Photovoltaic (PV) plants are a key technology in reducing greenhouse gases emissions, and their adoption has spread a lot, especially in building applications. In fact, those systems can produce significant energy savings. Despite their benefits, PV plants present some fire safety risks, which, though relatively low in percentage terms, can lead to serious incidents if not managed properly. Given the large number of installations, even a small percentage of incidents can result in a non-negligible absolute number of episodes. This work conducts a detailed analysis of the main risk factors and causes for PV-related fires. It includes a statistical evaluation of fire incidents in Italy, Germany and UK. A machine learning model was trained to predict PV-related fires in Italy from 2015 to 2022 to compensate for the lack of data. The analysis compares Italian regulations with those of Germany and the UK, focusing on building regulations and PV-specific safety guidelines. The comparison includes also best practices, guidelines and studies produced by companies and research institutes. Current Italian regulations offer some fire safety guidelines for PV plants but lack a unified, comprehensive directive. It has emerged, that often not all the regulations prescriptions are applied. The findings suggest that attention to maintenance and monitoring should be boosted. Particularly, it would be necessary to oblige users to perform regular maintenance. In fact, proper maintenance could significantly improve risk management and incidents prevention. Particularly, it is shown that a very effective inspection method is the infrared thermography. That technique can detect many of the most common faults in PV plants. The present work proposes some design and maintenance requirements that can be added to Italian regulations. Furthermore, the work wants to promote advanced real-time monitoring associating it to predictive maintenance techniques. This study reviews advanced monitoring and maintenance methods, including automated systems and machine learning techniques, to enhance safety aspects and possibly to optimize maintenance.
Gli impianti fotovoltaici (FV) sono una tecnologia chiave nella riduzione delle emissioni di gas serra. In particolare, la tecnologia si è diffusa molto, soprattutto nelle applicazioni edilizie. Infatti, questi sistemi possono produrre significativi risparmi energetici. Nonostante i loro benefici, gli impianti FV presentano alcuni rischi per la sicurezza antincendio che, sebbene siano relativamente bassi in termini percentuali, possono portare a gravi incidenti se non gestiti correttamente. Dato il gran numero di installazioni, anche una piccola percentuale di incidenti può comportare un numero assoluto di episodi non trascurabile. Questa tesi esegue un'analisi dettagliata dei principali fattori di rischio e cause degli incendi correlati al FV. Include una valutazione statistica degli incendi in Italia, Germania e Regno Unito. Un modello di machine learning è stato addestrato per prevedere gli incendi correlati al FV in Italia dal 2015 al 2022 per compensare la mancanza di dati. L'analisi confronta le normative italiane con quelle di Germania e Regno Unito, concentrandosi sulle normative edilizie e sulle linee guida di sicurezza specifiche per il FV. Il confronto include anche diverse pratiche consolidate, linee guida e alcuni studi prodotti da aziende e istituti di ricerca. Le attuali normative italiane offrono alcune linee guida per la sicurezza antincendio per gli impianti FV ma mancano di una direttiva unificata e completa. È emerso che spesso non tutte le prescrizioni normative vengono applicate. I risultati suggeriscono che l'attenzione alla manutenzione e al monitoraggio dovrebbe essere aumentata. In particolare, sarebbe necessario obbligare gli utenti ad effettuare una manutenzione regolare. Infatti, una corretta manutenzione potrebbe migliorare significativamente la gestione del rischio e la prevenzione degli incidenti. In particolare, un metodo di ispezione molto efficace e consolidato prevede l’ispezione con telecamere ad infrarossi (IRT). Questa tecnica può rilevare molti dei guasti più comuni negli impianti fotovoltaici. Sono proposti alcuni requisiti di progettazione e manutenzione che possono essere aggiunti alla normativa italiana. Inoltre, il lavoro vuole promuovere il monitoraggio avanzato in tempo reale associandolo a tecniche di manutenzione predittiva. Questo studio esamina alcuni metodi avanzati di monitoraggio e manutenzione, inclusi sistemi automatizzati e tecniche di machine learning, per migliorare gli aspetti di sicurezza ed eventualmente ottimizzare la manutenzione.
Photovoltaic (PV) plants in buildings fire safety: italian regulations analysis and possibilities for improvement
Tedesco, Lorenzo
2023/2024
Abstract
Photovoltaic (PV) plants are a key technology in reducing greenhouse gases emissions, and their adoption has spread a lot, especially in building applications. In fact, those systems can produce significant energy savings. Despite their benefits, PV plants present some fire safety risks, which, though relatively low in percentage terms, can lead to serious incidents if not managed properly. Given the large number of installations, even a small percentage of incidents can result in a non-negligible absolute number of episodes. This work conducts a detailed analysis of the main risk factors and causes for PV-related fires. It includes a statistical evaluation of fire incidents in Italy, Germany and UK. A machine learning model was trained to predict PV-related fires in Italy from 2015 to 2022 to compensate for the lack of data. The analysis compares Italian regulations with those of Germany and the UK, focusing on building regulations and PV-specific safety guidelines. The comparison includes also best practices, guidelines and studies produced by companies and research institutes. Current Italian regulations offer some fire safety guidelines for PV plants but lack a unified, comprehensive directive. It has emerged, that often not all the regulations prescriptions are applied. The findings suggest that attention to maintenance and monitoring should be boosted. Particularly, it would be necessary to oblige users to perform regular maintenance. In fact, proper maintenance could significantly improve risk management and incidents prevention. Particularly, it is shown that a very effective inspection method is the infrared thermography. That technique can detect many of the most common faults in PV plants. The present work proposes some design and maintenance requirements that can be added to Italian regulations. Furthermore, the work wants to promote advanced real-time monitoring associating it to predictive maintenance techniques. This study reviews advanced monitoring and maintenance methods, including automated systems and machine learning techniques, to enhance safety aspects and possibly to optimize maintenance.File | Dimensione | Formato | |
---|---|---|---|
2024_10_Tedesco_Thesis.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: Tesi
Dimensione
11.56 MB
Formato
Adobe PDF
|
11.56 MB | Adobe PDF | Visualizza/Apri |
2024_10_Tedesco_Executive_Summary.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: Executive summary
Dimensione
1.02 MB
Formato
Adobe PDF
|
1.02 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/10589/227918