme la prédiction de l'échec. Cependant, le journal de bord contient du texte écrit par des humains, il contient beaucoup de bruit qui doit être supprimé afin d'extraire les informations utiles. Nous avons abordé ce problème en proposant une approche basée sur des représentations vectorielles de mots. Notre approche exploite des similitudes sémantiques, apprises par des neural networks qui ont généré les représentations vectorielles, afin d'identifier et de corriger les fautes d'orthographe et les abréviations. Enfin, des mots-clés importants sont extraits à l'aide du Part of Speech Tagging.Dans la troisième partie, nous avons abordé le problème de l'évaluation de l'état des composants à bord en utilisant les mesures des capteurs. Dans les cas considérés, l'état du composant est évalué par l'ampleur de la fluctuation du capteur et une tendance à l'augmentation monotone. Dans notre approche, nous avons formulé un problème de décomposition des séries temporelles afin de séparer les fluctuations de la tendance en résolvant un problème convexe. Pour quantifier l'état du composant, nous calculons à l'aide de Gaussian Mixture Models une fonction de risque qui mesure l'écart du capteur par rapport à son comportement normalThe increase of available data in almost every domain raises the necessity of employing algorithms for automated data analysis. This necessity is highlighted in predictive maintenance, where the ultimate objective is to predict failures of hardware components by continuously observing their status, in order to plan maintenance actions well in advance. These observations are generated by monitoring systems usually in the form of time series and event logs and cover the lifespan of the corresponding components. Analyzing this history of observation in order to develop predictive models is the main challenge of data driven predictive maintenance.Towards this direction, Machine Learning has become ubiquitous since it provides the means of extracting knowledge from a variety of data sources with the minimum human intervention. The goal of this dissertation is to study and address challenging problems in aviation related to predicting failures of components on-board. The amount of data related to the operation of aircraft is enormous and therefore, scalability is a key requirement in every proposed approach.This dissertation is divided in three main parts that correspond to the different data sources that we encountered during our work. In the first part, we targeted the problem of predicting system failures, given the history of Post Flight Reports. We proposed a regression-based approach preceded by a meticulous formulation and data pre-processing/transformation. Our method approximates the risk of failure with a scalable solution, deployed in a cluster environment both in training and testing. To our knowledge, there is no available method for tackling this problem until the time this thesis was written.The second part consists analyzing logbook data, which consist of text describing aircraft issues and the corresponding maintenance actions and it is written by maintenance engineers. The logbook contains information that is not reflected in the post-flight reports and it is very essential in several applications, including failure prediction. However, since the logbook contains text written by humans, it contains a lot of noise that needs to be removed in order to extract useful information. We tackled this problem by proposing an approach based on vector representations of words (or word embeddings). Our approach exploits semantic similarities of words, learned by neural networks that generated the vector representations, in order to identify and correct spelling mistakes and abbreviations. Finally, important keywords are extracted using Part of Speech Tagging.In the third part, we tackled the problem of assessing the health of components on-board using sensor measurements. In the cases under consideration, the condition of the component is assessed by the magnitude of the sensor's fluctuation and a monotonically increasing trend. In our approach, we formulated a time series decomposition problem in order to separate the fluctuation from the trend by solving a convex program. To quantify the condition of the component, we compute a risk function which measures the sensor's deviation from it's normal behavior, which is learned using Gaussian Mixture Models
4
Karlsson, Lotta. "Predictive Maintenance for RM12 with Machine Learning." Thesis, Högskolan i Halmstad, Akademin för ekonomi, teknik och naturvetenskap, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:hh:diva-42283.
Full textAbstract:
Few components within mechanical engineering possess the fatigue resistance as of high-pressure turbine blades found in jet engines. This as they are designed to perform in extensively high temperatures under severe loading which causes degradation to be an important aspect despite a design, optimized for its environment. This study aims to find a method for predicting life consumption of those blades belonging to the turbine section of the jet engine in JAS 39 Gripen C/D called RM12. This was performed at GKN Aerospace, which holds the military type certificate for this engine as well as a patented solution that determines life consumption in components depending on operational history. With the help of machine learning in Matlab, flight sensor data and loading results, the method was to explore a variety of prediction models and find a selection of blades with varied utilization before reaching end of life for comparison. Followed by a search of understanding the life limiting fatigue conditions and the factors involved in the deterioration process. A similarity finding approach gave valuable meaning to the accuracy of regression analysis from flight data towards output in form of temperature predictions. Comparing known and reliable fatigue calculation results gave however no clear picture as inspected blades had reach their limit at very diverse accumulated values. The next approach was therefore to investigate if an initialization point of degradation could be found, from where the result could give an answer that matched for all blades and their different utilization. The result was that an accelerated degradation after high loading could give a prediction that could explain the total life consumption with an accuracy of 87% for 19 out of 21 investigated blades. The accelerated deterioration could in theory be explained by the fact that the fatigue resistance as well as different types of degradation, propagates each other and originates from thermal loading making them all contributors, whereas the conventional numerical methods only handles them separately. In order to get confidence, valuable and reliable predictions, the models do however need to be accompanied with more testing and adding of contributing factors before assumed as a proven method for life consumption determination of the high-pressure turbine blades.
5
Sedghi, Mahdieh. "Data-driven predictive maintenance planning and scheduling." Licentiate thesis, Luleå tekniska universitet, Industriell Ekonomi, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:ltu:diva-80828.
Full textAbstract:
The railway track network is one of the major modes of transportation and among a country’s most valuable infrastructure assets. Maintenance and renewal of railway infrastructure have a vital role in safety performance, the quality of the ride, train punctuality, and the life cycle cost of assets. Due to the large proportion of maintenance costs, increasing the efficiency of maintenance through optimised planning can result in high amounts of cost-saving. Moreover, from a safety perspective, late maintenance intervention can result in defective track and rollingstock components, which in severe cases, can cause severe accidents such as derailments. An effective maintenance management system is required to ensure the availability of the infrastructure system and meet the increasing capacity demand. The recent rapid technological revolution and increasing deployment of sensors and connected devices created new possibilities to increase the maintenance strategy effectiveness in the railway network. The purpose of this thesis is to expand the knowledge and methods for planning and scheduling of railway infrastructure maintenance. The research vision is to find quantitative approaches for integrated tactical planning and operational scheduling of predictive condition-based maintenance which can be put to practical use and improve the efficiency of the railway system. First, a thorough literature review study is performed to identify improvement policies for maintenance planning and scheduling in the literature and also to analyse the current approaches in optimising the maintenance planning and scheduling problem. Second, a novel data-driven multi-level decision-making framework to improve the efficiency of maintenance planning and scheduling is developed. The proposed framework aims to support the selection of track segments for maintenance by providing a practical degradation prediction model based on available condition measurement data. The framework considers the uncertainty of future predictions using the probability of surpassing a maintenance limit instead of using the predicted value. Moreover, an extensive total maintenance cost formulation is developed to include both direct and indirect preventive and corrective costs to observe the effect of using cost optimisation and grouping algorithms at the operational scheduling level. The performance of the proposed framework is evaluated through a case study based on data from a track section of the iron ore line between Boden and Luleå. The results indicate that the proposed approach can lead to cost savings in both optimal and grouping plans. This framework may be a useful decision support tool in the automated planning and scheduling of maintenance based on track geometry measurements.
6
Killeen, Patrick. "Knowledge-Based Predictive Maintenance for Fleet Management." Thesis, Université d'Ottawa / University of Ottawa, 2020. http://hdl.handle.net/10393/40086.
Full textAbstract:
In recent years, advances in information technology have led to an increasing number of devices (or things) being connected to the internet; the resulting data can be used by applications to acquire new knowledge. The Internet of Things (IoT) (a network of computing devices that have the ability to interact with their environment without requiring user interaction) and big data (a field that deals with the exponentially increasing rate of data creation, which is a challenge for the cloud in its current state and for standard data analysis technologies) have become hot topics. With all this data being produced, new applications such as predictive maintenance are possible. One such application is monitoring a fleet of vehicles in real-time to predict their remaining useful life, which could help companies lower their fleet management costs by reducing their fleet's average vehicle downtime. Consensus self-organized models (COSMO) approach is an example of a predictive maintenance system for a fleet of public transport buses, which attempts to diagnose faulty buses that deviate from the rest of the bus fleet. The present work proposes a novel IoT-based architecture for predictive maintenance that consists of three primary nodes: namely, the vehicle node (VN), the server leader node (SLN), and the root node (RN). The VN represents the vehicle and performs lightweight data acquisition, data analytics, and data storage. The VN is connected to the fleet via its wireless internet connection. The SLN is responsible for managing a region of vehicles, and it performs more heavy-duty data storage, fleet-wide analytics, and networking. The RN is the central point of administration for the entire system. It controls the entire fleet and provides the application interface to the fleet system. A minimally viable prototype (MVP) of the proposed architecture was implemented and deployed to a garage of the Soci\'et\'e de Transport de l'Outaouais (STO), Gatineau, Canada. The VN in the MVP was implemented using a Raspberry Pi, which acquired sensor data from a STO hybrid bus by reading from a J1939 network, the SLN was implemented using a laptop, and the RN was deployed using meshcentral.com. The goal of the MVP was to perform predictive maintenance for the STO to help reduce their fleet management costs.
The present work also proposes a fleet-wide unsupervised dynamic sensor selection algorithm, which attempts to improve the sensor selection performed by the COSMO approach. I named this algorithm the improved consensus self-organized models (ICOSMO) approach. To analyze the performance of ICOSMO, a fleet simulation was implemented. The J1939 data gathered from a STO hybrid bus, which was acquired using the MVP, was used to generate synthetic data to simulate vehicles, faults, and repairs. The deviation detection of the COSMO and ICOSMO approach was applied to the synthetic sensor data. The simulation results were used to compare the performance of the COSMO and ICOSMO approach. Results revealed that in general ICOSMO improved the accuracy of COSMO when COSMO was not performing optimally; that is, in the following situations: a) when the histogram distance chosen by COSMO was a poor choice, b) in an environment with relatively high sensor white noise, and c) when COSMO selected poor sensors. On average ICOSMO only rarely reduced the accuracy of COSMO, which is promising since it suggests deploying ICOSMO as a predictive maintenance system should perform just as well or better than COSMO . More experiments are required to better understand the performance of ICOSMO. The goal is to eventually deploy ICOSMO to the MVP.
7
Williamsson, Ia. "Total Quality Maintenance (TQMain) A predictive and proactive maintenance concept for software." Thesis, Blekinge Tekniska Högskola, Avdelningen för för interaktion och systemdesign, 2006. http://urn.kb.se/resolve?urn=urn:nbn:se:bth-2281.
Full textAbstract:
This thesis describes an investigation of the possibility to apply a maintenance concept originally developed for the industry, on software maintenance. Today a large amount of software development models exist but not many of them treat maintenance as a part of the software life cycle. In most cases maintenance is depicted as an activity towards the end of the software life cycle. The high cost ascribed to software maintenance motivates for improvements. The maintenance concept TQMain proposed in this thesis distinguishes from other maintenance concepts by its use of preventive, predictive and proactive maintenance strategies. TQMain uses a common database to store real-time data from various departments and uses it for analyse and assessment to track the development of deviations in the condition of the production process and product quality at an early stage. A continuous cyclic improvement of the maintenance strategy is reached by comparing the data from the real-time measurements with data from the database. The ISO/IEC Software engineering – Product qualities is used as a source of empiric data to conclude that the correct quality characteristics are used for identifying software product quality and its characteristics and compare them with the characteristics of industrial product quality. The results presented are that in the conceptual outline of TQMain measures are obviously not the same as in software maintenance, but the aspect of product quality is common for both. The continuous cyclic improvement of the product quality that TQMain features together with the aspect of detecting potential failures before they occur would, judging from the conceptual outline of TQMain be applicable on software maintenance.
8
Pryor, Jacqueline. "Earthwork maintenance : a geotechnical database and predictive model." Thesis, Cardiff University, 1998. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.266614.
Full text
9
De, Giorgi Marcello. "Tree ensemble methods for Predictive Maintenance: a case study." Master's thesis, Alma Mater Studiorum - Università di Bologna, 2021. http://amslaurea.unibo.it/22282/.
Full textAbstract:
Nel lavoro descritto in questa tesi sono stati creati modelli per la manutenzione predittiva di macchine utensili in ambito industriale; in particolare, i modelli realizzati sono stati addestrati sfruttando degli ensemble tree methods con le finalità di: predire il verificarsi di un guasto in macchina con un anticipo tale da permettere l'organizzazione delle squadre di manutenzione; predire la necessità della sostituzione anticipata dell'utensile utilizzato dalla macchina, per mantenere alti gli standard di qualità. Dopo aver dato uno sfondo al contesto industriale in esame, la tesi illustra i processi seguiti per la creazione e l'aggregazione di un dataset, e l'introduzione di informazioni relative agli eventi in macchina. Analizzato il comportamento di alcune variabili durante la lavorazione ed effettuata una distinzione tra cicli di lavorazione validi e non validi, si procede introducendo gli ensemble tree methods e il motivo della scelta di questa classe di algoritmi. Nel dettaglio, vengono presentati due possibili candidati al problema trattato: Random Forest ed XGBoost; dopo averne descritto il funzionamento, vengono presentati i risultati ottenuti dai modelli proponendo, per stimarne l'efficacia, un funzione di costo atteso come alternativa all'accuracy score. I risultati dei modelli allenati con i due algoritmi proposti vengono infine confrontati.
10
FURTADO, FELIPE MIANA DE FARIA. "NEURAL NETWORKS FOR PREDICTIVE MAINTENANCE ON OFF-HIGWAY TRUCKS." PONTIFÍCIA UNIVERSIDADE CATÓLICA DO RIO DE JANEIRO, 2009. http://www.maxwell.vrac.puc-rio.br/Busca_etds.php?strSecao=resultado&nrSeq=15673@1.
Full textAbstract:
CONSELHO NACIONAL DE DESENVOLVIMENTO CIENTÍFICO E TECNOLÓGICOCom o aumento da demanda por minério no mundo, a complexidade, o tamanho e o preço dos equipamentos de extração mineral aumentaram consideravelmente. Como estas máquinas possuem uma tecnologia de monitoramento embarcada no equipamento, a utilização desses dados para o aumento da confiabilidade e da disponibilidade do equipamento tornou-se fundamental, de modo a reduzir os custos de manutenção. O objetivo desta dissertação foi desenvolver um modelo de apoio à decisão de parada de equipamento, baseado na classificação por Redes Neurais Artificiais de padrões pré-falha de caminhões fora de estrada. O modelo proposto tem como objetivo identificar o estado de falha, ou padrão pré-falha de um equipamento, utilizando os dados armazenados nos equipamentos e seus respectivos registros de falha, para que seja possível avaliar o risco de falha deste equipamento e decidir se o mesmo deve ser parado ou aguardar uma nova parada programada. Essa dissertação foi desenvolvida em quatro partes: estudo dos principais modelos de manutenção atualmente utilizados; definição e desenvolvimento do modelo para abordar o problema, baseado em redes neurais artificiais; avaliação de desempenho do modelo proposto; e simulação do downtime da máquina utilizando o modelo de decisão proposto. No estudo dos principais modelos foi realizada uma pesquisa bibliográfica sobre a evolução da manutenção, passando por modelos de manutenção corretiva, manutenção preventiva e, por fim, chegando ao modelo de manutenção baseada no monitoramento de condições. Para os dois últimos tipos de manutenção, foram apresentados os principais modelos utilizados na abordagem do problema, seus benefícios e deficiências. O desenvolvimento do modelo foi segmentado em três etapas principais: tratamento das bases de dados, tanto de dados obtidos diretamente do equipamento quanto das bases de registro de falha dos equipamentos; seleção de variáveis, baseada no cálculo da influência de cada sensor do equipamento na determinação de seu estado de falha, assim como na definição do intervalo ideal para se agrupar os dados; e definição da topologia das redes. Na etapa de avaliação do desempenho do modelo proposto foram utilizados dados de falhas corretivas mais recorrentes para os dois componentes específicos de caminhões fora de estrada: motor e transmissão, sendo que o monitoramento eletrônico do motor é mais extenso do que o de transmissão, no que diz respeito ao número de sensores empregados no monitoramento. Para a comparação de desempenho entre os diferentes modelos avaliados, dois fatores tiveram maior relevância: melhor desempenho na classificação e maior intervalo entre a identificação do padrão pré-falha e a ocorrência da falha. Os resultados de classificação dos padrões pré-falha foram bastante satisfatórios para a maioria dos casos de estudos, com as taxas de acerto variando entre 85% e 95%. A partir do modelo de classificação determinado na etapa anterior, passou-se à simulação de diferentes cenários de falhas, calculando-se os tempos de máquina parada (downtimes) que teriam sido evitados se as intervenções definidas pelo modelo tivessem sido executadas, analisando-se, assim, o aumento de disponibilidade proporcionado pelo uso do modelo proposto.
With the increasing demand for ore in the world, the complexity, size and price of mining equipment have increased considerably. As these machines have embedded monitoring technology, the use of such data to increase the reliability and availability of the equipment has become essential in order to reduce maintenance costs. The objective of this work is developing a model that supports the decision of stopping an equipment, based on its actual state, using pattern recognition by neural networks. The proposed model aims to identify the state of equipment failure or pre-failure based on the data stored in the equipment and on the records of failure, so as to assess the risk of failure of equipment and to decide whether it should be stopped or wait for a new programmed shutdown. This dissertation was developed in four parts: study of the main models currently used for maintenance; design and implementation of the model to address this problem, based on artificial neural networks; performance evaluation of the proposed model; and simulation of equipment downtime using the proposed model. In the study of the main models a research was made about the evolution of maintenance techniques, through models of corrective maintenance, preventive maintenance and, finally, reaching the maintenance model based on condition monitoring. For the last two types of maintenance, it is presented the main models used in addressing the problem, its benefits and shortcomings. The development of the model was segmented into three main stages: processing of databases, from the data obtained directly from the equipment to the base of record of equipment failure; variable selection, based on the calculation of the influence of each equipment sensor to determine its failure state, as well as the definition of the ideal range of group data, and definition of the topology of networks. In the stage of assessing the performance of the proposed model we used data from corrective failures more often of two specific components of off-highway trucks: engine and transmission. To compare the performance between the different models evaluated, two factors were more important: classification performance and the longest interval between the identification of a pre-failure pattern and the occurrence of the failure. The results of classification of pre-failure patterns were quite satisfactory for most case studies, with hit rates ranging between 85% and 96%. From the classification model given in the previous step, we moved on to simulate different failure scenarios, calculating the equipment downtime that would have been avoided if the interventions defined by the model had been implemented, thus analyzing the increased availability provided by the use of the proposed model.
11
Gorman, Joe, Glenn Takata, Subhash Patel, and Dan Grecu. "A Constraint-Based Approach to Predictive Maintenance Model Development." International Foundation for Telemetering, 2008. http://hdl.handle.net/10150/606187.
Full textAbstract:
ITC/USA 2008 Conference Proceedings / The Forty-Fourth Annual International Telemetering Conference and Technical Exhibition / October 27-30, 2008 / Town and Country Resort & Convention Center, San Diego, CaliforniaPredictive maintenance is the combination of inspection and data analysis to perform maintenance when the need is indicated by unit performance. Significant cost savings are possible while preserving a high level of system performance and readiness. Identifying predictors of maintenance conditions requires expert knowledge and the ability to process large data sets. This paper describes a novel use of constraint-based data-mining to model exceedence conditions. The approach extends the extract, transformation, and load process with domain aggregate approximation to encode expert knowledge. A data-mining workbench enables an expert to pose hypotheses that constrain a multivariate data-mining process.
12
Odekvist, Joshua, and Robert Antar. "Predictive Maintenance and Data Analysis in Ellevio's Distribution Grid." Thesis, Uppsala universitet, Institutionen för teknikvetenskaper, 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-398597.
Full textAbstract:
As part of the international EU project Integrid, Ellevio has started a pilot project where measurement equipment has been installed in secondary substations in Stockholm Royal Seaport. The purpose of the pilot project is to improve operation, maintenance and future planning of the grid. This thesis focuses on predictive maintenance and data analysis of power quality parameters and aims to cover several areas within these two categories to answer the following research questions: -In what ways can the life-cycle and faults of components be used as a basis for predictive maintenance? -Are there any measurable parameters - currently being monitored or not - that can be useful for Ellevio in terms of maintenance and grid development? -Is the installation of more monitoring systems worthwhile, and if so, what should be monitored? -Is it possible to find indicators that allow for predictive maintenance to reduce the interruption time and maintenance cost? The main results are that the focus should lie on middle voltage (MV) cables due to its significant contribution to SAIDI and SAIFI. Monitoring partial discharges in MV cables could potentially lower the frequency of faults. A study needs to be conducted to establish the extent of partial discharges as a source of faults. Furthermore harmonic content was found to have a large impact on transformer losses and this parameter should be considered when dimensioning new secondary substations. Another parameter that had interesting results is the current unbalance which led to high neutral currents and caused high losses.
13
AHMED, Umair. "DECISION-MAKING MODELS FOR PREDICTIVE MAINTENANCE SERVICE SUPPORT SYSTEMS." Doctoral thesis, Università degli Studi di Palermo, 2023. https://hdl.handle.net/10447/579250.
Full textAbstract:
Nell'era digitale, la tecnologia è in continua evoluzione, con enormi progressi nell'automazione che consentono una gestione della manutenzione più efficiente ed economica. Le tecnologie digitali stanno convergendo e avanzando insieme alle industrie, determinando progressi significativi nella gestione della manutenzione. La tradizionale strategia di manutenzione preventiva gestita dall'uomo lascia progressivamente spazio alla manutenzione predittiva, che rappresenta un’ottima opportunità per migliorare significativamente la pianificazione della manutenzione del sistema, in particolare per i sistemi più complessi e dal significativo valore monetario. Tuttavia, l’implementazione di tecniche di manutenzione predittiva si trova ad affrontare una serie di sfide sostanziali, essendo richiesti l’utilizzo di tecnologie di tracciamento moderne, lo sviluppo di solidi sistemi di raccolta dati e l'esecuzione di una varietà di procedure complesse.
Considerando il ruolo chiave della gestione della manutenzione nelle industrie, la motivazione principale di questo lavoro di ricerca consiste nell’indagare le pratiche esistenti e proporre nuove metodologie in grado di fornire implicazioni pratiche che possono essere utili nel contribuire a questo campo di studio in termini di previsione dei guasti, efficienza e ottimizzazione dei costi. Il presente lavoro di tesi è organizzato in tre capitoli, che rappresentano le principali aree di studio: 1) panoramica sulla gestione della manutenzione, 2) modelli decisionali a supporto della manutenzione predittiva, 3) trasformazione digitale nella gestione della manutenzione. Gli obiettivi di ricerca relativi ai menzionati capitoli sono: 1) studiare le attuali pratiche di manutenzione predittiva e le sue applicazioni nell'industria per identificare la sua capacità di prevedere e controllare i guasti delle apparecchiature di sistemi complessi; 2) studiare vari metodi di decisione multi-criterio (MCDM) e le loro applicazioni in modo da sviluppare una metodologia decisionale di manutenzione predittiva integrata per sistemi complessi nell'industria 4.0; 3) studiare la trasformazione digitale della gestione della manutenzione e i fattori critici della digitalizzazione, nonché l'incertezza nel processo decisionale per la gestione della manutenzione nell'industria 4.0.
