Nowadays, the number of interconnected devices is increasing dramatically: devices used in everyday life are a source of data that can be used for any purpose. Gaining value from this data is the most important task: data can be used for understanding the interested environmental trends, predicting their behaviour and thus, generating new data. This view can be applied to the management of automated machines, providing with the possibility to analyse their working status, to understand and improve their throughput and to accomplish the necessary maintenance operations in time. Thus, in a so structured and interconnected environment, high volumes of data are being generated day by day, creating Big Data which are ready to serve deep analyses and develop advanced monitoring solutions. The work thesis is based on a project developed in collaboration with Elettric80 S.p.A., with the aim of developing a monitoring system for its laser guided AGV systems, to then be part of the company commercial offer. The thesis initially shows the state of the art of maintenance techniques, then introduces the theoretical concepts on which it is based, such as: • Big Data; • Message Brokers; • Software containerisation systems; • Hardware In the Loop; • Industry 4.0. The methodology used in the project is then illustrated: its requirements are collected and analysed. A first conceptual architecture is then defined: this must respect several constraints including the capability to manage a large amount of data, as well as being capable to save them easily on the database. In fact, Elettric80 has several customers who also 6 | P a g e manage hundreds of laser guided AGVs and the system must be able to easily handle this complexity. Based on this architecture, software solutions are chosen to meet the design requirements. The final solution is then explained in all its components: starting from the machine having the task of being able to send data from its sensors, a system installed on a server is responsible for acquiring such data, processing and showing it in a real-time fashion or in terms of batches. Solutions such as NoSQL databases (Apache Cassandra) and Message Brokers (Apache Kafka) are milestones of the architecture, as they allow you to easily manage huge amounts of data coming from all the machines of the customer, to analyse and save them safely. Certain types of analyses were defined with Elettric80 during the development of the software solution: analyses such as the quality of navigation and data coming from the machine sensors were implemented. A web dashboard will then have the task of showing the collected and analysed data. Finally, additional solutions have been implemented in order to make this architecture solid: checks are carried out so that all the components work properly and act in case they are not working as expected (automatic troubleshooting and technician alerting). Finally, the achieved results are explained and commented as well as the way the project has been leaded shown. Future works are explained in the last chapter.

Al giorno d’oggi, il numero di dispositivi interconnessi sta aumentando notevolmente: dispositivi utilizzati nella vita di tutti i giorni sono una importante sorgente di dati che può essere usata per qualunque scopo. Acquisire valore da questi dati è il compito più importante: i dati possono essere usati per conoscere i trend all’interno dell’ambiente di interesse, predirne il comportamento e perciò generare nuovi dati. Questa visione può essere applicata alla gestione delle macchine automatiche, dando la possibilità di analizzare il loro stato di funzionamento, comprenderne e migliorarne il throughput ed effettuare le necessarie operazioni di manutenzione in tempo. Perciò, in un ambiente così strutturato ed interconnesso, un alto volume di dati viene generato giorno dopo giorno, creando i Big Data: questi sono a loro volta utilizzati per effettuare analisi approfondite e realizzare avanzati sistemi di monitoraggio. Il lavoro di tesi si basa su un progetto sviluppato in collaborazione con Elettric80 S.p.A., con l’obiettivo di sviluppare un sistema di monitoraggio per sistemi AGV a guida laser, per poi corredare l’offerta commerciale dell’azienda. La tesi mostra inizialmente lo stato dell’arte ad oggi delle tecniche di manutenzione, dopodiché introduce i concetti teorici su cui essa si basa, quali: • Big Data; • Message Brokers; • Sistemi di containerizzazione software; • Hardware In the Loop; • Industria 4.0. La metodologia utilizzata nel progetto di tesi viene quindi illustrata: i requisiti di progetto vengono raccolti ed analizzati. Una prima architettura concettuale viene quindi definita: questa deve rispettare diversi vincoli tra cui anche quello di poter gestire una grande mole di dati, nonché poterli salvare agilmente su database. Infatti, Elettric80 ha diversi clienti che gestiscono anche centinaia di AGV a guida laser ed il 8 | P a g e sistema dovrà essere in grado di poter gestire agevolmente tale complessità. Basandosi su questa architettura, le soluzioni software sono scelte in modo da poter soddisfare i requisiti di progetto. La soluzione finale viene quindi spiegata in tutte le sue componenti: partendo dalla macchina avente il compito di poter inviare dati provenienti dai propri sensori, un sistema installato su un server ha l’onere di acquisire tali dati, processarli e poterli mostrare in real-time o in batches. Soluzioni come NoSQL databases (Apache Cassandra) e Message Brokers (Apache Kafka) sono punti cardini dell’architettura, in quanto permettono di poter gestire agevolmente le enormi moli di dati provenienti da tutte le macchine presenti presso il cliente, di poter quindi analizzarli e salvarli in modo sicuro. Determinate tipologie di analisi sono state definite con Elettric80 durante lo sviluppo della soluzione software: analisi quali la qualità della navigazione e dei dati provenienti dai sensori macchina sono state implementate. Una web dashboard avrà quindi il compito di poter mostrare i dati raccolti ed analizzati. Infine, soluzioni aggiuntive sono state implementate in modo da poter rendere tale architettura solida: controlli vengono effettuati affinché tutti i componenti funzionino correttamente ed azioni automatiche vengono intraprese nel caso in cui non funzioni inaspettatamente (risoluzione automatica delle problematiche ed allertamento dei tecnici). Infine, i risultati raggiunti vengono quindi spiegati e commentati, nonché l’organizzazione del lavoro mostrata. I lavori futuri vengono quindi illustrati nell’ultimo capitolo.

