In order to perform a robustness analysis of the controller, it is necessary to include uncertainties in the plant model. An effective approach to achieve this is by employing a Linear Fractional Transformation (LFT) to model the plant. Moreover, real-case applications foresee the implementation of discrete-time controllers, which need to be integrated with the continuous-time signals of the LFT plant. Consequently, the discretization of the LFT system becomes necessary and it shall be performed by means of a zero-order-hold actuation and sampling. However, obtaining an exact discrete-time (DT) projection of an LFT system poses inherent challenges. Therefore, this thesis aims to identify novel strategies that enable the adoption of various discretization methods. Two main approaches for discretization are explored and compared in terms of discretization stability and error. To validate the proposed methods, they are applied to two satellite models: a SISO 1D mass-spring-damper model and a MIMO 3D model of the meteorological satellite MetOp-SG SAT-B. The analysis demonstrates that it is possible to accurately discretize an LFT plant using approximation methods such as Padé and Hanselmann, but these methods can lead to increased LFT complexity.
Per effettuare un'analisi di robustezza del controllore, è necessario includere le incertezze nel modello del sistema dinamico. Un approccio efficace per ottenere ciò è utilizzare una Trasformazione Lineare Frazionaria (LFT) per modellare l'impianto. Inoltre, le applicazioni reali prevedono l'implementazione di controllori a tempo discreto, che devono essere integrati con i segnali a tempo continuo dell'impianto LFT. Di conseguenza, diventa necessaria la discretizzazione del sistema LFT, che viene effettuata mediante lo zero order hold. Tuttavia, ottenere una proiezione discreta esatta (DT) di un sistema LFT presenta sfide intrinseche. Perciò questa tesi si propone di identificare nuove strategie che consentano l'adozione di vari metodi di discretizzazione. Due approcci principali per la discretizzazione vengono esplorati e confrontati in termini di stabilità e errore di discretizzazione. Per convalidare i metodi proposti, essi vengono applicati a due modelli di satelliti: un modello massa-molla-smorzatore SISO (single-input single-output) 1D e un modello MIMO (multiple-input multiple-output) 3D del satellite meteorologico MetOp-SG SAT-B. L'analisi dimostra che è possibile discretizzare accuratamente un impianto LFT utilizzando metodi di approssimazione come Padé e Hanselmann, ma questi metodi possono portare a un aumento della complessità del sistema LFT.
Modelling of uncertain satellite plants for discrete control loop analysis
GIULIANI, VICTORIA KATIA
2022/2023
Abstract
In order to perform a robustness analysis of the controller, it is necessary to include uncertainties in the plant model. An effective approach to achieve this is by employing a Linear Fractional Transformation (LFT) to model the plant. Moreover, real-case applications foresee the implementation of discrete-time controllers, which need to be integrated with the continuous-time signals of the LFT plant. Consequently, the discretization of the LFT system becomes necessary and it shall be performed by means of a zero-order-hold actuation and sampling. However, obtaining an exact discrete-time (DT) projection of an LFT system poses inherent challenges. Therefore, this thesis aims to identify novel strategies that enable the adoption of various discretization methods. Two main approaches for discretization are explored and compared in terms of discretization stability and error. To validate the proposed methods, they are applied to two satellite models: a SISO 1D mass-spring-damper model and a MIMO 3D model of the meteorological satellite MetOp-SG SAT-B. The analysis demonstrates that it is possible to accurately discretize an LFT plant using approximation methods such as Padé and Hanselmann, but these methods can lead to increased LFT complexity.File | Dimensione | Formato | |
---|---|---|---|
2023_07_Giuliani_01.pdf
accessibile in internet solo dagli utenti autorizzati
Dimensione
6.16 MB
Formato
Adobe PDF
|
6.16 MB | Adobe PDF | Visualizza/Apri |
2023_07_Giuliani_02.pdf
accessibile in internet solo dagli utenti autorizzati
Dimensione
2.53 MB
Formato
Adobe PDF
|
2.53 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/10589/207537