The digital era in which we are living is changing the way manufacturing companies are managing their physical assets. The search for operational excellence asks for holistic and integrated methodologies for the management of machines as well as embracing the digital transformation to unseal the production system state. Hence, maintenance is more central in company strategy given its capability to gather insights from the shopfloor. This is causing maintenance to evolve towards a more modern practice. Nevertheless, this evolution should be accompanied by an adequate information and data management strategy. Maintenance makes the integration of data and information a pillar so to exploit asset-related decision-making as well as to support decisions of other organisation functions, which should be grounded on the knowledge of the current production system state. Therefore, this PhD Thesis aims at investigating the management and integration of information to boost a modern maintenance practice, driven by the digital transformation and by the Industrial Asset Management. To this end, data modelling and ontology engineering are applied, each for a specific goal, but with the underpinning objective of promoting suitable management and integration of relevant information. Data modelling is used to support maintenance managers for the strategic decision regarding the eventual restructuring of the company’s maintenance processes, and to plan the integration of the required information systems. This is based on the development of a methodology that is able, through the formalisation and instantiation of a reference data model, to depict the current state of the maintenance process in terms of: completeness of the process itself, integration of the information systems, and completeness of the needed data and information. Then, ontology engineering is applied to support tactical and operational decisions. The developed ontology, called ORMA (Ontology for Reliability-centred MAintenance), rooted in the developed data model, has a modular structure, which integrates product and process knowledge in addition to the asset-related concepts. ORMA is developed thanks to AMODO (Asset Management Ontology Development methOdology), which stems from available ontology building methodologies, accompanied by a compendium that fosters knowledge reuse in the maintenance domain. ORMA can support maintenance decisions at both tactical and operational levels, also given its modularity. At tactical level, the multi-attribute criticality analysis is formalised to semantically coordinate multiple facilities. At operational level, the ontology can infer product feasibility based on current asset state to promote a shopfloor-synchronised decision-making. The proposed data and ontological models are applied and verified in real industrial manufacturing contexts. The data model is applied for strategic decisions in an automotive company to guide them towards the restructuring of their information systems stack supporting the maintenance process. Instead, ORMA is firstly applied in a food company with multiple facilities to semantically align the realisation of the criticality analysis. Then, ORMA is also applied in a Flexible Manufacturing Line where the data from the shopfloor are elaborated via health state detection algorithms, proper of Prognostics and Health Management. After updating and reasoning of the ontological model ORMA, the augmented information is displayed on a web-based dashboard for cross-functional decisions, namely, maintenance and production. Therefore, both the data model and the ontological model empower the information management and integration for maintenance in manufacturing companies. The management is specifically supported by the data model that organises where the flows of data and information. This reflects in a better organisation of the maintenance process and of its information systems. A better management of the information is promoted also by ORMA, when the terminology is fixed so to unify multiple perspectives. Moreover, ORMA empowers the integration of information given its capability to integrate augmented information from the shopfloor to support a cross-functional decision-making. Concluding, the application of data modelling and ontology engineering unveils their potentialities in helping the evolution of maintenance towards a modern practice, leveraging upon data and information pushed by the digitalisation and the managerial changes asked by the Industrial Asset Management. In so doing, maintenance could address short, medium, and long-term decision-making, centred on the asset as relevant to generate value for the company.
L’era digitale che stiamo vivendo ha cambiato il modo in cui le aziende manifatturiere stanno gestendo i loro asset fisici. La ricerca dell’eccellenza nella gestione operativa richiede metodologie olistiche ed integrate di gestione delle macchine, combinate con l’adozione di tecnologie digitali atte a discriminare lo stato dei sistemi produttivi. Perciò, la manutenzione è centrale nelle strategie aziendali, data la sua capacità di migliorare la comprensione di ciò che avviene sul campo. Questo sta spingendo la manutenzione ad evolversi verso una pratica più moderna. Nonostante ciò, tale evoluzione deve essere accompagnata da un’adeguata strategia di gestione dei dati e delle informazioni. La manutenzione ha come pilastro l’integrazione delle informazioni al fine di supportare il processo decisionale relativo agli asset fisici e di sostenere altre funzioni organizzative in diverse decisioni, le quali devono essere fondate sulla conoscenza dell’attuale stato del sistema produttivo. Pertanto, questa Tesi di Dottorato ha come obiettivo quello di investigare la gestione e l’integrazione delle informazioni per promuovere una pratica di manutenzione più moderna, guidata dalla trasformazione digitale e dall’Asset Management industriale. A tal fine, saranno applicate, ciascuna per specifici obiettivi, la modellazione dati e la modellazione semantica (ontology engineering). La modellazione dati è utilizzata per supportare i dirigenti di manutenzione in decisioni strategiche riguardanti l’eventuale ristrutturazione del processo aziendale di manutenzione e la pianificazione delle attività di integrazione dei necessari sistemi informativi. Questo si basa sullo sviluppo di una metodologia che è capace, grazie alla formalizzazione ed istanziazione di un modello dati di riferimento, di descrivere l’attuale stato del processo manutentivo in termini di: completezza del processo stesso, integrazione dei sistemi informativi, completezza dei dati e delle informazioni necessari. Dopodiché, la modellazione semantica è applicata per supportare decisioni tattiche ed operative. L’ontologia sviluppata, chiamata ORMA (Ontology for Reliability-centred MAintenance) e che trova le sue radici nel modello dati sviluppato precedentemente, ha una struttura modulare, la quale integra conoscenza del processo e del prodotto in aggiunta ai concetti relativi agli asset fisici. ORMA è sviluppata grazie alla metodologia AMODO (Asset Management Ontology Development methOdology), che deriva dalla attuale metodologia di realizzazione delle ontologie, accompagnata da un compendio che abilita il riutilizzo di conoscenza nel dominio della manutenzione. ORMA può supportare decisioni sia tattiche che operative, anche grazie alla sua modularità. A livello tattico, un’analisi di criticità multi-attributo è stata formalizzata col fine di allineare semanticamente più impianti. A livello operativo, l’ontologia è capace di dedurre la fattibilità di un prodotto basandosi sullo stato attuale dell’asset, per garantire un processo decisionale sincronizzato con il campo. I modelli proposti, sia quello dati sia quello ontologico, sono stati applicati e verificati in contesti manufatturieri reali. Il modello dati è stato sfruttato per decisioni strategiche in un’azienda nel settore automobilistico per guidarla verso la ristrutturazione del loro insieme di sistemi informativi per la manutenzione. Invece, ORMA è stata dapprima applicata in un’azienda nel settore alimentare, con all’attivo numerosi stabilimenti nel mondo, per allineare semanticamente la realizzazione dell’analisi di criticità. Infine, ORMA è stata anche implementata in una linea manifatturiera flessibile, dove i dati dal campo sono stati elaborati attraverso algoritmi di identificazione dello stato di salute, propri della disciplina di prognostica delle macchine. Dopo un aggiornamento del modello ontologico ORMA, le informazioni aumentate, tramite le capacità di deduzione del suddetto modello, sono visualizzate su un’interfaccia web per decisioni congiunte di manutenzione e produzione. Entrambi i modelli, dati ed ontologico, potenziano la gestione e l’integrazione delle informazioni per la manutenzione nelle aziende manifatturiere. La gestione è specificatamente supportata dal modello dati che organizza i flussi di dati e di informazioni. Questo si riflette in una migliore organizzazione del processo manutentivo e dei sistemi informativi. Una migliore gestione delle informazioni è promossa anche da ORMA, dove la terminologia è fissata così da unificare diverse prospettive. Oltretutto, ORMA potenzia l’integrazione delle informazioni, grazie alla sua capacità di sintetizzare le informazioni aumentate dal campo per supportare un processo decisionale integrato tra le diverse funzioni organizzative. L’applicazione dei modelli dati e semantici ha svelato le loro potenzialità ai fini di supportare il processo decisionale di manutenzione e di AM, pianificando decisioni di corto, medio e lungo termine, centrate sull’asset come generatore di valore per l’azienda.
