No description available.

No description available.

The use of machine learning for predictive maintenance has been the focus of many studies, usually utilizing industrial setups consisting of actual industrial motors. This work examines the possibility of creating a simple setup to develop a machine learning model to detect electric motor failures, eliminating the need to rely on having access to industrial equipment in the early stages. The work conducted in this thesis leverages autoencoders, a specific type of neural network, to detect motor faults based on vibration readings from an accelerometer. The final model detected anomalies with 100% accuracy at three different speeds when a constant load was applied to the motor. However, it should be improved when a variation in load is introduced as it only had 85.1% Accuracy and 90.1% F1-score with 82.0% Recall and 99.8% Precision. In conclusion, the setup and the model developed show promise as an initial setup for testing and experimenting with electric motors and machine learning for predictive maintenance.

Digitaliseringen inom industrin medför ett antal utmaningar där manuella tillståndskontroller övergår till digitaliserade mätningar. Utmaningarna som uppstår med det nya arbetssättet är vilken data som ska samlas in samt hur den genererade data ska analyseras. Syftet med detta examensarbete är att förslå åtgärder för att effektivisera och implementera Industri 4.0 (Smart Maintenance) genom prediktivt underhåll på Scania. Således uppnås en mer kostnadseffektiv verksamhet samtidigt som det bidrar till att skapa ett hållbarare företag. Detta genom att utnyttja komponenters fulla livslängd och inte utföra onödigt underhåll. Det prediktiva underhållet kommer medföra en högre tillgänglighet och tillförlitlighet inom maskinparken på Scanias cylinderhuvudlinje. För att implementera underhållsstrategin i examensarbetet så används en mjukvara som skapats av styrsystems leverantör FANUC. Mjukvaran är ett mätverktyg som heter Servo Viewer och kan mäta maskinens status genom att avläsa procentsatsen utav den totala mängden vridmoment som bildas samt positionsfel under maskinens körning. Ett arbetssätt för att automatisera analysering av data som hämtas ifrån Servo Viewer är att mätningarna samlas i databasen MT-LINKi för att sedan kunna analyseras av ett program FANUC AI Servo Monitor. Den slutsats som kan dras av arbetet är att det går att använda FANUC Servo Viewer till att avläsa maskinens kondition och därmed prediktera när underhåll behöver utföras eftersom det går att avläsa avvikande faktorer under mätningarna. Vid dessa faktorer går det att sätta triggers som kommer larma i systemet när maskinen överstiger dem. De komponenter som mjukvaran kommer varna systemet för är alla komponenter som har en påverkande faktor på fleroperationsmaskinen. Dessa komponenter kan vara allt från kulskruvar, linjärstyrning, pulsgivare, remmar, servomotorer och servokort. Men arbetet är inte ett färdigt koncept i sig, det behövs fler mätningar över tid för att kunna skapa ett tydligare normalläge samt identifiera felutvecklingskurvor för att ställa in triggers i mjukvaran. Eftersom analys och insamling av mätdata blev mer tidkrävande än planerat så har ej utvärderingen av MT-LINKi samt AI Servo Monitor utförts och en vidare beskrivning av arbetet har lämnats. / Digitalization in the manufacturing industry involves many challenges due to moving from manual controls towards digitalized condition monitoring. The challenges that occur with the new way of working is what data should be collected and how it should be analyzed. This thesis aims to streamline the industry and implement Industry 4.0 and Smart Maintenance through predictive maintenance in Scania. In this way a more cost-effective business is achieved at the same time as it contributes to creating a more sustainable company. The predictive maintenance will lead to a higher availability and reliability within the machine park at Scania´s cylinder head line. To be able to implement the maintenance strategy a software created by FANUC, the system supplier, is used. The software is a measuring tool called Servo Viewer and it can analyze the status of the machine by measuring the percentage of the total amount of torque that is available and the position error while the machine is running. The thesis also aims to investigate how to automatize the measurements within a database called MT-LINKi and later be analyzed by a software called AI Servo Monitor. The conclusion that can be drawn from the thesis is that it is possible to use FANUC Servo Viewer to measure the condition of the machine and therefore being able to predict when the maintenance needs to be performed as it is possible to read deviating factors during the measurements. With these factors it is possible to set triggers that will alarm the system when the machine exceeds them. Some of the components that will be possible to monitor condition for, are ball screws, linear control, encoders, belts, servomotors and servo cards. However the work isn’t not a complete concept in itself, more measurements are needed to be performed over time to create a normal situation and identify error development graphs to set the triggers in the software. As analysis and collection of measurement data became more time consuming than planned, the evaluation of MT-LINKi and AI Servo Monitor has not been performed and a further description of the work has been provided.