Questi obiettivi di ricerca vengono perseguiti attraverso una metodologia mista, ovvero sia qualitativa e sia quantitativa, basata su un ampio studio della letteratura. È stata sviluppata una revisione della letteratura sulla manutenzione predittiva e le sue applicazioni industriali insieme ai suoi limiti per identificare le carenze negli approcci esistenti. Sono state inoltre studiate varie metodologie MCDM per analizzarne gli effetti nella gestione della manutenzione ed è stata sviluppata una pletora di casi reali per offrire spunti gestionali pratici.In the digital era, technology is continually evolving, with enormous advancements in automation enabling more efficient and cost-effective maintenance management. Digital technologies are converging and advancing in tandem with industries, resulting in significant progress in maintenance management. The traditionally human-managed preventive maintenance strategy is outclassed with predictive maintenance, something that represents a wonderful opportunity to significantly improve system maintenance planning, particularly for more complex systems with a significant monetary value. However, predictive maintenance methods face numerous substantial challenges in terms of their application, as they necessitate the use of contemporary tracking technologies, the development of robust data-gathering systems, and the execution of a variety of intricate procedures. Considering the significance of maintenance management in industries, the primary motivation for this research work is to investigate existing practices and propose new methodologies capable of providing practical implications that may be useful in contributing to this field of study in terms of predicting failures, efficiency, and cost optimization. The present work is organized through three chapters, representing the main areas of study: 1) overview on maintenance management, 2) decision-making models supporting predictive maintenance, and 3) digital transformation in maintenance management. The objectives of research linked to the defined chapters are; 1) to study current practices of predictive maintenance and its applications in industry to identify its capability to predict and control equipment failures of complex systems; 2) to investigate various Multi-Criteria Decision-Making (MCDM) methods and their applications so as to develop an integrated predictive maintenance decision-making methodology for complex systems in industry 4.0; 3) to study the digital transformation of maintenance management and critical factors of digitalization, as well as uncertainty in the decision-making process for maintenance management in industry 4.0. In achieving the objectives of this research, a mixed methodology, i.e., qualitative and quantitative research, is carried out on the basis of an extensive literature study. A literature review of predictive maintenance, its industrial applications along with its limitations is developed to identify the shortcomings in existing approaches. Various MCDM methodologies have been studied as well to investigate their effects on maintenance management and a plethora of real-world cases have been developed to offer practical managerial insights.
14
Murphy, Killian. "Predictive maintenance of network equipment using machine learning methods." Electronic Thesis or Diss., Institut polytechnique de Paris, 2024. http://www.theses.fr/2024IPPAS013.
Full textAbstract:
Avec la montée en puissance des capacités de calcul nécessaires pour les méthodes plus développées d'Apprentissage Machine (ML), la Prédiction des Incidents Réseau (NFP:Network Fault Prediction) connait un regain d'intérêt scientifique. La capacité de prédire les incidents des équipements réseau est de plus en plus fréquemment identifiée comme un moyen efficace d'améliorer la fiabilité du réseau. Cette capacité prédictive peut être utilisée pour atténuer ou mettre en œuvre une maintenance prédictive en prévision des cas d'incidents réseau imminents. Cela pourrait contribuer à la mise en œuvre de réseaux sans défaillance et sans pertes, et permettre aux applications critiques d'être exécutées sur des réseaux de plus grandes dimensions et hétérogènes. Dans ce manuscrit, nous nous proposons de contribuer au domaine du NFP en nous focalisant sur la prédiction des alertes réseau. Dans un premier temps, nous présentons une étude de l'état de l'art complet du NFP en utilisant des méthodes d'apprentissage machine (ML) entièrement dédiée aux réseaux de télécommunications. Ensuite, nous établissons de futures directions de recherche dans le domaine. Dans un deuxième temps, nous proposons et étudions un couple de métriques (Réduction des coûts de maintenance, et mesure des gains de Qualité de Service) de performances de ML adaptées au NFP dans le cadre de la maintenance des réseaux. Dans un troisième temps, nous décrivons l'architecture complète de traitement des données, incluant l'infrastructure réseau et logicielle, et la chaîne de prétraitement des données nécessaires au ML qui ont été mis en œuvre chez SPIE ICS, société d'intégration de réseaux et de systèmes. Nous décrivons également avec précision le modèle du problème d'alarme et d'incidents. Dans un quatrième temps, nous établissons une comparaison des différentes méthodes de ML appliquées à notre jeu de données. Nous considérons des méthodes conventionnelles de ML, basés sur des arbres de décision, des perceptrons multicouches et des Séparateurs à Vastes Marges. Nous testons la généralisation des performances des modèles par rapport aux différents types d'équipements, ainsi que les généralisations en ML des modèles de ML et des paramètres proposés. Ensuite, nous étudions avec succès les architectures de ML à entrée séquentielle - Réseaux de neurones convolutifs et Long Short Term Memory - dans le cas de données SNMP séquentielles sur notre ensemble de données. Finalement, nous étudions l'impact de la définition de l'horizon de prédiction (et des variables arbitraires associées) sur la performance de prédiction des modèles MLWith the improvement of computation power necessary for advanced applications of Machine Learning (ML), Network Fault Prediction (NFP) experiences a renewed scientific interest. The ability to predict network equipment failure is increasingly identified as an effective means to improve network reliability. This predictive capability can be used, to mitigate or to enact predictive maintenance on incoming network failures. This could contribute to establishing zero-failure networks and allow safety-critical applications to run over higher dimension and heterogeneous networks.In this PhD thesis, we propose to contribute to the NFP field by focusing on network alarm prediction. First, we present a comprehensive survey on NFP using Machine Learning (ML) methods entirely dedicated to telecommunication networks, and determine new directions for research in the field. Second, we propose and study a set of Machine Learning performance metrics (maintenance cost reduction and Quality of Service improvement) adapted to NFP in the context of network maintenance. Third, we describe the complete data processing architecture, including the network and software infrastructure, and the necessary data preprocessing pipeline that was implemented at SPIE ICS, Networks and Systems Integrator. We also describe the alarm or failure prediction problem model precisely. Fourth, we establish a benchmark of the different ML solutions applied to our dataset. We consider Decision Tree-based methods, Multi-Layer Perceptron and Support Vector Machines. We test the generalization of performance prediction across equipment types as well as normal ML generalization of the proposed models and parameters.Then, we apply sequential - Convolutional Neural Networks (CNN) and Long Short Term Memory (LSTM) - ML architectures with success on our sequential SNMP dataset. Finally, we study the impact of the definition of the prediction horizon (and associated arbitrary timeframes) on the ML model prediction performance
15
Kumbala, Bharadwaj Reddy. "Predictive Maintenance of NOx Sensor using Deep Learning : Time series prediction with encoder-decoder LSTM." Thesis, Blekinge Tekniska Högskola, Institutionen för tillämpad signalbehandling, 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:bth-18668.
Full textAbstract:
In automotive industry there is a growing need for predicting the failure of a component, to achieve the cost saving and customer satisfaction. As failure in a component leads to the work breakdown for the customer. This paper describes an effort in making a prediction failure monitoring model for NOx sensor in trucks. It is a component that used to measure the level of nitrogen oxide emission from the truck. The NOx sensor has chosen because its failure leads to the slowdown of engine efficiency and it is fragile and costly to replace. The data from a good and contaminated NOx sensor which is collated from the test rigs is used the input to the model. This work in this paper shows approach of complementing the Deep Learning models with Machine Learning algorithm to achieve the results. In this work LSTMs are used to detect the gain in NOx sensor and Encoder-Decoder LSTM is used to predict the variables. On top of it Multiple Linear Regression model is used to achieve the end results. The performance of the monitoring model is promising. The approach described in this paper is a general model and not specific to this component, but also can be used for other sensors too as it has a universal kind of approach.
16
Della, Penna Roberto. "Edge Cloud Computing Middleware per Predictive Maintenance in Industria 4.0." Master's thesis, Alma Mater Studiorum - Università di Bologna, 2020. http://amslaurea.unibo.it/20065/.
Full textAbstract:
Il paradigma Edge Cloud Computing ha assunto, negli ultimi anni, un ruolo sempre più importante, soprattutto nel contesto di Industria 4.0, in cui la nascita di nuovi modelli di business basati sull’acquisizione e l’analisi di dati si sono scontrati con le limitazioni imposte dal modello architetturale classico introdotto dal paradigma Cloud Computing. In particolare, la latenza, il consumo di banda, e la dipendenza da un’infrastruttura centralizzata sono stati i fattori abilitanti per l’evoluzione del modello a due livelli, composto da sorgenti dati e infrastruttura cloud, verso l’introduzione di un livello edge intermedio volto ad estendere e distribuire l’infrastruttura ed i servizi cloud nella rete perimetrale. Il presente progetto di tesi ha lo scopo di analizzare e confrontare due piattaforme di Edge Computing, una proprietaria, Azure IoT Edge, ed una open source, KubeEdge, dal punto di vista delle funzionalità offerte e delle performance in uno scenario applicativo che risponde ad un caso di studio reale: lo sviluppo e l’applicazione in tempo reale di modelli di prognosi su un centro di lavoro di un impianto produttivo Bonfiglioli, volti a predire la vita utile residua di uno o più componenti (Predictive Maintenance). Ad una analisi delle due piattaforme, seguirà l’implementazione di un Middleware che, sulla base di esse, fornisce il supporto al deployment e all’aggiornamento dinamico dei componenti dell’applicazione, ed estende le funzionalità di gestione dell’infrastruttura edge attraverso logiche di orchestrazione dinamica dei singoli componenti applicativi volte a garantire il bilanciamento del carico tra i nodi del cluster. I risultati sperimentali mostrano i benefici portati dal supporto di orchestrazione fornito.
17
Lee, Hock Guan. "A study on predictive analytics application to ship machinery maintenance." Thesis, Monterey California. Naval Postgraduate School, 2013. http://hdl.handle.net/10945/37659.
Full textAbstract:
Approved for public release; distribution is unlimitedEngine failures on ships are expensive, and affect operational readiness critically due to long turn-around times for maintenance. Prior to the engine failures, there are signs of engine characteristic changes, for example, exhaust gas temperature (EGT), to indicate that the engine is acting abnormally. This is used as a precursor towards the modeling of failures. There is a threshold limit of 520 degree Celsius for the EGT prior to the need for human intervention. With this knowledge, the use of time series forecasting technique, to predict the crossing over of threshold, is appropriate to model the EGT as a function of its operating running hours and load. This allows maintenance to be scheduled just in time. When there is a departure of result from the predictive model, Cumulative Sum (CUSUM) Control charts can then be used to monitor the change early before an actual problem arises. This paper discusses and demonstrates the proof of principle for one engine and a particular operating profile of a commercial vessel with the use of predictive analytics. The realization with time series forecasting coupled with CUSUM control chart allows this approach to be extended to other attributes beyond EGT.
18
Lindström, Johan. "Predictive maintenance for a wood chipper using supervised machine learning." Thesis, Linköpings universitet, Institutionen för datavetenskap, 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-149304.
