ETD

Archivio digitale delle tesi discusse presso l'Università di Pisa

Tesi etd-01272005-194330


Tipo di tesi
Tesi di laurea vecchio ordinamento
Autore
Vecchio, Enrica
URN
etd-01272005-194330
Titolo
SPACECRAFT FAULT ANALYSIS USING DATA MINING TECHNIQUES
Dipartimento
INGEGNERIA
Corso di studi
INGEGNERIA INFORMATICA
Relatori
relatore Donati, Alessandro
relatore Prof. Marcelloni, Francesco
relatore Prof.ssa Lazzerini, Beatrice
Parole chiave
  • Association Rules
  • Apriori
  • Spacecraft
  • CLUSTER-II Mission
  • Spacecraft Telemetries
  • K-means
  • Pattern Clustering
  • Data Mining
Data inizio appello
03/03/2005
Consultabilità
Non consultabile
Data di rilascio
03/03/2045
Riassunto
In this work I have demonstrated the possibility to apply Data Mining techniques to the analysis of spacecraft HKTM data, in order to find out typical patterns of different parameters, which can be considered correlated to the fault occurrence.
The proposed Fault Analysis Method has been applied to CLUSTER II Mission system, consisting of four spacecraft orbiting on an elliptical orbit around the Earth. In 2001 and 2002, during some manoeuvres of spacecraft, the ESOC’s Operation Team had sometimes observed anomalous temperature drop of one of the fired thrusters. The anomalous drop has been linked to a leakage of propellant through the thruster valve seals.
Moreover the prototype of a software tool, completely written in MatLab, has been designed and developed. The Tool implements automatically all the elaborations required by the phase of Data Preparation and by the used data mining algorithms. The Tool is provided with a user-friendly MMI, consisting of GUI Windows, which allows an easy usage.
Using the Tool, I have analysed the HKTM data archived during the execution of 179 spacecraft manoeuvres. As result, I have found a series of Association Rules between the fault condition and behaviour patterns of different spacecraft parameters.
Then I have successfully validated the new method, by applying the found association rules to a different set of HKTM data, corresponding to manoeuvers on a different time period.
The Tool has been customised to the characteristics of the CLUSTER’s mission and problem, but it can be easily modified and adapted to different missions and different types of spacecraft fault investigation. A possible future development of this work could be the application to the real-time HKTM data flow from the satellite.
The work has been developed in ESOC, the European Space Operation Control of the European Space Agency (ESA) based in Darmstadt, Germany. The project has been part of the technology promotional activities carried out at the Mission Control Technologies Unit (OPS-OSC).
File