The thesis presents the development of an automated system for analyzing test data (DHR) from electronic devices and corresponding technical intervention reports, with the goal of identifying potential correlations between initial configuration parameters and malfunctions observed in the field. The proposed approach is divided into two main phases: the extraction of test data and their transformation into a structured format, and the training of a Machine Learning model capable of detecting recurring patterns between test parameters and documented failure events. The thesis provides also a detailed description of the test procedures performed on the devices, the recorded technical parameters (power, energy, and pulse duration), and the methods used to acquire and structure the data, also highlighting challenges related to the unique identification of physical devices. The results demonstrate the feasibility of a predictive system based on historical data, setting the stage for future applications in quality monitoring and predictive maintenance.
Questa tesi presenta lo sviluppo di un sistema automatico per l’analisi dei dati di collaudo (DHR) degli apparati e dei relativi report tecnici di intervento, allo scopo di individuare eventuali correlazioni tra configurazioni iniziali e malfunzionamenti riscontrati sul campo. L’approccio proposto si divide in due fasi principali: l’estrazione dei dati di collaudo e la loro trasformazioni in un formato strutturato, e l’addestramento di un modello di Machine Learning in grado di rilevare pattern ricorrenti tra i parametri di test e gli eventi di guasto documentati. La tesi descrive in dettaglio le tipologie di test effettuati sugli apparati, i parametri tecnici rilevati (potenza, energia, durata dell’impulso) e le modalità di acquisizione dei dati, vengono inoltre analizzate le criticità legate all’identificazione univoca dei dispositivi. I risultati ottenuti provano la fattibilità di un sistema predittivo basato sui dati storici, aprendo la strada a future applicazioni nel monitoraggio della qualità e nella manutenzione preventiva degli apparati.
AI-Driven Technical Report Analysis and Failure Prediction: A Case Study in Intelligent Maintenance System
BRUNETTI, LEONARDO
2024/2025
Abstract
The thesis presents the development of an automated system for analyzing test data (DHR) from electronic devices and corresponding technical intervention reports, with the goal of identifying potential correlations between initial configuration parameters and malfunctions observed in the field. The proposed approach is divided into two main phases: the extraction of test data and their transformation into a structured format, and the training of a Machine Learning model capable of detecting recurring patterns between test parameters and documented failure events. The thesis provides also a detailed description of the test procedures performed on the devices, the recorded technical parameters (power, energy, and pulse duration), and the methods used to acquire and structure the data, also highlighting challenges related to the unique identification of physical devices. The results demonstrate the feasibility of a predictive system based on historical data, setting the stage for future applications in quality monitoring and predictive maintenance.| File | Dimensione | Formato | |
|---|---|---|---|
|
Brunetti_Leonardo_2054123.pdf
accesso aperto
Dimensione
2.88 MB
Formato
Adobe PDF
|
2.88 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/86903