Full textAbstract:
With a predictive model that can predict failures of a manufacturing machine, many benefits can be obtained. Unnecessary downtime and accidents can be avoided. In this study a wood chipper which has 12 replaceable knives was examined. The specific task was to create a predictive model that can predict if a knife change is needed or not. To create a predictive model, supervised machine learning was used. Decision forest was the algorithm used in this study. Data samples were collected from vibration measurements. Each sample was labeled with help of ocular inspections of the knives. Microsoft Azure learning studio was the workspace used to train all models. The data set acquired consist of 106 samples, were only 9 samples belongs to the minority class. Two strategies of training a model were used, with and without oversampling. The result for the best model without oversampling obtained 87.5% precision and 77.8% recall. The best model with oversampling achieved 79% precision and 86.7% recall. This result indicates that the trained models can be useful. However, the validity of the result has been hurt by a small data set and many uncertainness of acquiring the data set.
19
Wang, Katherine(Katherine Yuchen). "A machine learning framework for predictive maintenance of wind turbines." Thesis, Massachusetts Institute of Technology, 2020. https://hdl.handle.net/1721.1/129927.
Full textAbstract:
Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, February, 2020Cataloged from student-submitted PDF of thesis.
Includes bibliographical references (pages 73-75).
Wind energy is one of the fastest growing energy sources in the world. However, the failure to detect the breakdown of turbine parts can be very costly. Wind energy companies have increasingly turned to machine learning to improve wind turbine reliability. Thus, the goal of this thesis is to create a flexible and extensible machine learning framework that enables wind energy experts to define and build models for the predictive maintenance of wind turbines. We contribute two libraries that provide experts with the necessary tools to solve prediction problems in the wind energy industry. The first is GPE, which translates and uses the desired prediction problem to generate machine learning training examples from turbine operations data. The other library, CMS-ML, provides the architecture for building machine learning models using vibration data generated by turbine sensors within the Condition Monitoring System (CMS). With this architecture, we can easily create modular feature engineering and machine learning pipelines for the CMS signal data. Finally, we demonstrate the application of these two libraries on proprietary wind turbine data and analyze the effects of their parameters.
by Katherine Wang.
M. Eng.
M.Eng. Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science
20
Shahi, Durlabh, and Ankit Gupta. "Forecasting Components Failure Using Ant Colony Optimization For Predictive Maintenance." Thesis, Högskolan i Halmstad, Akademin för informationsteknologi, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:hh:diva-42457.
Full textAbstract:
Failures are the eminent aspect of any machine and so is true for vehicle as it is one of the sophisticated machines of today’s time. Early detection of faults and prioritized maintenance is a necessity of vehicle manufactures as it enables them to reduce maintenance cost and increase customer satisfaction. In our research, we have proposed a method for processing Logged Vehicle Data (LVD) that uses Ant-Miner algorithm which is a Ant Colony Optimization (ACO) based Algorithm. It also utilizes processes like Feature engineering, Data preprocessing. We tried to explore the effectiveness of ACO for solving classification problem in the form of fault detection and prediction of failures which would be used for predictive maintenance by manufacturers. From the seasonal and yearly model that we have created, we have used ACO to successfully predict the time of failure which is the month with highest likelihood of failure in vehicle’s components. Here, we also validated the obtained results. LVD suffers from data imbalance problem and we have implemented balancing techniques to eliminate this issue, however more effective balancing techniques along with feature engineering is required to increase accuracy in prediction.
21
Dinh, Duc-Hanh. "Opportunistic Predictive Maintenance for Multi-Component Systems with Multiple Dependences." Electronic Thesis or Diss., Université de Lorraine, 2021. http://www.theses.fr/2021LORR0171.
Full textAbstract:
Les dépendances (économiques, stochastiques et/ou structurelles) entre composants influencent de manière significative le processus de dégradation des composants ainsi que le processus de prise de décision en maintenance. En ce sens, la non prise en compte des dépendances entre composants dans la modélisation de la maintenance pourrait entraîner des surcoûts de maintenance et un planning de maintenance sous-optimal. En lien avec ces considérations de nombreux travaux en maintenance prédictive de systèmes multi-composants avec des dépendances entre composants ont été récemment faits. Cependant, la plupart des modèles de maintenance prédictive existants ne permettent de prendre en compte qu'un seul type de dépendances, car la considération de plusieurs dépendances entraîne une complexité plus importante lors de la modélisation de la dégradation mais aussi la formalisation des processus de décision et d’optimisation de la maintenance. Cependant, dans les cas réels de systèmes industriels, plusieurs types de dépendances peuvent exister ensemble, notamment les dépendances économiques et structurelles. Par exemple, la plupart des systèmes mécaniques sont construits sur une structure hiérarchique impliquant que la maintenance d'un composant nécessite le démontage d'autres composants. L’objectif de cette thèse est donc d’intégrer à la fois des dépendances économiques et structurelles dans le processus de modélisation de la dégradation et le processus de décision en maintenance d'un système à composants multiples dans le cadre de la maintenance prédictive. Plus précisément, cet objectif repose sur deux axes scientifiques majeurs. Le premier consiste à étudier l'impact des dépendances structurelles et économiques sur le processus de dégradation des composants et sur la structure des coûts de maintenance. Le deuxième axe de recherche a pour objet d’intégrer les impacts des dépendances économiques et structurelles dans les processus de décision et d'optimisation de la maintenance. Face à ces problématiques, dans cette thèse nous avons proposé trois contributions principales : (1)-Formalisation et proposition de modèles mathématiques permettant de modéliser les dépendances structurelles et économiques entre composants; (2)-Développement d'un modèle de dégradation considérant les impacts de la dépendance structurelle entre composants; (3)-Développement d'une politique de maintenance prédictive opportuniste adaptée permettant de prendre en considération les impacts des dépendances économiques et structurelles dans les processus de prise de décision et d'optimisation de la maintenance. Enfin, pour évaluer la faisabilité et la valeur ajoutée ainsi que les limites des modèles proposées dans un cadre d'optimisation de la maintenance, une étude numérique sur un convoyeur industriel est investiguéeRecently, maintenance modeling for multi-component systems with dependences (economic, stochastic, and/or structural dependences) has been extensively studied. However, most of the existing studies only consider one type of dependence since combining more than one makes the models too complicated to analyze and solve. However, in practice, several types of dependences, especially, the economic and structural dependences, may exist together in the system. To face this issue, the main objective of this thesis is to consider both economic and structural dependences in maintenance modeling and optimization for multi-component systems in framework of predictive maintenance. For this purpose, the impacts of economic and structural dependences on the maintenance cost, duration and the degradation process of the components are firstly investigated. Mathematical models for quantifying the impacts of the economic and structural dependences are then developed. Finally, a multi-level opportunistic maintenance policy is proposed to consider the impacts of these dependences between components.Due to the structural dependence between components, when a maintenance (preventive or corrective action) occurs, only few components need to be disassembled. The disassembled components are subjected to both economic and structural dependences while the non-disassembled components are subjected to only economic dependence. In that way, the proposed maintenance policy is characterized by one preventive threshold, that is used to select survival components for preventive maintenance, and two opportunistic maintenance thresholds, that are used for opportunistic maintenance. When a maintenance occurs, the first opportunistic threshold is defined to select the non-disassembled components (with only economic dependence) while the second opportunistic threshold is then developed to consider the disassembled components for opportunistic maintenance (with both economic and structural dependences). To evaluate the performance of the proposed opportunistic maintenance policy, a cost model is developed. Particle swarm optimization algorithm is then implemented to find the optimal decision variables. Finally, the proposed opportunistic maintenance policy is illustrated through a conveyor system to show its feasibility and added value in maintenance optimization framework
22
Le, Nguyen Minh Huong. "Online machine learning-based predictive maintenance for the railway industry." Electronic Thesis or Diss., Institut polytechnique de Paris, 2023. http://www.theses.fr/2023IPPAT027.
Full textAbstract:
En tant que moyen de transport en commun efficace sur de longues distances, le chemin de fer continuera de prospérer pour son empreinte carbone limitée dans l'environnement. Assurer la fiabilité des équipements et la sécurité des passagers fait ressortir la nécessité d'une maintenance efficace. Outre la maintenance corrective et périodique courante, la maintenance prédictive a pris de l'importance ces derniers temps. Les progrès récents de l'apprentissage automatique et l'abondance de données poussent les praticiens à la maintenance prédictive basée sur les données. La pratique courante consiste à collecter des données pour former un modèle d'apprentissage automatique, puis à déployer le modèle pour la production et à le conserver inchangé par la suite. Nous soutenons qu'une telle pratique est sous-optimale sur un flux de données. Le caractère illimité du flux rend le modèle sujet à un apprentissage incomplet. Les changements dynamiques sur le flux introduisent de nouveaux concepts invisibles pour le modèle et diminuent sa précision. La vitesse du flux rend l'étiquetage manuel impossible et désactive les algorithmes d'apprentissage supervisé. Par conséquent, il est nécessaire de passer d'un paradigme d'apprentissage statique et hors ligne à un paradigme adaptatif en ligne, en particulier lorsque de nouvelles générations de trains connectés générant en continu des données de capteurs sont déjà une réalité. Nous étudions l'applicabilité de l'apprentissage automatique en ligne pour la maintenance prédictive sur des systèmes complexes typiques du secteur ferroviaire. Tout d'abord, nous développons InterCE en tant que framework basé sur l'apprentissage actif pour extraire des cycles d'un flux non étiqueté en interagissant avec un expert humain. Ensuite, nous implémentons un auto-encodeur à mémoire longue et courte durée pour transformer les cycles extraits en vecteurs de caractéristiques plus compacts tout en restant représentatifs. Enfin, nous concevons CheMoc comme un framework pour surveiller en permanence l'état des systèmes en utilisant le clustering adaptatif en ligne. Nos méthodes sont évaluées sur les systèmes d'accès voyageurs sur deux flottes de trains gérés par la société nationale des chemins de fer SNCF de la FranceBeing an effective long-distance mass transit, the railway will continue to flourish for its limited carbon footprint in the environment. Ensuring the equipment's reliability and passenger safety brings forth the need for efficient maintenance. Apart from the prevalence of corrective and periodic maintenance, predictive maintenance has come into prominence lately. Recent advances in machine learning and the abundance of data drive practitioners to data-driven predictive maintenance. The common practice is to collect data to train a machine learning model, then deploy the model for production and keep it unchanged afterward. We argue that such practice is suboptimal on a data stream. The unboundedness of the stream makes the model prone to incomplete learning. Dynamic changes on the stream introduce novel concepts unseen by the model and decrease its accuracy. The velocity of the stream makes manual labeling infeasible and disables supervised learning algorithms. Therefore, switching from a static, offline learning paradigm to an adaptive, online one is necessary, especially when new generations of connected trains continuously generating sensor data have already been a reality. We investigate the applicability of online machine learning for predictive maintenance on typical complex systems in the railway. First, we develop InterCE as an active learning-based framework that extracts cycles from an unlabeled stream by interacting with a human expert. Then, we implement a long short-term memory autoencoder to transform the extracted cycles into feature vectors that are more compact yet remain representative. Finally, we design CheMoc as a framework that continuously monitors the condition of the systems using online adaptive clustering. Our methods are evaluated on the passenger access systems on two fleets of passenger trains managed by the national railway company SNCF of France
23
Koeneman, Peter William. "An analysis of sensor effectiveness to inform a predictive maintenance policy." Thesis, Monterey, Calif. : Naval Postgraduate School, 2009. http://handle.dtic.mil/100.2/ADA502234.
Full textAbstract:
Thesis (M.S. in Operations Research)--Naval Postgraduate School, June 2009.Thesis Advisor(s): Jacobs, P. A. ; Gaver, D. P. "June 2009." Description based on title screen as viewed on Ju;ly 10, 2009. DTIC Identifiers: CBM (Condition Based Maintenance), warning time of impending failure, renewal reward process, failure time, maintenance models, predictive maintenance. Author(s) subject terms: Condition Based Maintenance ; CBM+; renewal reward process; correlated failure time and warning time of impending failure. Includes bibliographical references (p. 77-78). Also available in print.
