Nowadays, assembly systems are facing major challenges as efficiency levels are not satisfactory due to constantly evolving market needs and greater customer involvement. Markets evolution towards higher personalization has forced the emergence of innovative production paradigms such as smart manufacturing and industry 4.0. In the face of increasing competition and mass customization, digitalization is becoming increasingly important to meet market challenges and achieve sufficient production efficiency. Over the past years, the industrial environment has evolved dramatically as a result of numerous technological breakthroughs. Digital Twin (DT) is at the heart of Industry 4.0. concepts and it is one of the main digital breakthroughs that is still at the beginning of its growth stage. It can be simply described as a virtual replica interacting with its physical counterpart that allows to reproduce and analyze production systems in real time. A typical DT mainly consists of three parts: physical world, digital world, and an interaction interface that connects both worlds. However, literature is lacking a universal definition of the concept as the DT construction and development is still the main focus of research. This work explored the state-of-the-art of DT applications in assembly systems through a literature review. The literature resulted ample and granular at the same time. From the 55 papers selected for the review, 34 case studies have been found in various industrial fields and laboratories. Different topics have been treated in these implementations, among which there is scheduling as the most treated topic. Furthermore, literature provides a variety of frameworks. Simulation is usually the core part of digital twins together with optimisation models as well as Big Data and Deep Learning based methods. As revealed by the literature review, the DT can provide various services for manufacturing companies. Yet, the benefits of implementation are still unclear and questionable. The presented case studies are very few and do not show the great potential of DT application. A lack of a complete case study is one of the shortcomings of the literature. Most of the applications have been mainly in high volume environments. However, when dealing with a high mix low volume (HMLV), the assembly process becomes quite challenging. Developing a cost-effective assembly system for a wide range of products is still a challenge today. It requires more flexibility and the use of breakthrough technologies that allow a wide range of products to be manufactured on the same system. Machine breakdowns, quality defects, material shortages, and variations in processing times are all examples of disruptive events in a production line. Thus, having active decision-making support is fundamental to optimally manage assembly processes and enhance resilience levels. To cover the identified gap, an industrial case study was selected for the development phase of DT. The system under study is an assembly system of a company operating in the heat generation industry for both residential and industrial uses. Consequently, based on the results of the literature review, a specific DT framework has been constructed to adapt to the requirements of the case study. The framework consists of a physical layer representing the real assembly system, a digital layer consisting of a discrete event simulation based DT and a statistical process control service, and an interaction layer composed mainly of an ERP system that enables communication and data exchange between the physical and digital layers. Utilization rate has been used as the main indicator to monitor and control the performance of the system. Individuals and moving range charts have been chosen for statistical process control as they are adequate to monitor individual measurements and their variation over time. The DT of the system has been developed using Simpy which is an open-source python-based framework for discrete-event simulation. Experiments have been conducted to assess the effectiveness of DT application to prevent capacity losses and enhance system resilience to breakdowns. Results showed that DT is an effective tool to mitigate the impact of failures on the system as both overall worktime and system throughput resulted to be 4,61% and 8,59%, respectively, higher than when the DT was not used. The differences have been verified to be statistically significant using a paired t-test. The contribution of this thesis work is the development of a practical case study in which DT has been used adopting a predictive strategy. Thanks to DT, it was possible to predict system performance and act in advance to avoid performance deterioration and prevent unrecoverable damages. This implementation has also enabled to quantify the benefits of DT application. Thus, it could be a contribution to fill the identified scientific gap in literature.
