Soft-landing control of short-stroke reluctance actuators
- Autores: Eduardo Moya Lasheras
- Directores de la Tesis: Carlos Sagüés (dir. tes.)
- Lectura: En la Universidad de Zaragoza ( España ) en 2021
- Idioma: español
- Tribunal Calificador de la Tesis: Gonzalo López-Nicolás (presid.), Diego Puyal Puente (secret.), Johannes Maria Schellekens (voc.)
- Programa de doctorado: Programa de Doctorado en Ingeniería de Sistemas e Informática por la Universidad de Zaragoza
- Materias:
- Texto completo no disponible (Saber más ...)
- Resumen
Reluctance actuators are widely used due to their high force densities and low heat dissipation. In particular, simple short-stroke single-coil reluctance actuators, such as electromechanical relays and solenoid valves, are the best choice for on-off switching operations in many applications because of their low cost, size and mass. However, a major drawback is the strong impact at the end of each commutation, which provokes bouncing, mechanical wear and acoustic noise. They are very undesirable phenomena that detract from the evident advantages of these actuators and limit their range of potential applications.
This thesis focuses on the development and study of soft-landing control solutions for short-stroke reluctance actuators, aiming at minimizing their impact velocities. It is important to indicate that the efficiency of the aforementioned devices comes at the cost of serious theoretical and practical challenges regarding their control, e.g., fast, hybrid and highly nonlinear dynamics, complex electromagnetic phenomena, unit-to-unit variability and lack of position measurements during motion.
The starting point is the system modeling, accounting for their interconnected electrical, magnetic and mechanical subsystems. The main purpose of the models is to be used for the development of control and estimation methods. Therefore, they are lumped-parameter models expressed as state-space representations. Different electromagnetic phenomena are specified, with special attention to the magnetic hysteresis. Two model types of different complexities are proposed depending on whether the magnetic hysteresis phenomenon is incorporated or neglected.
The first approach for soft-landing control is the optimal design of position trajectories and their corresponding input signals. The proposal considers uncertainty in the contact position, and hence, the obtained solutions are more robust. While the generated input signals are effective for open-loop control strategies, the generated position trajectories can be used in feedforward or feedback control.
In order to improve the robustness of open-loop controllers, we also propose a run-to-run strategy that iteratively adapts the input signals. Specifically, it is designed to work in conjunction with a feedforward controller based on the aforementioned optimally constructed position trajectories. For the cycle-to-cycle learning algorithm, an optimization technique is chosen, adjusted and compared to two alternatives.
Another explored approach is feedback control for tracking predefined position trajectories. The proposed solution is a purely switching sliding-mode controller. The focus is on simplicity to facilitate its implementation, while also taking into account the hybrid dynamics. Theoretical and simulated analyses show that soft landing is achievable with reasonable sampling rates.
Feedback and other tracking controllers require accurate measurements or position estimations. As measuring the position is rarely practical, part of the research is devoted to the design of state estimators. The main proposal is an extended Rauch–Tung–Striebel smoother, which includes several new ideas regarding the discrete model, the inputs and the outputs. Simulated analyses demonstrate that the combined effect of the novel additions results in much better position estimations.