Underhåll har alltid varit viktigt för industriella företag, med nya teknologier som prediktivt underhåll på väg fram. Insamlad sensordata och annan data används för att förbättra underhållet och för att minska kostnader och öka konkurrenskraften för företaget. Små och medelstora företag (SME) är viktiga för ekonomin, med 99 % av alla företag i Europa betraktade som SMEs idag. Det är viktigt att små och medelstora företag också inkluderas i implementeringen av prediktivt underhåll på grund av deras stora antal och betydelse för ekonomin. Syftet med denna avhandling har varit att undersöka utmaningar i implementering för SME som vill använda prediktivt underhåll. Problemet har varit att identifiera utmaningar och möjliga alternativ för mindre företag genom att utvärdera och jämföra flera maskininlärningsalgoritmer. Problemet som undersökts i avhandlingen har varit att identifiera en maskininlärningsalgoritm som optimerar både prestanda och resursanvändning. Genom att göra prediktivt underhåll mer tillgängligt för SME kan de också dra nytta av minskade driftskostnader, förlängd livslängd på sin utrustning och förbli konkurrenskraftiga i framtiden. Fem maskininlärningsmodeller tränades och testades. Algoritmen som presterade bäst var XGBoost, med Random Forest som en nära utmanare. Om resurserna är mycket begränsade presterar Decision Trees bäst av de enklare modellerna. / Maintenance has always been an important aspect for industrial companies, with new technologies allowing for predictive maintenance to be an option. Collected sensor data, and other data, can in this case be used to improve maintenance to reduce costs and increase competitiveness for the company. Small and medium-sized enterprises (SMEs) are important to the economy with 99% of all businesses in Europe being considered an SME. It’s important for small and mediumsized companies to also be included in the implementations of predictive maintenance due to their great number and importance to the economy. The aim of this thesis has been to investigate implementation challenges for SMEs that want to use predictive maintenance. The problem has been to identify challenges and possible options for smaller companies by evaluating and comparing several machine learning algorithms. The problem investigated in this thesis is identifying a machine learning algorithm that optimizes both performance and resource use. By making predictive maintenance more approachable for SMEs, they can also benefit from a reduction in operational costs, extended lifespan of their equipment and remain competitive in the future. Five machine learning models underwent training and testing. The algorithm that performed best was XGBoost, with Random Forest being a close contender. However, if resources are very limited, Decision Trees perform best out of the simpler models.

Unternehmen stehen zunehmend vor der Herausforderung, dem drohenden Wissensverlust durch demografischen Wandel und Mitarbeiterabgang zu begegnen. In Zeiten voranschreitender Digitalisierung gilt es, große Datenmengen beherrschbar und nutzbar zu machen, mit dem Ziel, einerseits die Ressourceneffizienz innerhalb des Unternehmens zu erhöhen und anderseits den Kunden zusätzliche Dienstleistungen anbieten zu können. Vor dem Hintergrund, ein effizientes Qualitätsmanagement und eine vorausschauende Instandhaltung mit ein und demselben System zu realisieren, sind zunächst technologische Kennzahlen und die Prozessführung zu bestimmen. Im Bereich der intelligenten Instandhaltung ist es jedoch nicht immer möglich, Fehlerzustände von physischen Anlagen im Serienbetrieb als Datensatz abzufassen. Das bewusste Zulassen von Fehlern unter realen Produktionsbedingungen könnte zu fatalen Ausfällen bis hin zur Zerstörung der Anlage führen. Auch das gezielte Erzeugen von Fehlern unter stark kontrollierten Bedingungen kann zeitaufwendig, kostenintensiv oder sogar undurchführbar sein.