Al giorno d'oggi, i sistemi di assemblaggio si trovano ad affrontare sfide importanti, poiché i livelli di efficienza non sono soddisfacenti a causa della costante evoluzione delle esigenze del mercato e del maggiore coinvolgimento dei clienti. L'evoluzione dei mercati verso una maggiore personalizzazione ha imposto l'emergere di paradigmi di produzione innovativi, come la smart manufacturing e l'industria 4.0. Di fronte alla crescente concorrenza e alla personalizzazione di massa, la digitalizzazione sta diventando sempre più importante per affrontare le sfide del mercato e raggiungere una sufficiente efficienza produttiva. Negli ultimi anni, l'ambiente industriale si è evoluto radicalmente grazie a numerose innovazioni tecnologiche. Il Digital Twin (DT) è al centro dei concetti dell'Industria 4.0. ed è una delle principali innovazioni digitali che è ancora all'inizio della sua fase di crescita. Può essere semplicemente descritto come una replica virtuale che interagisce con la sua controparte fisica e che permette di riprodurre e analizzare i sistemi di produzione in tempo reale. Un tipico DT è composto principalmente da tre parti: mondo fisico, mondo digitale e interfaccia di interazione che collega entrambi i mondi. Tuttavia, in letteratura manca una definizione universale del concetto, poiché la costruzione e lo sviluppo del DT sono ancora al centro della ricerca. Con l'obiettivo di studiare le applicazioni del DT nei sistemi di assemblaggio, questo lavoro ha esplorato lo stato dell'arte delle applicazioni del DT nei sistemi di assemblaggio attraverso una revisione della letteratura. La letteratura è risultata ampia e granulare allo stesso tempo. Dai 55 articoli selezionati per la revisione, sono stati individuati 34 casi studio in vari settori industriali e laboratori. In queste implementazioni sono stati trattati diversi argomenti, tra i quali la schedulazione è quello più trattato. Inoltre, la letteratura fornisce una varietà di framework. La simulazione è di solito la parte centrale dei gemelli digitali, insieme ai modelli di ottimizzazione e ai metodi basati su Big Data e Deep Learning. Come emerge dall'analisi della letteratura, il DT può fornire diversi servizi alle aziende manifatturiere. Tuttavia, i benefici dell'implementazione sono ancora poco chiari e discutibili. I casi studio presentati sono limitati in numero e non mostrano il grande potenziale dell'applicazione del DT. La mancanza di casi di studio completi è una delle carenze della letteratura. La maggior parte delle applicazioni è stata realizzata principalmente in ambienti ad alto volume. Tuttavia, quando si ha a che fare con un volume basso ad alta mix (HMLV), il processo di assemblaggio diventa piuttosto impegnativo. Lo sviluppo di un sistema di assemblaggio economicamente vantaggioso per un'ampia gamma di prodotti è ancora oggi una sfida. Richiede una maggiore flessibilità e l'uso di tecnologie innovative che consentano di fabbricare un'ampia gamma di prodotti sullo stesso sistema. I guasti alle macchine, i difetti di qualità, le carenze di materiale e le variazioni nei tempi di lavorazione sono tutti esempi di eventi di disturbo in una linea di produzione. Pertanto, un supporto decisionale attivo è fondamentale per gestire in modo ottimale i processi di assemblaggio e migliorare i livelli di resilienza. Per colmare la lacuna individuata, per la fase di sviluppo del DT è stato selezionato un caso studio industriale. Il sistema oggetto di studio è un sistema di assemblaggio di un'azienda che opera nel settore della generazione di calore per usi sia residenziali che industriali. Di conseguenza, sulla base dei risultati della revisione della letteratura, è stato costruito un framework DT specifico per adattarsi ai requisiti del caso studio. Il framework è composto da un livello fisico che rappresenta il sistema di assemblaggio reale, un livello digitale che consiste in un DT basato sulla simulazione di eventi discreti e un servizio di controllo statistico dei processi, e un livello di interazione composto principalmente da un sistema ERP che consente la comunicazione e lo scambio di dati tra i livelli fisico e digitale. Il tasso di utilizzo è stato scelto come indicatore principale per monitorare e controllare le prestazioni del sistema. Per il controllo statistico dei processi sono stati scelti i grafici individuali e i grafici a intervallo mobile, in quanto adeguati a monitorare le singole misure e la loro variazione rispetto a un processo nel tempo. Il DT del sistema è stato sviluppato utilizzando Simpy, un framework open source basato su Python per la simulazione a eventi discreti. Sono stati condotti esperimenti per valutare l'efficacia dell'applicazione del DT per prevenire perdite di capacità e migliorare la resilienza del sistema ai guasti. I risultati hanno dimostrato che il DT è uno strumento efficace per mitigare l'impatto dei guasti sul sistema, in quanto il tempo di lavoro complessivo e il throughput del sistema sono risultati rispettivamente del 4,61% e dell'8,59% superiori rispetto a quando il DT non viene utilizzato. Le differenze sono state verificate in modo statisticamente significativo con un paired t-test. Il contributo di questo lavoro di tesi è lo sviluppo di un caso pratico in cui il DT è stato utilizzato adottando una strategia predittiva. Grazie al DT, è stato possibile prevedere le prestazioni del sistema e agire in anticipo per evitare il deterioramento delle prestazioni e prevenire danni irrecuperabili. Questa implementazione ha anche permesso di quantificare i benefici dell'applicazione del DT. Pertanto, questo lavoro di tesi potrebbe essere un contributo per colmare la lacuna scientifica identificata in letteratura.