24
Burrows, John H. (John Henry). "Predictive and preventive maintenance of mobile mining equipment using vibration data." Thesis, McGill University, 1996. http://digitool.Library.McGill.CA:80/R/?func=dbin-jump-full&object_id=24052.
Full textAbstract:
This thesis discusses approaches to evaluate the health of mining machinery, based on monitored vibration data. The objective was to develop a means to determine machine health, while operating on-line, without reference to an expert. This approach is based on processing acquired vibration data with artificial neural networks (ANN's). A case study, based on data obtained from the monitoring of locomotives at the Iron Ore Company (IOCC). Real time data patterns, profiles and trends, obtained by processing vibration signals from various points on locomotives, were used to test the developed technique. The results indicate that observed patterns and trends can be classified into categories that reliably indicate the mechanical state of the equipment. An implemented system will assist maintenance personnel at this mine to identify the trends of a developing component problem in advance of catastrophic failure. In addition the system will be able to predict its remaining life prior to catastrophic failure. Thus, a machine could be reliably and safely operated until just prior to failure of a component.The thesis work is a sub-component of a larger project at IOCC, to implement a mine-wide predictive/preventative maintenance program for pumps, locomotives, trucks, shovels and drills at their open-pit mine in Labrador City, Newfoundland. This system will use intermittent on- and off-line, condition monitoring based on ANNs and expert systems (ES). A functional overview is discussed. The data would identify where and what is the particular machine alarm condition. Such an approach would allow improved fault detection of machine components, especially in mines where trained personnel are not readily available. (Abstract shortened by UMI.)
25
Ruiz, Cárcel Cristóbal. "Predictive condition monitoring of industrial systems for improved maintenance and operation." Thesis, Cranfield University, 2014. http://dspace.lib.cranfield.ac.uk/handle/1826/9305.
Full textAbstract:
Maintenance strategies based on condition monitoring of the different machines
and devices in an industrial process can minimize downtime, increase the
safety of plant operations and help in the process of decision-taking for control
and maintenance actions in order to reduce maintenance and operating costs.
Multivariate statistical methods are widely used for process condition monitoring
in modern industrial sites due to the quantity of data available and the difficulties
of building analytical models in complex facilities.
Nevertheless, the performance of these methodologies is still far away from
being ideal, due to different issues such as process nonlinearities or varying
operational conditions. In addition application of the latest approaches
developed for process monitoring is not widely extended in real industry.
The aim of this investigation is to develop new and improve existing
methodologies for predictive condition monitoring through the use of
multivariate statistical methods. The research focuses on demonstrating the
applicability of multivariate algorithms in real complex cases, the improvement
of these methods in terms of fault detection and diagnosis by means of data
fusion and the estimation of process performance degradation caused by faults.Marie Curie
26
Akindele, Babatunde Babajide. "Model-based fault diagnosis framework for effective predictive maintenance / B.B. Akindele." Thesis, North-West University, 2010. http://hdl.handle.net/10394/4435.
Full textAbstract:
Predictive maintenance is a proactive maintenance strategy that is aimed at preventing the unexpected failure of equipment through condition monitoring of the health and performance of the equipment.
Incessant equipment outage resulting in low availability of production facilities is a major issue in the Nigerian manufacturing environment. Improving equipment availability in Nigeria industry through institution of a full featured predictive maintenance has been suggested by many authors. The key to instituting a full-featured predictive maintenance is condition monitoring.
Primarily, this research is focused on how to reduce the prevalent of equipment downtime in the Nigerian manufacturing industry, through the application of Model Based Fault Diagnosis technology as a condition monitoring tool for enhancing predictive maintenance practices in Nigerian manufacturing industry.
The following objectives underscore the aim of this research work:
* To assess the implementation and performance of predictive maintenance practices in some selected manufacturing companies in Nigeria and verify if there is need for improvement in these practices.
* To identify the challenges and barriers to the implementation and performance of full-featured predictive maintenance practice in the Nigerian manufacturing industry.
* To develop a framework for enhancing quality of Predictive Maintenance practices in the manufacturing industry in Nigeria through a Model Based Fault Diagnosis and Decision Support System.
* To validate that the developed framework meets the Nigerian manufacturing industry needs through the implementation of a prototype in one of the selected manufacturing companies in the case study.
The empirical investigation undertaken as part of this research revolves around five (5) of the Nigerian manufacturing companies. Personal interviews were also adopted as means of data collection.
The research outcomes reveal the followings:
* Top management commitment to the implementation of predictive maintenance strategies in the Nigerian manufacturing industry is inadequate.
* Many of the manufacturing companies lack a tool for carrying out continuous condition monitoring in their predictive maintenance program. This is responsible for poor performance of most predictive maintenance programs in Nigerian manufacturing industry.
* Inadequate training on the implementation of predictive maintenance principles is adversely affecting the proficiency of personnel in adopting philosophy that underlies practices of predictive maintenance.
* The size of equipment part inventory, maintenance work backlog and machine scraps are also enormous in the maintenance yards of the companies.
* Nevertheless, the implementation of predictive maintenance program has a positive impact on the equipment availability of one of the case studies. Management commitment in Chemical and Allied Products (CAP) Plc is outstanding. Application of intelligent condition monitoring system, and personnel training and competence are vital to the success of Predictive Maintenance implementation in CAP Plc.
The specific deliverable from this research is a proposed framework (MBFDF) for effective implementation of predictive maintenance strategy through application of model based fault diagnosis technology, which can be adopted to improve performance of predictive maintenance practices in the Nigerian manufacturing industry.
The deliverable also includes a soft copy of data in Excel spreadsheet obtained during experimental test of the proposed framework in a small manufacturing company in Nigeria.
In this research, a model based fault diagnosis framework (MBFDF) to serve as a condition monitoring and decision support tool for predictive maintenance programs in Nigerian manufacturing industry was developed. Implementation to verify the real-life implementability and effectiveness of the proposed framework was performed in one of the companies used for the case study.
A comparison of results with pre-integration predictive maintenance program is presented, showing the implementability and the effectiveness of the proposed MBFDF for condition monitoring in predictive maintenance programs in the Nigerian manufacturing company.
Recommendations presented in this dissertation are also vital to the success of implementing predictive maintenance program in Nigerian manufacturing companies.Thesis (M.Ing. (Development and Management))--North-West University, Potchefstroom Campus, 2011.
27
Flint, Anthony David. "The development of predictive maintenance systems based on the Hough transform." Thesis, University of Huddersfield, 1995. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.307836.
Full text
28
Ye, Chen S. M. Massachusetts Institute of Technology. "A system approach to implementation of predictive maintenance with machine learning." Thesis, Massachusetts Institute of Technology, 2018. http://hdl.handle.net/1721.1/118502.
Full textAbstract:
Thesis: S.M. in Engineering and Management, Massachusetts Institute of Technology, System Design and Management Program, 2018.Cataloged from PDF version of thesis.
Includes bibliographical references (pages 87-91).
Digital technology is changing the industrial sector, yet how to make rational use of some technologies and create considerable value in a variety of industrial scenarios is an issue. Many digital industrial companies have stated that they have helped clients with their digital transformation, create much value, but the real effects have not been shown in public. Venture capitals firms have made huge investment in potential digital industrial startups. Numerous industrial IoT platforms are emerging in the market, but a number of them fade soon after. Many people have heard about industrial maintenance technology, but they have difficulty in differentiate concepts such as reactive maintenance, planned maintenance, proactive maintenance, and predictive maintenance. Many people know that big data and Al are essential in industrial sector, but they do not know how to process, analyze, and extract value from industrial data and how to use Al algorithms and tools to implement a research project. This thesis analyzes the entire digital industrial ecosystem in various dimensions such as initiatives, technologies in related domains, stakeholders, markets, and strategies. This work also analyzes of the predictive maintenance solution in various dimensions such as background, importance, suitable scenarios, market, business model, and technology. The author plans an experiment for the predictive maintenance solution, including goal, data source and description, methods and steps, and flow and tools. Then author uses a baseline approach and an optimal approach to implement the experiment, including data preparation, selection and evaluation of both regression and classification models, and deep learning practice through neural network building and optimization. Finally, contributions and expectations, and limitations and future research are discussed. This work uses a system approach, including system architecting, system engineering, and project management, to complete the process of analysis, design, and implementation.
by Chen Ye.
S.M. in Engineering and Management
29
Numanovic, Kerim. "Advanced Clinical Data Processing: A Predictive Maintenance Model for Anesthesia Machines." Thesis, KTH, Tillämpad fysik, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-283323.
Full textAbstract:
The maintenance of medical devices is of great importance to ensure that the devices are stable, well-functioning, and safe to use. The current method of maintenance, which is called preventive maintenance, has its advantages but can be problematic both from an operators’ and a manufacturers’ side. Developing a model that will predict failure in anesthesia machines can be of great use for the manufacturer, the customers, and the patients. This thesis sets to examine the possibility of creating a predictive maintenance model for anesthesia machines by utilizing device data and machine learning. This thesis also investigates the influence of the data on the model performance and compare different lag sizes and future horizons to model performance. The time-series data collected came from 87 unique devices and a specific test was chosen to be the output variable of the model. A whole pipeline was created, which included pre-processing of the data, feature engineering, and model development. Feature extraction was done on the time series data, with the help of a library called tsfresh, which transformed time series characteristics into features that would enable supervised learning. Two models were developed: logistic regression and XGBoost. The logistic regression model acted as a baseline model and the result of its performance was as expected, quite poor. The XGBoost yielded an AUCPR score of 0.21 on the full dataset and 0.32 on a downsampled dataset. Although a quite low score, it was surprisingly high considering the extreme class imbalance that existed in the dataset. No clear pattern was found between the lag sizes and future horizons with the model performance. Something that could be seen was that the data imbalance had a great impact on the model performance, which was discovered when the downsampled dataset with less class imbalance yielded a higher AUCPR score.Underhållet av medicintekniska produkter är mycket viktigt för att säkerställa att enheterna är stabila, välfungerande och säkra att använda. Den nuvarande underhållsmetoden, som kallas förebyggande underhåll, har sina fördelar men kan vara problematisk både från operatörens och tillverkarsidan. Att utveckla en modell som förutsäger fel i anestesimaskiner kan vara till stor nytta för tillverkaren, kunderna och patienterna. Denna avhandling syftar till att undersöka möjligheten att skapa en förutsägbar underhållsmodell för anestesimaskiner genom att använda enhetsdata och maskininlärning. Denna avhandling undersöker också påverkan av data på modellprestanda och jämför olika fördröjningsstorlekar och framtida horisonter med modellprestanda. Tidsseriedata som samlats in kom från 87 unika enheter och ett specifikt test valdes för att vara modellens outputvariabel. En hel pipeline skapades, som inkluderade förbehandling av data, funktionsteknik och modellutveckling. Funktionsextraktion gjordes på tidsseriedata med hjälp av ett bibliotek som heter tsfresh, som förvandlade tidsserieegenskaper till funktioner som skulle möjliggöra övervakat lärande. Två modeller utvecklades: logistisk regression och XGBoost. Den logistiska regressionsmodellen fungerade som en basmodell och resultatet av dess prestanda var som förväntat ganska dåligt. XGBoost gav en AUCPR-poäng på 0,21 på hela datamängden och 0,32 på en nedmonterad datamängd. Även om det var en ganska låg poäng, var det överraskande högt med tanke på den extrema klassobalansen som fanns i datasetet. Inget tydligt mönster hittades mellan fördröjningsstorlekarna och framtida horisonter med modellprestanda. Något som kunde ses var att dataobalansen hade stor inverkan på modellens prestanda, vilket upptäcktes när den nedprovade datamängden med mindre obalans i klassen gav en högre AUCPR-poäng.