Big Data per la diagnosi avanzata dei guasti e sistemi di monitoraggio:gestire la complessità di sistemi in ambienti distribuiti / Claudio Santo Longo , 2020 Mar 05. 32. ciclo, Anno Accademico 2018/2019.

Big Data per la diagnosi avanzata dei guasti e sistemi di monitoraggio:gestire la complessità di sistemi in ambienti distribuiti

LONGO, CLAUDIO SANTO
2020

Abstract

Nowadays, the number of interconnected devices is increasing dramatically: devices used in everyday life are a source of data that can be used for any purpose. Gaining value from this data is the most important task: data can be used for understanding the interested environmental trends, predicting their behaviour and thus, generating new data. This view can be applied to the management of automated machines, providing with the possibility to analyse their working status, to understand and improve their throughput and to accomplish the necessary maintenance operations in time. Thus, in a so structured and interconnected environment, high volumes of data are being generated day by day, creating Big Data which are ready to serve deep analyses and develop advanced monitoring solutions. The work thesis is based on a project developed in collaboration with Elettric80 S.p.A., with the aim of developing a monitoring system for its laser guided AGV systems, to then be part of the company commercial offer. The thesis initially shows the state of the art of maintenance techniques, then introduces the theoretical concepts on which it is based, such as: • Big Data; • Message Brokers; • Software containerisation systems; • Hardware In the Loop; • Industry 4.0. The methodology used in the project is then illustrated: its requirements are collected and analysed. A first conceptual architecture is then defined: this must respect several constraints including the capability to manage a large amount of data, as well as being capable to save them easily on the database. In fact, Elettric80 has several customers who also 6 | P a g e manage hundreds of laser guided AGVs and the system must be able to easily handle this complexity. Based on this architecture, software solutions are chosen to meet the design requirements. The final solution is then explained in all its components: starting from the machine having the task of being able to send data from its sensors, a system installed on a server is responsible for acquiring such data, processing and showing it in a real-time fashion or in terms of batches. Solutions such as NoSQL databases (Apache Cassandra) and Message Brokers (Apache Kafka) are milestones of the architecture, as they allow you to easily manage huge amounts of data coming from all the machines of the customer, to analyse and save them safely. Certain types of analyses were defined with Elettric80 during the development of the software solution: analyses such as the quality of navigation and data coming from the machine sensors were implemented. A web dashboard will then have the task of showing the collected and analysed data. Finally, additional solutions have been implemented in order to make this architecture solid: checks are carried out so that all the components work properly and act in case they are not working as expected (automatic troubleshooting and technician alerting). Finally, the achieved results are explained and commented as well as the way the project has been leaded shown. Future works are explained in the last chapter.
Big Data for advanced fault diagnosis and monitoring systems: managing system complexity in a distributed environment
5-mar-2020
FANTUZZI, Cesare
File in questo prodotto:
File Dimensione Formato  
Claudio Santo Longo - PhD Thesis.pdf

Open access

Descrizione: tesi di dottorato
Dimensione 7.94 MB
Formato Adobe PDF
7.94 MB Adobe PDF Visualizza/Apri
Pubblicazioni consigliate

Licenza Creative Commons
I metadati presenti in IRIS UNIMORE sono rilasciati con licenza Creative Commons CC0 1.0 Universal, mentre i file delle pubblicazioni sono rilasciati con licenza Attribuzione 4.0 Internazionale (CC BY 4.0), salvo diversa indicazione.
In caso di violazione di copyright, contattare Supporto Iris

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11380/1200382
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact