The work is set in a context in which digitization has become a driving force for change and improvement in many areas as a result of the fourth industrial revolution. The study aims to develop a new methodology for extracting information from the manufacturing systems based on a comparison between the nominal behavior of the system and the real one. Process Mining techniques, such as Conformance Checking and Model Enhancement, can be used to make such a comparison. The existing methodology coupled these techniques with Process Discovery, a technique that is advantageous when applied in contexts in which the process model is not known a priori. Since manufacturing processes are known a priori, the work proposes to replace Process Discovery with a system modeling phase based on the study of available process knowledge. The knowledge is translated into a Petri Net model representing the manufacturing system and it is used to highlight deviations with the actual process. In this study, it is intended to demonstrate that the developed methodology can avoid classical problems presented by the existing methodology by better extracting information from process data. This demonstration is developed by studying the results obtained applying the methodology to a case study and through an experiment aimed at comparing the two methodologies.
Il lavoro si inserisce in un contesto in cui, a seguito della quarta rivoluzione industriale, la digitalizzazione è diventata una forza motrice di cambiamento e miglioramento in molti ambiti. Lo studio vuole sviluppare una nuova metodologia di estrazione delle informazioni dai sistemi manifatturieri, basata sul confronto tra il comportamento nominale del sistema e quello reale. Per effettuare tale confronto è possibile sfruttare tecniche di Process Mining come il Conformance Checking ed il Model Enhancement. La procedura attuale accoppia queste tecniche al Process Discovery, tecnica che risulta vantaggiosa se applicata in contesti in cui il modello di processo non è noto a priori (metodologia esistente). Essendo i processi manifatturieri noti a priori, lo studio propone di sostituire il Process Discovery con una fase di modellazione del sistema basato sullo studio della conoscenza aziendale disponibile. Tale conoscenza viene tradotta in un modello di rete di Petri che rappresenta il sistema di produzione e viene usata per evidenziare deviazioni tra il processo reale e la conoscenza nominale del processo. In questo studio si vuole dimostrare che la metodologia sviluppata può evitare problematiche classiche che la metodologia esistente presenta, migliorando l'estrazione di informazioni dai dati di processo. Tale dimostrazione viene sviluppata studiando i risultati ottenuti dall'applicazione della metodologia ad un caso studio e attraverso un esperimento volto al confronto delle due metodologia.
Knowledge-based Model Enhancement through Conformance Checking techniques in Manufacturing Systems
Spilotros, Lorenzo;Urbinati, Daniele
2021/2022
Abstract
The work is set in a context in which digitization has become a driving force for change and improvement in many areas as a result of the fourth industrial revolution. The study aims to develop a new methodology for extracting information from the manufacturing systems based on a comparison between the nominal behavior of the system and the real one. Process Mining techniques, such as Conformance Checking and Model Enhancement, can be used to make such a comparison. The existing methodology coupled these techniques with Process Discovery, a technique that is advantageous when applied in contexts in which the process model is not known a priori. Since manufacturing processes are known a priori, the work proposes to replace Process Discovery with a system modeling phase based on the study of available process knowledge. The knowledge is translated into a Petri Net model representing the manufacturing system and it is used to highlight deviations with the actual process. In this study, it is intended to demonstrate that the developed methodology can avoid classical problems presented by the existing methodology by better extracting information from process data. This demonstration is developed by studying the results obtained applying the methodology to a case study and through an experiment aimed at comparing the two methodologies.File | Dimensione | Formato | |
---|---|---|---|
2022_12_Executive_Summary_Spilotros_Urbinati.pdf
accessibile in internet per tutti
Descrizione: Executive Summary
Dimensione
2.35 MB
Formato
Adobe PDF
|
2.35 MB | Adobe PDF | Visualizza/Apri |
2022_12_Master_Thesis_Spilotros_Urbinati.pdf
accessibile in internet per tutti
Descrizione: Master Thesis
Dimensione
12.39 MB
Formato
Adobe PDF
|
12.39 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/10589/198654