30
Bergström, Joakim. "Transfer Learning on Ultrasound Spectrograms of Weld Joints for Predictive Maintenance." Thesis, Uppsala universitet, Signaler och system, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-424726.
Full textAbstract:
A big hurdle for many companies to start using machine learning is that trending techniques need a huge amount of structured data. One potential way to reduce the need for data is taking advantage of previous knowledge from a related task. This is so called transfer learning. A basic description of it would be when you take a model trained on existing data and reuse that for another problem. The purpose of this master thesis is to investigate if transfer learning can reduce the need for data when faced with a new machine learning task which is, in particular, to use transfer learning on ultrasound spectrograms of weld joints for predictive maintenance. The base for transfer learning is VGGish, a convolutional neural network model trained on audio samples collected from YouTube videos. The pre-trained weights are kept, and the prediction layer is replaced with a new prediction layer consisting of two neurons. The whole model is re-trained on the ultrasound spectrograms. The dataset is restricted to a minimum of ten and a maximum of 100 training samples. The results are evaluated and compared to a regular convolutional neural network trained on the same data. The results show that transfer learning improves the test accuracy compared to the regular convolutional neural network when the dataset is small. This thesis project concludes that transfer learning can reduce the need for data when faced with a new machine learning task. The results indicate that transfer learning could be useful in the industry.
31
Nordberg, Andreas. "Evaluation of Neural Networks for Predictive Maintenance : A Volvo Penta Study." Thesis, Linköpings universitet, Institutionen för datavetenskap, 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-176390.
Full textAbstract:
As part of Volvo Penta's initiative to further the development of predictive maintenance in their field test environments, this thesis compares neural networks in an effort to predict the occurrence of three common diagnostics trouble codes using field test data. To quantify the neural networks' performances for comparison a number of evaluation metrics were used. By training a multitude of differently configured feedforward neural networks with the processed field test data and evaluating the resulting models, it was found that the resulting models perform better than that of a baseline classifier. As such it is possible to use Volvo Penta's field test data along with neural networks to achieve predictive maintenance. It was also found that Long Short-Term Memory (LSTM) networks with methodically selected hyperparameters were able to predict the diagnostic trouble codes with the greatest performance among all the tested neural networks.
32
Hedkvist, Adam. "Predictive maintenance with machine learning on weld joint analysed by ultrasound." Thesis, Uppsala universitet, Avdelningen för beräkningsvetenskap, 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-396059.
Full textAbstract:
Ever since the first industrial revolution industries have had the goal to increase their production. With new technology such as CPS, AI and IoT industries today are going through the fourth industrial revolution denoted as industry 4.0. The new technology not only revolutionises production, but also maintenance, making predictive maintenance possible. Predictive maintenance seeks to predict when failure would occur, instead of having scheduled maintenance or maintenance after failure already occurred. In this report a convolutional neural network (CNN) will analyse data from an ultrasound machine scanning a weld joint. The data from the ultrasound machine will be transformed by the short time Fourier transform in order to create an image for the CNN. Since the data from the ultrasound is not complete, simulated data will be created and investigated as another option for training the network. The results are promising, however the lack of data makes it hard to show any concrete proof.
33
LAKSHMANAN, KAYALVIZHI. "Predictive Maintenance of an External Gear Pump using Machine Learning Algorithms." Doctoral thesis, Università degli studi di Pavia, 2021. http://hdl.handle.net/11571/1447613.
Full textAbstract:
The thesis describes a novel computational strategy of Predictive Maintenance (fault diagnosis
and fault prognosis) with ML and Deep Learning applications for an FG304 series external gear
pump, also known as a domino pump. Due to the unavailability of a sufficient amount of
experimental data, a novel approach of generating a high-fidelity in-silico dataset via a
Computational Fluid Dynamic model of the gear pump in a healthy and various faulty working
conditions (e.g., clogging, radial gap variations, viscosity variations, etc.). The synthetic data
generation technique is implemented by perturbing the frequency content of the time series to
recreate other environmental conditions. These synthetically generated datasets are used to train
the underlying ML metamodel. In addition, various types of feature extraction methods considered
to extract the most discriminatory information from the data. For fault diagnosis, three types of ML
classification algorithms are employed, namely Multilayer Perceptron (MLP), Support Vector
Machine (SVM) and Naive Bayes algorithms. For prognosis, ML regression algorithms, such as
MLP and SVM, are utilised. Hyper-parameters of the ML algorithms is optimised with a staggered
approach. In addition, a real case study of fault diagnosis and fault prognosis of an external gear
pump considering noisy measurements to understand the sensitivity of the employed ML
algorithms by adding noise on the training dataset and test dataset. A series of numerical examples
are presented, enabling us to conclude that for fault diagnosis, the use of wavelet features and a
MLP algorithm can provide the best accuracy and for fault prognosis, the use of MLP algorithm
provides the best prediction results.
34
CONSILVIO, ALICE. "Risk-Based Optimal Scheduling for the Predictive Maintenance of Railway Infrastructure." Doctoral thesis, Università degli studi di Genova, 2018. http://hdl.handle.net/11567/929594.
Full textAbstract:
In this thesis a risk-based decision support system to schedule the predictive maintenance activities, is proposed. The model deals with the maintenance planning of a railway infrastructure in which the due-dates are defined via failure risk analysis.The novelty of the approach consists of the risk concept introduction in railway maintenance scheduling, according to ISO 55000 guidelines, thus implying that the maintenance priorities are based on asset criticality, determined taking into account the relevant failure probability, related to asset degradation conditions, and the consequent damages.
35
Oliveira, Rafael José Gomes de [UNESP]. "Implementação de técnicas de processamento de sinais para o monitoramento da condição de mancais de rolamento." Universidade Estadual Paulista (UNESP), 2005. http://hdl.handle.net/11449/97131.
Full textAbstract:
Made available in DSpace on 2014-06-11T19:28:35Z (GMT). No. of bitstreams: 0
Previous issue date: 2005-05Bitstream added on 2014-06-13T20:58:36Z : No. of bitstreams: 1
oliveira_rjg_me_guara.pdf: 1311511 bytes, checksum: 7c57fbdb099a4b3d6123bb38b37813e3 (MD5)Universidade Estadual Paulista (UNESP)
Na indústria moderna o monitoramento da condição de operação de máquinas rotativas é essencial para se determinar o surgimento de falhas em mancais de rolamentos. Este trabalho apresenta uma técnica de análise adotada para a identificação de falhas em mancais de rolamento em seus estágios iniciais, utilizando procedimentos de análise de sinais no domínio do tempo e da freqüência, com especial atenção para a técnica do HFRT (High Frequency Resonance Technique), também conhecida como Técnica do Envelope. Este método de análise de sinais foi escolhido em razão de ser uma ferramenta apropriada para identificar falhas em mancais de rolamentos na sua fase inicial. A teoria das técnicas foi discutida e os passos para a implementação computacional foram apresentados. As rotinas foram implementadas através da linguagem de programação MATLAB e um sinal simulado representativo de um sinal coletado de um mancal de rolamento com defeito pontual na pista externa foi desenvolvido para verificar a eficácia dos métodos implementados. Os experimentos foram desenvolvidos utilizando-se uma bancada de testes aplicada para testar mancais de rolamento com defeitos pontuais produzidos em laboratório. A aquisição dos dados foi desenvolvida com instrumentação comercial. Os resultados obtidos mostraram ser efetivos para identificar falhas em rolamentos para os dados simulados e dados experimentais.
In the modern industries, the condition monitoring of the rotational machinery operation is important to evidence the beginning of the fails in bearings. This work presents a technique of analysis applied to identify fails in bearing during the initial phases, using techniques of signal analysis in time and frequency domain with special attention for the High Frequency Resonance Technique, also called envelope technique. This method for signal analysis was chosen because is an appropriated tool to identify fails in bearings during initial phases. The theory for the techniques was discussed and the steps for the computational implementation were showed. The routines were implemented through MATLAB programming language and it was prepared a representative signal of a bearing with a single point defect in the outer race in order to verify the capability of the method implemented in the routine. The experiments were performed using a experimental test rig applied to test bearings with single point defects performed in laboratory. The data acquisition were performed with commercial instrumentation. The results obtained shown to be effective to identify fails in bearings for both numerically simulated data and experimental data.
36
Ayandokun, O. K. "The incremental motion encoder : a sensor for the integrated condition monitoring of rotating machinery." Thesis, Nottingham Trent University, 1997. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.245075.
Full text
37
Rossi, Tisbeni Simone. "Big data analytics towards predictive maintenance at the INFN-CNAF computing centre." Master's thesis, Alma Mater Studiorum - Università di Bologna, 2019. http://amslaurea.unibo.it/18430/.
Full textAbstract:
La Fisica delle Alte Energie (HEP) è da lungo tra i precursori nel gestire e processare enormi dataset scientifici e nell'operare alcuni tra i più grandi data centre per applicazioni scientifiche.
HEP ha sviluppato una griglia computazionale (Grid) per il calcolo al Large Hadron Collider (LHC) del CERN di Ginevra, che attualmente coordina giornalmente le operazioni di calcolo su oltre 800k processori in 170 centri di calcolo e gestendo mezzo Exabyte di dati su disco distribuito in 5 continenti.
Nelle prossime fasi di LHC, soprattutto in vista di Run-4, il quantitativo di dati gestiti dai centri di calcolo aumenterà notevolmente.
In questo contesto, la HEP Software Foundation ha redatto un Community White Paper (CWP) che indica il percorso da seguire nell'evoluzione del software moderno e dei modelli di calcolo in preparazione alla fase cosiddetta di High Luminosity di LHC. Questo lavoro ha individuato in tecniche di Big Data Analytics un enorme potenziale per affrontare le sfide future di HEP.
Uno degli sviluppi riguarda la cosiddetta Operation Intelligence, ovvero la ricerca di un aumento nel livello di automazione all'interno dei workflow. Questo genere di approcci potrebbe portare al passaggio da un sistema di manutenzione reattiva ad uno, più evoluto, di manutenzione predittiva o addirittura prescrittiva.
La tesi presenta il lavoro fatto in collaborazione con il centro di calcolo dell'INFN-CNAF per introdurre un sistema di ingestione, organizzazione e processing dei log del centro su una piattaforma di Big Data Analytics unificata, al fine di prototipizzare un modello di manutenzione predittiva per il centro. Questa tesi contribuisce a tale progetto con lo sviluppo di un algoritmo di clustering dei messaggi di log basato su misure di similarità tra campi testuali, per superare il limite connesso alla verbosità ed eterogeneità dei log raccolti dai vari servizi operativi 24/7 al centro.
38
Faraj, Dina. "Using Machine Learning for Predictive Maintenance in Modern Ground-Based Radar Systems." Thesis, KTH, Matematisk statistik, 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-299634.
