Modeling and control of reluctance actuators
- Autores: Edgar Jorge Ramírez Laboreo
- Directores de la Tesis: Carlos Sagüés (dir. tes.)
- Lectura: En la Universidad de Zaragoza ( España ) en 2019
- Idioma: español
- Tribunal Calificador de la Tesis: Jesús Acero Acero (presid.), David Paesa Garcia (secret.), Jorge Luiz Duarte (voc.)
- Programa de doctorado: Programa de Doctorado en Ingeniería de Sistemas e Informática por la Universidad de Zaragoza
- Materias:
- Enlaces
- Tesis en acceso abierto en: Zaguán
- Resumen
Reluctance actuators are characterized by having a high force density, good efficiency, high fault tolerance and reduced cost. These features make them a promising alternative to other electromagnetic actuators for high-speed and high-precision applications. In addition, reluctance actuators are also ideal for small switch-type devices that require a modest performance because of their compactness, low cost, reduced mass and low energy dissipation. In particular, electromechanical switches and solenoid valves are devices whose operation is based on the force created by a small reluctance actuator.
Despite their advantages, reluctance actuators are systems with highly nonlinear dynamics. One of their most distinctive features is that the magnetic force that produces the motion is always attractive and varies greatly with the position of the armature. In essence, the nature of this force explains why switch-type devices like relays and valves are subject to strong impacts and wear each time they are operated. In addition to that, electromagnetic phenomena such as magnetic hysteresis and eddy currents make the dynamic modeling of reluctance actuators even more difficult. The work of this thesis aims to investigate the capabilities of reluctance actuators and, in particular, to analyze the dynamic behavior and propose estimation and control algorithms for electromechanical switches and solenoid valves.
The first objective of the investigation is the development of control-oriented dynamical models for reluctance actuators, i.e., low-order models that can be used to perform accurate transient simulations with low computational requirements. For that, the electromagnetic behavior of these systems is firstly studied. The magnetic equivalent circuit (MEC) methodology is selected as the primary modeling technique. Simulations from finite element models are also used for some specific purposes, e.g., to verify the assumptions of the MEC approach or to calculate the reluctance of the air gap. Then, the main electromagnetic phenomena that occur in reluctance actuators are studied. Analytic expressions are derived to model magnetic saturation, hysteresis, flux fringing and eddy currents, and an energy balance is used to obtain the expression for the magnetic force that produces the motion.
After that, the motion of the armature is incorporated to the analysis. Given that reluctance actuators usually have a limited range of motion, two different techniques are proposed to model the limits of the armature stroke and the bouncing phenomenon. Then, the electromagnetic equations and the mechanical models are combined to describe the overall dynamic behavior of the actuator. Five different dynamical models are presented, ranging from a computationally inexpensive structure to a comprehensive model that includes saturation, hysteresis, eddy currents and flux fringing. The models are compared in terms of accuracy and computational requirements.
Measurements play an important role in the analysis and characterization of dynamical systems. Thus, another objective of this thesis is the evaluation of different measurement methodologies that may improve the understanding of the dynamic behavior of reluctance actuators and, if possible, be used as part of a feedback controller. In this regard, three optical instruments are explored in order to record the motion of switch-type actuators. The results show that, even though in some cases it is possible to measure the position of several components of the device, none of the instruments could be applied in a practical situation due to their low flexibility and high cost. For that reason, other variables that are much more easily obtainable are also explored.
Another significant part of the research is devoted to estimation in reluctance actuators. Two different algorithms are proposed to estimate the magnetic flux, the resistance and the inductance of the device, both of which can be implemented in real time. The algorithms rely only on measurements of the coil voltage and current, which represents a clear advantage because no additional hardware is required. Simulation and experiments are presented to show the performance of the estimators. Furthermore, the estimation of the armature position is also investigated in this work. In particular, special focus is put on highlighting the effects of magnetic hysteresis on the performance of different estimation approaches.
Control strategies are then proposed to achieve soft landing in reluctance actuators, i.e., a controlled motion without impacts or bounces. As a first step, the basic properties of control systems theory---stability, controllability and stability---are investigated for a nominal actuator. Then, feedback linearization is explored as a method to design a trajectory tracking controller for the armature position. The obtained results show that soft landing can be accomplished by means of feedback control provided that accurate measurements or estimates of the position are available. Since this situation is rare in practice, open-loop optimal control is proposed as an alternative technique when the position is not accessible. Different time-optimal and energy-optimal solutions are derived for a nominal actuator and then compared in terms of robustness using a Monte Carlo analysis.
Finally, Run-to-Run (R2R) control is explored as another method that may be used to improve the performance of reluctance actuators. These techniques are specifically designed for systems that perform a repetitive operation and, hence, they are very well suited to being applied to switch-type devices. In particular, implicit R2R methods are based on the idea of building a function that evaluates the performance of the system at the end of each repetition. In this way, the dynamic behavior of the actuator can be gradually improved along the repetitions by conducting a black-box search. Considering that the possibilities to design a R2R controller are almost endless, practical advice is given on how to select and parameterize the input profile, how to use measurements to evaluate the system performance and how to compare different search algorithms. The performed experiments show that the designed R2R controller is able to improve greatly the behavior of a switch-type device and that, after a few cycles, it outperforms other methodologies in the literature.