Towards a modern maintenance practice in manufacturing by empowering information management and integration
Adalberto, Polenghi
2021
Abstract
The digital era in which we are living is changing the way manufacturing companies are managing their physical assets. The search for operational excellence asks for holistic and integrated methodologies for the management of machines as well as embracing the digital transformation to unseal the production system state. Hence, maintenance is more central in company strategy given its capability to gather insights from the shopfloor. This is causing maintenance to evolve towards a more modern practice. Nevertheless, this evolution should be accompanied by an adequate information and data management strategy. Maintenance makes the integration of data and information a pillar so to exploit asset-related decision-making as well as to support decisions of other organisation functions, which should be grounded on the knowledge of the current production system state. Therefore, this PhD Thesis aims at investigating the management and integration of information to boost a modern maintenance practice, driven by the digital transformation and by the Industrial Asset Management. To this end, data modelling and ontology engineering are applied, each for a specific goal, but with the underpinning objective of promoting suitable management and integration of relevant information. Data modelling is used to support maintenance managers for the strategic decision regarding the eventual restructuring of the company’s maintenance processes, and to plan the integration of the required information systems. This is based on the development of a methodology that is able, through the formalisation and instantiation of a reference data model, to depict the current state of the maintenance process in terms of: completeness of the process itself, integration of the information systems, and completeness of the needed data and information. Then, ontology engineering is applied to support tactical and operational decisions. The developed ontology, called ORMA (Ontology for Reliability-centred MAintenance), rooted in the developed data model, has a modular structure, which integrates product and process knowledge in addition to the asset-related concepts. ORMA is developed thanks to AMODO (Asset Management Ontology Development methOdology), which stems from available ontology building methodologies, accompanied by a compendium that fosters knowledge reuse in the maintenance domain. ORMA can support maintenance decisions at both tactical and operational levels, also given its modularity. At tactical level, the multi-attribute criticality analysis is formalised to semantically coordinate multiple facilities. At operational level, the ontology can infer product feasibility based on current asset state to promote a shopfloor-synchronised decision-making. The proposed data and ontological models are applied and verified in real industrial manufacturing contexts. The data model is applied for strategic decisions in an automotive company to guide them towards the restructuring of their information systems stack supporting the maintenance process. Instead, ORMA is firstly applied in a food company with multiple facilities to semantically align the realisation of the criticality analysis. Then, ORMA is also applied in a Flexible Manufacturing Line where the data from the shopfloor are elaborated via health state detection algorithms, proper of Prognostics and Health Management. After updating and reasoning of the ontological model ORMA, the augmented information is displayed on a web-based dashboard for cross-functional decisions, namely, maintenance and production. Therefore, both the data model and the ontological model empower the information management and integration for maintenance in manufacturing companies. The management is specifically supported by the data model that organises where the flows of data and information. This reflects in a better organisation of the maintenance process and of its information systems. A better management of the information is promoted also by ORMA, when the terminology is fixed so to unify multiple perspectives. Moreover, ORMA empowers the integration of information given its capability to integrate augmented information from the shopfloor to support a cross-functional decision-making. Concluding, the application of data modelling and ontology engineering unveils their potentialities in helping the evolution of maintenance towards a modern practice, leveraging upon data and information pushed by the digitalisation and the managerial changes asked by the Industrial Asset Management. In so doing, maintenance could address short, medium, and long-term decision-making, centred on the asset as relevant to generate value for the company.| File | Dimensione | Formato | |
|---|---|---|---|
|
PhD_Thesis_final_Adalberto Polenghi.pdf
accesso solo da BNCF e BNCR
Licenza:
Tutti i diritti riservati
Dimensione
11.08 MB
Formato
Adobe PDF
|
11.08 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/204473
URN:NBN:IT:POLIMI-204473