Companies are increasingly challenged by the impending loss of knowledge due to demographic change and employee loss. In times of advancing digitalization, it is important to make large datasets accessible and usable, aiming at increasing resource efficiency within the company on the one hand and being able to offer customers additional services on the other. Given the background of implementing efficient quality management and predictive maintenance with the same system, technological key figures and process control must first be determined. In the field of intelligent maintenance, however, it is not always possible to record error states of physical systems in series operation as a data set. Deliberately allowing faults to occur under real production conditions could lead to fatal failures or even the destruction of the system. The targeted generation of faults under highly controlled conditions can also be timeconsuming, cost-intensive, or even impractical.

A non-invasive, on-line method for detection of mechanical (rotor, bearings eccentricity) and stator winding faults in a 3-phase induction motors from observation of motor line current supply input. The main aim is to avoid the consequence of unexpected failure of critical equipment which results in extended process shutdown, costly machinery repair, and health and safety problems. This thesis looks into the possibility of utilizing machine learning techniques in the field of condition monitoring of electromechanical systems. Induction motors are chosen as an example for such application. Electrical motors play a vital role in our everyday life. Induction motors are kept in operation through monitoring its condition in a continuous manner in order to minimise their off times. The author proposes a model free sensor-less monitoring system, where the only monitored signal is the input to the induction motor. The thesis considers different methods available in literature for condition monitoring of induction motors and adopts a simple solution that is based on monitoring of the motor current. The method proposed use the feature extraction and Support Vector Machines (SVM) to set the limits for healthy and faulty data based on the statistical methods. After an extensive overview of the related literature and studies, the motor which is the virtual sensor in the drive system is analysed by considering its construction and principle of operation. The mathematical model of the motor is used for analysing the system. This is followed by laboratory testing of healthy motors and comparing their output signals with those of the same motors after being intentionally failed, concluding with the development of a full monitoring system. Finally, a monitoring system is proposed that can detect the presence of a fault in the monitored machine and diagnose the fault type and severity

Den nya generationens digitalisering har slagit rot i samhället. Algoritmer och datamodeller styr nyhetsflödet i social media, röststyr mobilen genom att tolka rösten och självstyr bilen, helt och hållet i autonoma fordon. Inom industrierna finns det också en pågående process där machine learning kan appliceras för att öka drifttillgänglighet och minska kostnader. Det nuvarande paradigmet för att underhålla icke-säkerhetsklassade maskiner i kärnkraftindustrin är en kombination av Avhjälpande Underhåll och Förebyggande Underhåll. Avhjälpande underhåll innebär att underhålla maskinen när fel inträffar, förebyggande underhåll innebär att underhålla med periodiska intervall. Båda sätten är kostsamma för att de riskerar att under- respektive över-underhålla maskinen och blir därmed resurskrävande. Ett paradigmskifte är på väg, det stavas Prediktivt Underhåll - att kunna förutspå fel innan de inträffar och  planera underhåll därefter. Den här rapporten utforskar möjligheten att använda sig av de neurala nätverken LSTM och GRU för att kunna prognostisera eventuella skador på maskiner. Det här baseras på mätdata och historiska fel på maskinen. / The new generation of digitalization has been ingrained into society. Algorithms and data models are controlling the news feed of social media, controlling the phone by interpreting voices and controlling the car, altogether with automonous vehicles. In the industries there is also an ongoing process where machine learning is applied to increase availability and reduce costs. The current paradigm for maintaining non-critical machines in the nuclear power industry is a combination of corrective maintenance and preventive maintenance. Corrective maintenance means doing repairs on the machine upon faults, preventive maintenance means doing repairs periodically. Both ways are costly because they run the risk of under- and over-maintaining the machine and therefore becoming resource-intensive. A paradigm shift is on it's way, and it's spelled Predictive Maintenance - being able to predict faults before they happen and plan maintenance thence. This report explores the possibilities of using LSTM and GRU to forecast potential damage on machines. This is based on data from measurements and historical issues on the machine.

