This work presents a Modelica library for the simulation of autotuning PI controllers built along the so called “contextual” approach, i.e., a technique that provides both the tuned controller and a process model suitable for simulating the closed-loop thanks to particular fidelity near the cutoff frequency. After introducing the necessary theoretical background, the backbone of the Autotuning PI (ATPI) block, with all the inputs and outputs, is defined. Plus, the procedure which is implemented in OpenModelica as an algorithm is explained in detail, and some specific implementation-related points are discussed. In the next section, the algorithm is illustrated in the form of an automaton; the different states of the system are defined and the way the system evolves between the said states is addressed. Subsequently, the possibility for the user to either accept or reject the proposed controller tuning is mentioned, by illustrating how the tuning procedure can be safely aborted. Then, two important questions regarding the implementation are addressed, namely, how to detect the permanence of an oscillation and to measure its period and amplitude. Finally, the OpenModelica implementation with the relative code is part by part shown and illustrated. Afterward, as the last chapter, some tests built upon the K.J. Åström, T. Hägglund benchmark systems are carried out and their results, which were obtained by three different control rules within the contextual approach, are described.
Questo lavoro presenta una libreria Modelica per la simulazione di controllori PI autotuning costruita secondo l’approccio detto “contestuale”, ovvero una tecnica che fornisce sia il controllore sintonizzato che un modello di processo adatto a simulare il sistema di controllo in anello chiuso grazie alla sua particolare fedeltà vicino alla frequenza di taglio. Dopo aver introdotto le necessarie basi teoriche, viene definita la struttura portante del blocco AutoTuning PI (ATPI), con tutti I suoi ingressi e le sue uscite. Inoltre, viene spiegata in dettaglio la procedura di sintesi oggetto del lavoro, implementata in OpenModelica come algoritmo; vengono anche discussi alcuni punti specifici relativi all'implementazione. Nella sezione successiva, l'algoritmo è illustrato sotto forma di automa; vengono definiti i diversi stati del sistema e viene affrontato il modo in cui il suddetto sistema evolve tra tali stati. Successivamente, viene menzionata la possibilità per l'utente di accettare o rifiutare la parametrizzazione del controlore proposta dell’autotuner, illustrando come la procedura di taratura può essere interrotta in sicurezza. Vengono quindi affrontate due importanti questioni relative all'implementazione, ovvero come rilevare la permanenza di un'oscillazione e come misurarne il periodo e l'ampiezza. Infine viene mostrata ed illustrata punto per punto l'implementazione di quento descritto in OpenModelica con il relativo codice. In seguito, come ultimo capitolo, vengono presentati alcuni test basati sul bechmark proposto da K.J. Åström e T. Hägglund e vengono descritti i loro risultati, che sono stati ottenuti da tre diverse regole di controllo nell'ambito dell'approccio contestuale.
A Modelica library for the contextual autotuning of PI controllers and process models
Amidi, Barzin
2020/2021
Abstract
This work presents a Modelica library for the simulation of autotuning PI controllers built along the so called “contextual” approach, i.e., a technique that provides both the tuned controller and a process model suitable for simulating the closed-loop thanks to particular fidelity near the cutoff frequency. After introducing the necessary theoretical background, the backbone of the Autotuning PI (ATPI) block, with all the inputs and outputs, is defined. Plus, the procedure which is implemented in OpenModelica as an algorithm is explained in detail, and some specific implementation-related points are discussed. In the next section, the algorithm is illustrated in the form of an automaton; the different states of the system are defined and the way the system evolves between the said states is addressed. Subsequently, the possibility for the user to either accept or reject the proposed controller tuning is mentioned, by illustrating how the tuning procedure can be safely aborted. Then, two important questions regarding the implementation are addressed, namely, how to detect the permanence of an oscillation and to measure its period and amplitude. Finally, the OpenModelica implementation with the relative code is part by part shown and illustrated. Afterward, as the last chapter, some tests built upon the K.J. Åström, T. Hägglund benchmark systems are carried out and their results, which were obtained by three different control rules within the contextual approach, are described.File | Dimensione | Formato | |
---|---|---|---|
Tesina_001.pdf
accessibile in internet solo dagli utenti autorizzati
Dimensione
1.41 MB
Formato
Adobe PDF
|
1.41 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/10589/182305