This thesis is concerned with the study of the Airborne Wind Energy systems (AWE): devices that use a wing connected with tethers to the ground to generate energy from the wind. The idea is to develop a realistic model of the system and, based on this model, to realize an innovative control structure for the production cycle, where the system is able to automatically adapt its operating parameters to the incoming wind speed, without measuring the latter. To perform this task, at first a multi-body model of the tether is realized and merged with a point mass model of the kite. Two different configurations of an AWE system are considered in the modelling phase: a kite with two tethers and a kite with one tether. In the first one the steering input of the system is accomplished by a part of the winch control, while in the second one the two are completely decoupled (the engine at the ground station will just reel in and out the tether and an on-board kite steering unit controls the steering angle). As a second step, considering a kite with one tether, the control structure of the production cycle has been designed, with a power cycle composed of four phases: traction, transition 1, retraction and transition 2. The control architecture features a supervisor (that decides which phase is active and the velocity reference to give to the winch) and two parallel controller loops: one is the winch control, which is responsible for the reeling movement of the tether; the other is the trajectory control, whose aim is to make the wing perform a prescribed path. After the realization of the production cycle control structure, the model has been simulated several times with different wind conditions and reeling velocities to create a dataset. The measurements of the reeling velocities, of the wind and the average power have been stored and then they have been used to compute the optimal velocities for the system (the velocities so that the corresponding average power is maximized) in different wind conditions. To do so, a radial basis function (RBF) network has been identified (for each wind condition given to the system) to model the effects of reeling speed on the generated power. Then such a function has been included in an optimization algorithm to compute the optimal reeling speed as a function of wind speed, both for the traction and the retraction phases. Finally from the optimal velocities and the corresponding tether forces,a reeling control law has been derived which achieves the final goal of the thesis project, that is to have a control system able to optimally operate the plant while not relying on wind measurements.
La tesi si concentra sullo studio di sistemi Airborne Wind Energy: dispositivi che utilizzano una vela collegata a terra con dei cavi per creare energia eolica. L’idea `e di sviluppare un modello realistico del sistema sul quale realizzare una struttura di controllo innovativa del ciclo produttivo dove il sistema `e capace di adattare automaticamente i suoi parametri operativi alla velocit`a del vento, non avendo la misura di quest’ultima. Per raggiunger questo obbiettivo, prima un modello multibody del cavo `e stato creato ed `e stato unito con un modello del kite come punto materiale. Due diverse configurazioni del sistema AWE sono state considerate nella fase di modellazione: una vela con due cavi e una vela con un cavo. Nel primo l’input di sterzo del sistema `e una parte del controllo del cavo mentre nel secondo modello sono completamente slegati (il motore nella stazione a terra arrotoler`a solo dentro e fuori il cavo mentre l’unit`a di sterzo si occuper`a dell’angolo di sterzo). In una seconda fase, considerando il modello della vela con un cavo, si realizza la struttura di controllo per il ciclo produttivo, che `e composto da 4 fasi: trazione, transizione 1, recupero, transizione 2. L’archittettura di controllo include un supervisore (che decide quale fase sar`a attiva e i riferimenti di velocit`a per il motore) e due loop di controllo paralleli: uno `e il controllo cavo che `e responsabile del movimento di srotolamento e arrotolamento del cavo; l’altro `e il controllo di traiettoria, il cui compito `e far compiere alla vela un determinato percorso. Dopo aver realizzato la struttura di controllo del ciclo produttivo, il modello `e stato simulato diverse volte con condizioni di vento e velocit`a di srotolamento/arrotolamento cavo diverse per creare un dataset. Le misure di velocit`a, vento e la potenza media generata sono state poi salvate e quindi usate per calcolare le velocit`a di srotolamento e arrotolamento ottimali per il sistema (le velocit`a tale per cui la potenza media `e massimizzata) per diverse condizioni vento. Per farlo `e stata realizzata una radial basis function (RBF) network (per ogni condizioni di vento data al sistema) per modellare gli effetti delle velocit`a del cavo sulla potenza generata. Quindi questa funzione `e stata inclusa in un algoritmo di ottimizzazione per calcolare la velocit`a ottima del cavo come funzione della velocit`a del vento, sia per la fase di trazione che per quella di recupero. Alla fine dalle velocit`a ottime e le corrispondenti forze sul cavo si ricava una legge di controllo per lo srotolamento/arrotolamento che permette di raggiungere l’obbiettivo finale della tesi, che `e avere un sistema di controllo capace di far funzionare l’impianto in modo ottimale non basandosi sulle misure del vento.
Optimal control of pumping airborne wind energy systems without wind speed feedback
BERRA, ANDREA
2019/2020
Abstract
This thesis is concerned with the study of the Airborne Wind Energy systems (AWE): devices that use a wing connected with tethers to the ground to generate energy from the wind. The idea is to develop a realistic model of the system and, based on this model, to realize an innovative control structure for the production cycle, where the system is able to automatically adapt its operating parameters to the incoming wind speed, without measuring the latter. To perform this task, at first a multi-body model of the tether is realized and merged with a point mass model of the kite. Two different configurations of an AWE system are considered in the modelling phase: a kite with two tethers and a kite with one tether. In the first one the steering input of the system is accomplished by a part of the winch control, while in the second one the two are completely decoupled (the engine at the ground station will just reel in and out the tether and an on-board kite steering unit controls the steering angle). As a second step, considering a kite with one tether, the control structure of the production cycle has been designed, with a power cycle composed of four phases: traction, transition 1, retraction and transition 2. The control architecture features a supervisor (that decides which phase is active and the velocity reference to give to the winch) and two parallel controller loops: one is the winch control, which is responsible for the reeling movement of the tether; the other is the trajectory control, whose aim is to make the wing perform a prescribed path. After the realization of the production cycle control structure, the model has been simulated several times with different wind conditions and reeling velocities to create a dataset. The measurements of the reeling velocities, of the wind and the average power have been stored and then they have been used to compute the optimal velocities for the system (the velocities so that the corresponding average power is maximized) in different wind conditions. To do so, a radial basis function (RBF) network has been identified (for each wind condition given to the system) to model the effects of reeling speed on the generated power. Then such a function has been included in an optimization algorithm to compute the optimal reeling speed as a function of wind speed, both for the traction and the retraction phases. Finally from the optimal velocities and the corresponding tether forces,a reeling control law has been derived which achieves the final goal of the thesis project, that is to have a control system able to optimally operate the plant while not relying on wind measurements.File | Dimensione | Formato | |
---|---|---|---|
tesi.pdf
accessibile in internet per tutti
Descrizione: Testo della tesi
Dimensione
5.39 MB
Formato
Adobe PDF
|
5.39 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/10589/164438