Full textAbstract:
Military systems are often part of critical operations where unplanned downtime should be avoided at all costs. Using modern machine learning algorithms it could be possible to predict when, where, and at what time a fault is likely to occur which enables time for ordering replacement parts and scheduling maintenance. This thesis is a proof of concept study for anomaly detection in monitoring data, i.e., sensor data from a ground based radar system as an initial experiment to showcase predictive maintenance. The data in this thesis was generated by a Giraffe 4A during normal operation, i.e., no anomalous data with known failures was provided. The problem setting is originally an unsupervised machine learning problem since the data is unlabeled. Speculative binary labels are introduced (start-up state and steady state) to approximate a classification accuracy. The system is functioning correctly in both phases but the monitoring data looks differently. By showing that the two phases can be distinguished, it is possible to assume that anomalous data during break down can be detected as well. Three different machine learning classifiers, i.e., two unsupervised classifiers, K-means clustering and isolation forest and one supervised classifier, logistic regression are evaluated on their ability to detect the start-up phase each time the system is turned on. The classifiers are evaluated graphically and based on their accuracy score. All three classifiers recognize a start up phase for at least four out of seven subsystems. By only analyzing their accuracy score it appears that logistic regression outperforms the other models. The collected results manifests the possibility to distinguish between start-up and steady state both in a supervised and unsupervised setting. To select the most suitable classifier, further experiments on larger data sets are necessary.Militära system är ofta en del av kritiska operationer där oplanerade driftstopp bör undvikas till varje pris. Med hjälp av moderna maskininlärningsalgoritmer kan det vara möjligt att förutsäga när och var ett fel kommer att inträffa. Detta möjliggör tid för beställning av reservdelar och schemaläggning av underhåll. Denna uppsats är en konceptstudie för detektion av anomalier i övervakningsdata från ett markbaserat radarsystem som ett initialt experiment för att studera prediktivt underhåll. Datat som används i detta arbete kommer från en Saab Giraffe 4A radar under normal operativ drift, dvs. ingen avvikande data med kända brister tillhandahölls. Problemställningen är ursprungligen ett oövervakat maskininlärningsproblem eftersom datat saknar etiketter. Spekulativa binära etiketter introduceras (uppstart och stabil fas) för att uppskatta klassificeringsnoggrannhet. Systemet fungerar korrekt i båda faserna men övervakningsdatat ser annorlunda ut. Genom att visa att de två faserna kan urskiljas, kan man anta att avvikande data också går att detektera när fel uppstår. Tre olika klassificeringsmetoder dvs. två oövervakade maskininlärningmodeller, K-means klustring och isolation forest samt en övervakad modell, logistisk regression utvärderas utifrån deras förmåga att upptäcka uppstartfasen varje gång systemet slås på. Metoderna utvärderas grafiskt och baserat på deras träffsäkerhet. Alla tre metoderna känner igen en startfas för minst fyra av sju delsystem. Genom att endast analysera deras noggrannhetspoäng, överträffar logistisk regression de andra modellerna. De insamlade resultaten demonstrerar möjligheten att skilja mellan uppstartfas och stabil fas, både i en övervakad och oövervakad miljö. För att välja den bästa metoden är det nödvändigt med ytterligare experiment på större datamängder.
39
Bansal, Dheeraj. "An advanced real-time predictive maintenance framework for large scale machine systems." Thesis, Aston University, 2005. http://publications.aston.ac.uk/12235/.
Full textAbstract:
This thesis introduces and develops a novel real-time predictive maintenance system to estimate the machine system parameters using the motion current signature. A crucial concept underpinning this project is that the motion current signature contains information relating to the machine system parameters and that this information can be extracted using nonlinear mapping techniques, such as neural networks. Towards this end, a proof of concept procedure is performed, which substantiates this concept. A simulation model, TuneLearn, is developed to simulate the large amount of training data required by the neural network approach. Statistical validation and verification of the model is performed to ascertain confidence in the simulated motion current signature. Validation experiment concludes that, although, the simulation model generates a good macro-dynamical mapping of the motion current signature, it fails to accurately map the micro-dynamical structure due to the lack of knowledge regarding performance of higher order and nonlinear factors, such as backlash and compliance. Failure of the simulation model to determine the micro-dynamical structure suggests the presence of nonlinearity in the motion current signature. This motivated us to perform surrogate data testing for nonlinearity in the motion current signature. Results confirm the presence of nonlinearity in the motion current signature, thereby, motivating the use of nonlinear techniques for further analysis. Outcomes of the experiment show that nonlinear noise reduction combined with the linear reverse algorithm offers precise machine system parameter estimation using the motion current signature for the implementation of the real-time predictive maintenance system. Finally, a linear reverse algorithm, BJEST, is developed and applied to the motion current signature to estimate the machine system parameters.
40
Jorge, Inès. "Machine-learning-based predictive maintenance for lithium-ion batteries in electric vehicles." Electronic Thesis or Diss., Strasbourg, 2023. http://www.theses.fr/2023STRAD056.
Full textAbstract:
La batterie est un élément central des véhicules électriques, soumis à de nombreux enjeux en termes de performances, sécurité et coût. La durée de vie des batteries en particulier fait l’objet d’une grande attention, car elle doit s’aligner avec la durée de vie d’un véhicule. Dans ce contexte, la maintenance prévisionnelle vise à prédire de manière fiable la durée de vie utile restante (RUL) et l’évolution de l’état de santé (SOH) d'une batterie Lithium-Ion (Li-Ion) en utilisant les données d'utilisation passées et présentes, de manière à anticiper les opérations de maintenance. L’objectif de cette thèse est de tirer profit de l’information contenue dans les séries temporelles de courant, tension et température via des algorithmes d’apprentissage automatique. Plusieurs modèles prédictifs ont étés développés à partir de jeux de données publics, afin de prédire le RUL d’une batterie ou l’évolution de son SOH à plus ou moins long termeThe battery is a central component of electric vehicles, and is subject to numerous challenges in terms of performance, safety and cost. The life of batteries in particular is the subject of a great deal of attention, as it needs to be aligned with the life of a vehicle. In this context, predictive maintenance aims to reliably predict the remaining useful life (RUL) and the evolution of the state of health (SOH) of a Lithium-Ion (Li-Ion) battery using past and present operating data, so as to anticipate maintenance operations. The objective of this thesis is to take advantage of the information contained in the time series of current, voltage and temperature via machine learning algorithms. Several predictive models have been developed from public datasets, in order to predict the RUL of a battery or the evolution of its SOH in the more or less long term
41
Berg, Martin, and Albin Eriksson. "Toward predictive maintenance in surface treatment processes : A DMAIC case study at Seco Tools." Thesis, Luleå tekniska universitet, Institutionen för ekonomi, teknik, konst och samhälle, 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:ltu:diva-84923.
Full textAbstract:
Surface treatments are often used in the manufacturing industry to change the surface of a product, including its related properties and functions. The occurrence of degradation and corrosion in surface treatment processes can lead to critical breakdowns over time. Critical breakdowns may impair the properties of the products and shorten their service life, which causes increased lead times or additional costs in the form of rework or scrapping. Prevention of critical breakdowns due to machine component failure requires a carefully selected maintenance policy. Predictive maintenance is used to anticipate equipment failures to allow for maintenance scheduling before component failure. Developing predictive maintenance policies for surface treatment processes is problematic due to the vast number of attributes to consider in modern surface treatment processes. The emergence of smart sensors and big data has led companies to pursue predictive maintenance. A company that strives for predictive maintenance of its surface treatment processes is Seco Tools in Fagersta. The purpose of this master's thesis has been to investigate the occurrence of critical breakdowns and failures in the machine components of the chemical vapor deposition and post-treatment wet blasting processes by mapping the interaction between its respective process variables and their impact on critical breakdowns. The work has been conducted as a Six Sigma project utilizing the problem-solving methodology DMAIC. Critical breakdowns were investigated combining principal component analysis (PCA), computational fluid dynamics (CFD), and statistical process control (SPC) to create an understanding of the failures in both processes. For both processes, two predictive solutions were created: one short-term solution utilizing existing dashboards and one long-term solution utilizing a PCA model and an Orthogonal Partial Least Squares (OPLS) regression model for batch statistical process control (BSPC). The short-term solutions were verified and implemented during the master's thesis at Seco Tools. Recommendations were given for future implementation of the long-term solutions. In this thesis, insights are shared regarding the applicability of OPLS and Partial Least Squares (PLS) regression models for batch monitoring of the CVD process. We also demonstrate that the prediction of a certain critical breakdown, clogging of the aluminum generator in the CVD process, can be accomplished through the use of SPC. For the wet blasting process, a PCA methodology is suggested to be effective for visualizing breakdowns.
42
Sjöström, William. "Comparing SKF and Erbessd sensor integration for predictivemaintenance." Thesis, Linköpings universitet, Institutionen för datavetenskap, 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-176718.
Full textAbstract:
The purpose of this thesis was to compare two integration’s of sensors, into a system called Enlight, but could in theoryhave been integrated to most systems. As a pre-study, the specifications and availability of five sensors were researched.From the pre-study, Smart Edge 4.0 and Phantom EPH-V11/10 from Erbessd, were chosen and then integrated. Usability andperformance of the integrations were then compared usingcognitive dimensions and stopwatch. Phantom from Erbessdwas deemed to be more usable, and the integration of SmartEdge 4.0, had better performance.
43
Butylin, Sergei. "Predictive Maintenance Framework for a Vehicular IoT Gateway Node Using Active Database Rules." Thesis, Université d'Ottawa / University of Ottawa, 2018. http://hdl.handle.net/10393/38568.
Full textAbstract:
This thesis describes a proposed design and implementation of a predictive maintenance engine developed to fulfill the requirements of the STO Company (Societe de transport de l'Outaouais) for maintaining vehicles in the fleet. Predictive maintenance is proven to be an effective approach and has become an industry standard in many fields. However, in the transportation industry, it is still in the stages of development due to the complexity of moving systems and the high level dimensions of involved parameters. Because it is almost impossible to cover all use cases of the vehicle operational process using one particular approach to predictive maintenance, in our work we take a systematic approach to designing a predictive maintenance system in several steps. Each step is implemented at the corresponding development stage based on the available data accumulated during system funсtioning cycle.
% by dividing the entire system into modules and implementing different approaches.
This thesis delves into the process of designing the general infrastructural model of the fleet management system (FMS), while focusing on the edge gateway module located on the vehicle and its function of detecting maintenance events based on current vehicle status. Several approaches may be used to detect maintenance events, such as a machine learning approach or an expert system-based approach. While the final version of fleet management system will use a hybrid approach, in this thesis paper we chose to focus on the second option based on expert knowledge, while machine learning has been left for future implementation since it requires extensive training data to be gathered prior to conducting experiments and actualizing operations.
Inspired by the IDEA methodology which promotes mapping business rules as software classes and using the object-relational model for mapping objects to database entities, we take active database features as a base for developing a rule engine implementation. However, in contrast to the IDEA methodology which seeks to describe the specific system and its sub-modules, then build active rules based on the interaction between sub-systems, we are not aware of the functional structure of the vehicle due to its complexity. Instead, we develop a framework for creating specific active rules based on abstract classifications structured as ECA rules (event-condition-action), but with some expansions made due to the specifics of vehicle maintenance. The thesis describes an attempt to implement such a framework, and particularly the rule engine module, using active database features making it possible to encapsulate the active behaviour inside the database and decouple event detection from other functionalities. We provide the system with a set of example rules and then conduct a series of experiments analyzing the system for performance and correctness of events detection.
44
Marastoni, Gabriele. "Towards predictive maintenance at LHC computing centers: exploration of monitoring data at CNAF." Bachelor's thesis, Alma Mater Studiorum - Università di Bologna, 2018. http://amslaurea.unibo.it/16923/.
Full textAbstract:
Nel campo delle applicazioni industriali e scientifiche, l'emergere del machine learning sta cambiando il modo in cui la raccolta dei dati e la loro analisi è concepita e realizzata. Nel dettaglio, l'analisi di volumi massivi di file log compiute da unità computazionali facenti parte di infrastrutture estremamente grandi e complesse sta diventando un modo impegnativo ma promettente per estrarre informazioni fruibili nell'ambito del miglioramento e ottimizzazione nell'uso delle risorse e per un risparmio economico.
