The manufacturing industry makes up a significant proportion of many of the developed countries, including Italy and Sweden. In this sector, maintenance is a critical factor both from an economic and a management point of view, due to the high expenses and the difficulty in optimising it. While there has been a continuous growth in academic interest for machine condition monitoring and its integration with advanced maintenance strategies, its adoption by small and large enterprises is null or ineffective for most of the industrial applications based on industrial feedback. The lack of knowledge, the absence and low reliability of the data and the inadequate technology infrastructure that often affect the companies, are the main reason at the base of the low success rate of improvement projects in maintenance management. The research gap has been identified after a literature review of the state of the art of different maintenance strategies and performance monitoring, moreover, several interviews with industry experts have been conducted in order to _ll the gap between the academic approach and the practical needs of the companies. The largest part of the models showed in the literature has a deterministic approach, with the related assumptions of simplification, while others perform a cost-benefit analysis on already implemented or tested projects, obtaining useful results but losing the future outlook. This thesis aims to provide a useful methodology for supporting the decision processes related to the investment within machinery equipment. The model developed carries out an economic assessment of the impact that the selected maintenance strategy has on the Life Cycle Cost of the machine while handling the inevitable uncertainty that characterises the input data. The output and the statistical insights will help the managers during the decision process providing a long term analysis performed through the Monte Carlo method. The input data, together with the necessary estimations have been provided by the companies and their experts or, for the model-verification scope, obtained from the literature. On the output of the simulation, multiple sensitivity analyses have been carried out on the most relevant parameter to present a complete overview of the different scenarios and to help the manager in setting specific goals to be reached in a certain amount of time. Furthermore, the productivity performances are computed through a Markov Chain model, so that a further statistical insight could be provided to the management. Finally, the developed model has been implemented in a case study in collaboration with a large manufacturing company. The data acquisition process showed the lack of data available for this type of analysis, as well as the inefficient or absent methodical systems to monitor several important cost parameters as, for example, the overall downtime cost. Further development possibilities include the correlation analysis and modelling of the cost elements, the reduction of the data aggregation level and the application of the methodology for every single failure mode.
L'industria manifatturiera costituisce una parte significativa del PIL di molti dei paesi sviluppati, tra cui l'Italia e la Svezia. In questo settore, la manutenzione è un fattore critico sia da un punto di vista economico che gestionale, a causa delle alte spese e della difficoltà nella sua ottimizzazione. Mentre c'è stata una continua crescita dell’interesse accademico per il monitoraggio delle condizioni delle macchine e la sua integrazione con strategie di manutenzione avanzate, la sua adozione da parte di piccole e grandi imprese è nulla o inefficace per la maggior parte delle applicazioni industriali. La mancata conoscenza, l'assenza e la bassa affidabilità dei dati oltre che l'inadeguata infrastrutture tecnologiche che spesso colpiscono le aziende, sono la motivazione principale del basso tasso di successo dei progetti di miglioramento delle operazioni manutentive. Dopo una revisione della letteratura sullo stato dell'arte delle diverse strategie di manutenzione e del monitoraggio delle prestazioni, oltre che dopo diverse interviste con esperti del settore un grande divario è stato evidenziato tra l'approccio accademico e le esigenze pratiche delle aziende. La maggior parte parte dei modelli mostrati in letteratura si basano su modelli deterministici, i quali necessitano molteplici processi di semplificazione, mentre altri eseguono un'analisi costi-benefici su progetti già implementati, ottenendo risultati utili ma perdendo l’applicabilità su prospettive future. Questa tesi mira a fornire una metodologia utile per supportare i processi decisionali relative alle politiche di investimento nell’ambito delle attrezzature per la produzione manufatturiera. Il modello sviluppato consente di effettuare una valutazione economica dell’impatto che la strategia manutentiva ha sul costo del ciclo di vita della macchina, gestendo l’incertezza che affligge i dati in input. I risultati e molteplici approfondimenti statistici sono di support al management, fornendo un sistema analitico che fornisca un’analisi sulle probabilità di implementazione profittevole di una determinate strategia. Per ottenere gli output appena descritti, il metodo Monte Carlo e la modellizzazione Markoviana vengono eseguiti tramite un codice matematico scritto in MatLab. Infine, il modello sviluppato è stato implementato in un caso studio redatto con la collaborazione di una grande impresa manufatturiera Svedese, attiva nell’ambito delle machine utensili. Il processo di acquisizione delle informazioni necessarie, ha mostrato la grande carenza di competenze e dati disponibili per questo tipo di analisi, così come l’inefficiente o assente utilizzo di sistema di monitoraggio dei parametri di costo. Ulteriori possibilità di sviluppo includono l’analisi di correlazione tra gli elementi di costo, la riduzione del livello di aggregazione dei dati e l’applicazione della metodologia presentata per ogni tipologia di guasto.
Uncertainty propagation for machine tool life cycle cost assessment in case of different maintenance strategies
LUCCHESE, PIETRO
2020/2021
Abstract
The manufacturing industry makes up a significant proportion of many of the developed countries, including Italy and Sweden. In this sector, maintenance is a critical factor both from an economic and a management point of view, due to the high expenses and the difficulty in optimising it. While there has been a continuous growth in academic interest for machine condition monitoring and its integration with advanced maintenance strategies, its adoption by small and large enterprises is null or ineffective for most of the industrial applications based on industrial feedback. The lack of knowledge, the absence and low reliability of the data and the inadequate technology infrastructure that often affect the companies, are the main reason at the base of the low success rate of improvement projects in maintenance management. The research gap has been identified after a literature review of the state of the art of different maintenance strategies and performance monitoring, moreover, several interviews with industry experts have been conducted in order to _ll the gap between the academic approach and the practical needs of the companies. The largest part of the models showed in the literature has a deterministic approach, with the related assumptions of simplification, while others perform a cost-benefit analysis on already implemented or tested projects, obtaining useful results but losing the future outlook. This thesis aims to provide a useful methodology for supporting the decision processes related to the investment within machinery equipment. The model developed carries out an economic assessment of the impact that the selected maintenance strategy has on the Life Cycle Cost of the machine while handling the inevitable uncertainty that characterises the input data. The output and the statistical insights will help the managers during the decision process providing a long term analysis performed through the Monte Carlo method. The input data, together with the necessary estimations have been provided by the companies and their experts or, for the model-verification scope, obtained from the literature. On the output of the simulation, multiple sensitivity analyses have been carried out on the most relevant parameter to present a complete overview of the different scenarios and to help the manager in setting specific goals to be reached in a certain amount of time. Furthermore, the productivity performances are computed through a Markov Chain model, so that a further statistical insight could be provided to the management. Finally, the developed model has been implemented in a case study in collaboration with a large manufacturing company. The data acquisition process showed the lack of data available for this type of analysis, as well as the inefficient or absent methodical systems to monitor several important cost parameters as, for example, the overall downtime cost. Further development possibilities include the correlation analysis and modelling of the cost elements, the reduction of the data aggregation level and the application of the methodology for every single failure mode.File | Dimensione | Formato | |
---|---|---|---|
Uncertainty propagation for machine tool life cycle cost assessment in case of different maintenance strategies.pdf
accessibile in internet solo dagli utenti autorizzati
Dimensione
8.14 MB
Formato
Adobe PDF
|
8.14 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/10589/183347