Submitted by FERNANDA ALMEIDA MARCHINI GAYER null (fer_marchini@yahoo.com.br) on 2017-12-12T16:34:52Z No. of bitstreams: 1 dissertacao.pdf: 8726620 bytes, checksum: 66b744fe24afaf15e89574cfd8c23af1 (MD5) / Approved for entry into archive by Adriana Aparecida Puerta null (dripuerta@rc.unesp.br) on 2017-12-12T18:20:47Z (GMT) No. of bitstreams: 1 gayer_fam_me_rcla.pdf: 8697307 bytes, checksum: ef684be2abf1d5c9b37f2d77c9ab5ce8 (MD5) / Made available in DSpace on 2017-12-12T18:20:47Z (GMT). No. of bitstreams: 1 gayer_fam_me_rcla.pdf: 8697307 bytes, checksum: ef684be2abf1d5c9b37f2d77c9ab5ce8 (MD5) Previous issue date: 2017-12-01 / Neste trabalho, apresentamos um estudo introdutório à Teoria de Conjuntos Fuzzy e Lógica Fuzzy, mostrando o seu potencial de aplicação através da análise de um problema numa indústria química e da introdução de alguns conceitos dessa teoria a alunos do Ensino Médio. Quanto ao problema da indústria química, o objetivo é assessorar uma equipe de gestão de manutenção de uma certa indústria química na tomada de decisão quanto ao momento oportuno para realização de intervenção em uma bomba industrial. Esta equipe observou como dificuldade em sua gestão de custos a manutenção preventiva de uma certa bomba de seu processo. Para isso, dados dos processos são analisados e modelados matematicamente utilizando a lógica fuzzy, produzindo um resultado que direciona corretamente os profissionais no momento da tomada de decisão, sobrepondo a manutenção preventiva existente pela manutenção preditiva, baseada em dados do processo. Foi criado um sistema computacional que favorece este processo de forma automática e simplificada, evitando burocracias legais quanto à questão de licenciamento de softwares do mercado, de forma a realizar os cálculos, além de facilitar a incorporação de dados e análises de forma intuitiva, não necessitando de maiores treinamentos para tanto. Por fim, durante o estudo da lógica fuzzy e pesquisas relacionadas, foi detectada a possibilidade de aplicação prática para estudantes do Ensino Médio. Dessa forma, uma aula expositiva com atividades originais foi realizada para apresentar os conjuntos e a lógica fuzzy, mostrando a capacidade dos alunos do Ensino Médio em assimilar os conteúdos já citados. / In this work, we present an introductory study to Fuzzy Set Theory and Fuzzy Logic, showing its potential application through the analysis of a problem in a chemical industry and the introduction of some concepts of this theory to high school students. As for the problem of the chemical industry, the objective is to advise a manufacturing management team of a certain chemical industry in the decision making for an opportune time for an intervention in an industrial pump. This team observed as a difficulty in its cost management a preventive maintenance of a certain pump of its process. For this, the data of the processes are analysed and modelled mathematically using a fuzzy logic, producing a result that correctly directs the professionals at the moment of the decision making, overlapping the existing preventive maintenance, by the base predictive maintenance in process data. It was created a computer system that helps this process in an automatic and simplified way, avoiding legal bureaucracies regarding the licensing of software in the market, in order to perform the calculations, besides facilitating the incorporation of data and analyses in an intuitive way, not requiring of greater training for both. Finally, during the study of fuzzy logic and related research, it was detected as a possibility of practical application for high school students. Thus, an expository class with original activities was performed to present the sets and a fuzzy logic, showing the ability of the high school students to assimilate the referred contents.