Research on satellite anomaly detection is growing, especially to enhance anomaly detection methods. To improve the understanding of a satellite's behaviour under both nominal and anomalous conditions is an essential step, given the limited possibility for direct intervention during missions. This work presents the development of a comprehensive digital model of a satellite that helps bridge the gap between theoretical analysis and real-life satellite telemetry. Starting from an existing CubeSat lumped-parameter thermal model implemented in Matlab and Simulink, the framework was redesigned and extended to include key subsystems such as propulsion, reaction wheels, energy storage units, and solar cells. A fictitious mission scenario was presented to assess the satellite’s response to injected anomalies, with a focus on their impact on telemetry data. This approach demonstrates the model’s potential to support both operational optimization and anomaly analysis, providing insights into the type, severity, and consequences of anomalous events. The resulting datasets of artificial telemetry can serve as a valuable resource for the development and testing of anomaly detection algorithms, including machine-learning-based approaches. The study highlights that while the current model enables meaningful exploration of fault scenarios, further refinement could increase model fidelity, which in turn would improve the model’s representativeness of real satellite behaviour and broaden its applicability to more complex missions and anomaly types.
La ricerca sul rilevamento di anomalie nei satelliti artificiali è in crescita, soprattutto per migliorare i metodi di rilevamento delle anomalie stesse. Migliorare la comprensione del comportamento di un satellite in condizioni sia nominali che anomale è un passo essenziale, data la limitata possibilità di intervento diretto durante le missioni. Questo lavoro presenta lo sviluppo di un modello digitale completo di un satellite che aiuta a colmare il divario tra l'analisi teorica e la telemetria satellitare reale. Partendo da un modello termico a parametri concentrati CubeSat esistente implementato in Matlab e Simulink, il framework è stato riprogettato ed esteso per includere sottosistemi chiave quali propulsione, ruote di reazione, unità di accumulo di energia e celle solari. La risposta del satellite alle anomalie iniettate è stata osservata in una missione fittizia, con particolare attenzione all'impatto delle anomalie sui dati telemetrici. Questo approccio dimostra il potenziale del modello di supportare sia l'ottimizzazione operativa che l'analisi delle anomalie, fornendo informazioni sul tipo, la gravità e le conseguenze degli eventi anomali. I set di dati telemetrici sintetici risultanti possono costituire una risorsa preziosa per lo sviluppo e il collaudo di algoritmi di rilevamento delle anomalie, compresi gli approcci basati sul machine learning. Lo studio evidenzia che, sebbene il modello attuale consenta un'analisi significativa degli scenari di guasto, un ulteriore perfezionamento potrebbe aumentarne la fedeltà, migliorandone la rappresentatività del comportamento reale dei satelliti e ampliandone l'applicabilità a missioni e tipi di anomalie più complessi.
Modeling of a satellite for generation of artificial time series telemetry data
Rune, Joel
2024/2025
Abstract
Research on satellite anomaly detection is growing, especially to enhance anomaly detection methods. To improve the understanding of a satellite's behaviour under both nominal and anomalous conditions is an essential step, given the limited possibility for direct intervention during missions. This work presents the development of a comprehensive digital model of a satellite that helps bridge the gap between theoretical analysis and real-life satellite telemetry. Starting from an existing CubeSat lumped-parameter thermal model implemented in Matlab and Simulink, the framework was redesigned and extended to include key subsystems such as propulsion, reaction wheels, energy storage units, and solar cells. A fictitious mission scenario was presented to assess the satellite’s response to injected anomalies, with a focus on their impact on telemetry data. This approach demonstrates the model’s potential to support both operational optimization and anomaly analysis, providing insights into the type, severity, and consequences of anomalous events. The resulting datasets of artificial telemetry can serve as a valuable resource for the development and testing of anomaly detection algorithms, including machine-learning-based approaches. The study highlights that while the current model enables meaningful exploration of fault scenarios, further refinement could increase model fidelity, which in turn would improve the model’s representativeness of real satellite behaviour and broaden its applicability to more complex missions and anomaly types.| File | Dimensione | Formato | |
|---|---|---|---|
|
2025_10_Rune.pdf
accessibile in internet per tutti
Dimensione
5.75 MB
Formato
Adobe PDF
|
5.75 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/10589/243909