Questo è particolarmente interessante per il calcolo computazionale al LHC anche perchè si prevede che le attività future saranno svolte in "flat budget" per maggior parte dei finanziamenti elargiti nei prossimi anni. La Worldwide LHC Computing Grid coordina le operazioni di una grande quantità di centri di calcolo nel mondo, ognuno di questi è una composizione coerente di spazi di storage, potere di elaborazione e connessioni network sopra la quali vengono eseguite una grande quantità di applicazioni che lavorano con livelli di software comuni o specifici di determinati esperimenti. Questi servizi producono grandissimi volumi di log eterogenei e non strutturati i quali possono essere digeriti attraverso approcci tipici del Big Data Analitics. Presso il data center Tier-1 di Bologna il CNAF ha iniziato un'attività in quest'ambito con tentativi di analisi dati; si propone nel lungo periodo di ingegnerizzare e schierare una soluzione di manutenzione predittiva basata su tecniche di machine learning.
Questo lavoro di tesi si propone di raccogliere, manipolare ed esplorare i log prodotti da un servizio specifico del CNAF - StoRM, il servizio di gestione dello storage. L'obiettivo è gettare le basi di questa indagine collezionando osservazioni e producendo strumenti che possano essere utilizzati da altri su differenti tipi di log; quindi fare un primo piccolo passo verso un approccio basato sulla manutenzione predittiva per i centri computazionali di LHC.
45
Minarini, Francesco. "Anomaly detection prototype for log-based predictive maintenance at INFN-CNAF tier-1." Master's thesis, Alma Mater Studiorum - Università di Bologna, 2019. http://amslaurea.unibo.it/19304/.
Full textAbstract:
Splitting the evolution of HEP from the one of computational resources needed to perform analyses is, nowadays, not possible. Each year, in fact, LHC produces dozens of PetaBytes of data (e.g. collision data, particle simulation, metadata etc.) that need orchestrated computing resources for storage, computational power and high throughput networks to connect centers. As a consequence of the LHC upgrade, the Luminosity of the experiment will increase by a factor of 10 over its originally designed value, entailing a non negligible technical challenge at computing centers: it is expected, in fact, an uprising in the amount of data produced and processed by the experiment. With this in mind, the HEP Software Foundation took action and released a road-map document describing the actions needed to prepare the computational infrastructure to support the upgrade. As a part of this collective effort, involving all computing centres of the Grid, INFN-CNAF has set a preliminary study towards the development of AI driven maintenance paradigm. As a contribution to this preparatory study, this master thesis presents an original software prototype that has been developed to handle the task of identifying critical activity time windows of a specific service (StoRM). Moreover, the prototype explores the viability of a content extraction via Text Processing techniques, applying such strategies to messages belonging to anomalous time windows.
46
Kaiser, Kevin Michael. "A simulation study of predictive maintenance policies and how they impact manufacturing systems." Thesis, University of Iowa, 2007. http://ir.uiowa.edu/etd/152.
Full text
47
Gauchel, Wolfgang, Thilo Streichert, and Yannick Wilhelm. "Predictive maintenance with a minimum of sensors using pneumatic clamps as an example." Technische Universität Dresden, 2020. https://tud.qucosa.de/id/qucosa%3A71204.
Full textAbstract:
In standard pneumatics, the available signals for data analytics are very limited. As a rule, no continuous status information is available. Usually only the reaching of the end position is indicated - by means of a digital signal of a proximity sensor. This paper examines whether these limited data can be used to derive usable and useful information for predictive maintenance. Pneumatic clamps in bodyin- white construction were chosen as application example. The paper describes a continuous run to investigate the basic feasibility of predictibility. In the following chapters, possibilities for error classification are discussed. Finally, the implementation of the findings in a field test is described.
48
Guillen, Rosaperez Diego Alonso. "Self-Learning Methodology for Failure Detection in an Oil- Hydraulic Press : Predictive maintenance." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-289371.
Full textAbstract:
Deep Learning methods have dramatically improved the state-of-the-art across multiple fields, such as speech recognition, object detection, among others. Nevertheless, its application on signal processing, where data is frequently unlabelled, has received relatively little attention. In this field, nowadays, a set of sub-optimal techniques are often used. They usually require an expert to manually extract features to analyse, which is a knowledge and labour intensive process. Thus, a self-learning technique could improve current methods. Moreover, certain machines in a factory are particularly complex, such as an oil-hydraulic press. Here, its sensors can only identify few failures by setting up some thresholds, but they commonly cannot detect wear on its internal components. So, a self-learning technique would be required to detect anomalies related to deterioration. The concept is to determine the condition of a machine and to predict breakdowns by analysing patterns in the measurements from their sensors. This document proposes a self-learning methodology that uses a deep learning model to predict failures in such a machine. The core idea is to train an algorithm that can identify by itself the relevant features to extract on a work cycle, and to relate them to a part which will breakdown. The conducted evaluation focuses on an example case where a hydraulic accumulator fails. As result, it was possible to forecast its breakdown two weeks in advance. Finally, the proposed method provides explanations at every step, after acknowledging their importance in industrial applications. Also, some considerations and limitations of this technique are stated to support guiding the expectation of some stakeholders in a factory, i.e. a (Global) Process Owner.Deep Learning-metoder har dramatiskt förbättrat det senaste inom flera fält, såsom taligenkänning, objektdetektering, bland andra. Ändå har dess tillämpning på signalbehandling, där data ofta är omärkt, fått relativt lite uppmärksamhet. I detta fält används numera ofta en uppsättning suboptimala tekniker. De kräver vanligtvis en expert för att manuellt extrahera funktioner för att analysera, vilket är en kunskaps och arbetsintensiv process. Således kan en självlärande teknik förbättra nuvarande metoder. Dessutom är vissa maskiner i en fabrik särskilt komplexa, såsom en oljehydraulisk press. Här kan dess sensorer bara identifiera några fel genom att ställa in vissa trösklar, men de kan vanligtvis inte upptäcka slitage på dess interna komponenter. Så, en självlärande teknik skulle krävas för att upptäcka avvikelser relaterade till försämring. Konceptet är att bestämma maskinens tillstånd och att förutsäga haverier genom att analysera mönster i mätningarna från deras sensorer. Detta dokument föreslår en självlärningsmetodik som använder en djupinlärningsmodell för att förutsäga fel i en sådan maskin. Kärnidén är att träna en algoritm som i sig kan identifiera de relevanta funktionerna som ska extraheras i en arbetscykel och att relatera dem till en del som kommer att bryta ner. Den genomförda utvärderingen fokuserar på ett exempel på fall där en hydraulisk ackumulator misslyckas. Som ett resultat var det möjligt att förutse dess fördelning två veckor i förväg. Slutligen ger den föreslagna metoden förklaringar i varje steg, efter att ha erkänt deras betydelse i industriella applikationer. Några överväganden och begränsningar av denna teknik anges också som stöd för att vägleda förväntningarna hos vissa intressenter i en fabrik, dvs. en (global) processägare.
49
Calabrese, Francesca <1992>. "Integrating Machine Learning Paradigms for Predictive Maintenance in the Fourth Industrial Revolution era." Doctoral thesis, Alma Mater Studiorum - Università di Bologna, 2022. http://amsdottorato.unibo.it/10133/1/Tesi_CalabreseFrancesca.pdf.
Full textAbstract:
In the last decade, manufacturing companies have been facing two significant challenges. First, digitalization imposes adopting Industry 4.0 technologies and allows creating smart, connected, self-aware, and self-predictive factories. Second, the attention on sustainability imposes to evaluate and reduce the impact of the implemented solutions from economic and social points of view. In manufacturing companies, the maintenance of physical assets assumes a critical role. Increasing the reliability and the availability of production systems leads to the minimization of systems’ downtimes; In addition, the proper system functioning avoids production wastes and potentially catastrophic accidents. Digitalization and new ICT technologies have assumed a relevant role in maintenance strategies. They allow assessing the health condition of machinery at any point in time. Moreover, they allow predicting the future behavior of machinery so that maintenance interventions can be planned, and the useful life of components can be exploited until the time instant before their fault.
This dissertation provides insights on Predictive Maintenance goals and tools in Industry 4.0 and proposes a novel data acquisition, processing, sharing, and storage framework that addresses typical issues machine producers and users encounter. The research elaborates on two research questions that narrow down the potential approaches to data acquisition, processing, and analysis for fault diagnostics in evolving environments. The research activity is developed according to a research framework, where the research questions are addressed by research levers that are explored according to research topics. Each topic requires a specific set of methods and approaches; however, the overarching methodological approach presented in this dissertation includes three fundamental aspects: the maximization of the quality level of input data, the use of Machine Learning methods for data analysis, and the use of case studies deriving from both controlled environments (laboratory) and real-world instances.
50
Falcomer, Carlo <1993>. "Big data analytics for proactive and predictive maintenance in electric car battery packs." Doctoral thesis, Alma Mater Studiorum - Università di Bologna, 2022. http://amsdottorato.unibo.it/10183/1/TesiDottorato.pdf.
Full textAbstract:
The idea behind the project is to develop a methodology for analyzing and developing techniques for the diagnosis and the prediction of the state of charge and health of lithium-ion batteries for automotive applications.
For lithium-ion batteries, residual functionality is measured in terms of state of health; however, this value cannot be directly associated with a measurable value, so it must be estimated. The development of the algorithms is based on the identification of the causes of battery degradation, in order to model and predict the trend. Therefore, models have been developed that are able to predict the electrical, thermal and aging behavior.
In addition to the model, it was necessary to develop algorithms capable of monitoring the state of the battery, online and offline. This was possible with the use of algorithms based on Kalman filters, which allow the estimation of the system status in real time. Through machine learning algorithms, which allow offline analysis of battery deterioration using a statistical approach, it is possible to analyze information from the entire fleet of vehicles. Both systems work in synergy in order to achieve the best performance.
Validation was performed with laboratory tests on different batteries and under different conditions. The development of the model allowed to reduce the time of the experimental tests. Some specific phenomena were tested in the laboratory, and the other cases were artificially generated.L'idea alla base del progetto è stata quella di sviluppare una metodologia di analisi e di sviluppo di tecniche per la diagnosi e la previsione dello stato di carica e di salute delle batterie agli ioni di litio per applicazioni automobilistiche. Per le batterie agli ioni di litio, la funzionalità residua è misurata in termini di stato di salute, tuttavia questo valore non può essere direttamente associato ad un valore misurabile, di conseguenza è necessario stimarlo. Lo sviluppo degli algoritmi è basato sull'identificazione delle cause di degrado delle batterie, al fine di modellarne e prevederne il comportamento. Sono stati dunque sviluppati modelli in grado di prevedere il comportamento elettrico e termico, e di invecchiamento della batteria. Oltre al modello, è stato necessario sviluppare algoritmi in grado di monitorare lo stato della batteria, online e offline, questo è stato possibile con l'utilizzo di algoritmi basati su filtri di Kalman, che permettono la stima dello stato del sistema in tempo reale. Attraverso algoritmi di machine learning, che consentono di analizzare offline il deterioramento della batteria con un approccio statistico, è possibile analizzare le informazioni dell'intera flotta di veicoli. Entrambi i sistemi lavorano in sinergia al fine di ottenere le migliori prestazioni. La validazione è stata eseguita con test di laboratorio su diverse batterie e in diverse condizioni. Lo sviluppo del modello ha permesso di ridurre il tempo delle prove sperimentali. Alcuni fenomeni specifici sono stati testati in laboratorio, e gli altri casi sono stati generati artificialmente.