Digital twin applications in MTO assembly systems : review and a practical implementation
CHAHBAN, MOHAMED
2021/2022
Abstract
Nowadays, assembly systems are facing major challenges as efficiency levels are not satisfactory due to constantly evolving market needs and greater customer involvement. Markets evolution towards higher personalization has forced the emergence of innovative production paradigms such as smart manufacturing and industry 4.0. In the face of increasing competition and mass customization, digitalization is becoming increasingly important to meet market challenges and achieve sufficient production efficiency. Over the past years, the industrial environment has evolved dramatically as a result of numerous technological breakthroughs. Digital Twin (DT) is at the heart of Industry 4.0. concepts and it is one of the main digital breakthroughs that is still at the beginning of its growth stage. It can be simply described as a virtual replica interacting with its physical counterpart that allows to reproduce and analyze production systems in real time. A typical DT mainly consists of three parts: physical world, digital world, and an interaction interface that connects both worlds. However, literature is lacking a universal definition of the concept as the DT construction and development is still the main focus of research. This work explored the state-of-the-art of DT applications in assembly systems through a literature review. The literature resulted ample and granular at the same time. From the 55 papers selected for the review, 34 case studies have been found in various industrial fields and laboratories. Different topics have been treated in these implementations, among which there is scheduling as the most treated topic. Furthermore, literature provides a variety of frameworks. Simulation is usually the core part of digital twins together with optimisation models as well as Big Data and Deep Learning based methods. As revealed by the literature review, the DT can provide various services for manufacturing companies. Yet, the benefits of implementation are still unclear and questionable. The presented case studies are very few and do not show the great potential of DT application. A lack of a complete case study is one of the shortcomings of the literature. Most of the applications have been mainly in high volume environments. However, when dealing with a high mix low volume (HMLV), the assembly process becomes quite challenging. Developing a cost-effective assembly system for a wide range of products is still a challenge today. It requires more flexibility and the use of breakthrough technologies that allow a wide range of products to be manufactured on the same system. Machine breakdowns, quality defects, material shortages, and variations in processing times are all examples of disruptive events in a production line. Thus, having active decision-making support is fundamental to optimally manage assembly processes and enhance resilience levels. To cover the identified gap, an industrial case study was selected for the development phase of DT. The system under study is an assembly system of a company operating in the heat generation industry for both residential and industrial uses. Consequently, based on the results of the literature review, a specific DT framework has been constructed to adapt to the requirements of the case study. The framework consists of a physical layer representing the real assembly system, a digital layer consisting of a discrete event simulation based DT and a statistical process control service, and an interaction layer composed mainly of an ERP system that enables communication and data exchange between the physical and digital layers. Utilization rate has been used as the main indicator to monitor and control the performance of the system. Individuals and moving range charts have been chosen for statistical process control as they are adequate to monitor individual measurements and their variation over time. The DT of the system has been developed using Simpy which is an open-source python-based framework for discrete-event simulation. Experiments have been conducted to assess the effectiveness of DT application to prevent capacity losses and enhance system resilience to breakdowns. Results showed that DT is an effective tool to mitigate the impact of failures on the system as both overall worktime and system throughput resulted to be 4,61% and 8,59%, respectively, higher than when the DT was not used. The differences have been verified to be statistically significant using a paired t-test. The contribution of this thesis work is the development of a practical case study in which DT has been used adopting a predictive strategy. Thanks to DT, it was possible to predict system performance and act in advance to avoid performance deterioration and prevent unrecoverable damages. This implementation has also enabled to quantify the benefits of DT application. Thus, it could be a contribution to fill the identified scientific gap in literature.File | Dimensione | Formato | |
---|---|---|---|
Digital twin applications in MTO assembly systems - review and practical implementation.pdf
non accessibile
Descrizione: Version 01
Dimensione
2.98 MB
Formato
Adobe PDF
|
2.98 MB | Adobe PDF | Visualizza/Apri |
Digital twin applications in MTO assembly systems - review and practical implementation.pdf
non accessibile
Descrizione: Final version
Dimensione
2.99 MB
Formato
Adobe PDF
|
2.99 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/189969