Electromagnetic actuators are an interesting option for inducing vibrations in structures. However, their performance when attached to a flexible structure is relatively unknown. Therefore, this thesis aims to investigate the coupled dynamic behaviour of an electromagnetic actuator and a flexible structure. The goal is to see if the actuator can induce large displacements in the structure, in a controllable manner, and using a small input power.
Physical experiments were performed with an actuator placed on top of a flexible beam. In addition, a computer model was made that could simulate the system and predict its response. In the experiments, multiple input settings for the actuator were tested using frequency sweeps. Two different control settings were compared: the open-loop setting, which controls the current that is sent through the actuator, and the closed-loop setting, which controls the motion of the moving cylinder in the actuator. For both settings, an input signal is sent to the system. Respectively the current or the cylinder motion, relative to the tip of the beam, has to follow that input signal. The amplitude and frequency of the signal can be adjusted.
The experiments showed that there is no perfect input setting. Each setting has its advantages and disadvantages. Therefore, the best input setting to use depends on the situation. Using the open-loop setting at the resonance frequency of the system resulted in large beam tip displacements, a high effectiveness. However, this coincided with a low predictability of the displacements. On the other hand, the closed-loop setting gave a high predictability with a low effectiveness.
Looking at the efficiency of the system, the beam tip displacements normalised by the electrical input power, also did not give an ideal input setting. This was a result of the dynamic behaviour of the beam and the actuator counteracting each other a little. The beam vibrated most efficiently at its resonance frequency. However, this coincided with large relative displacements of the moving cylinder in the actuator. This generated a large Back EMF, causing the actuator to use significantly more electrical power. Thereby, the Back EMF cancels out the efficiency of the resonance.
In the closed-loop setting, the relative displacement is being controlled, and therefore it cannot increase to large values. This kills the resonance in the system, preventing the beam tip displacement from increasing.
The computer model was reasonably capable of predicting the response of the system. Due to a few inaccuracies in the model, it often slightly overestimated the displacements of the beam. However, the model showed patterns comparable to the results of the experiments, with resonance peaks at the same frequencies.
The model was also used to simulate the system response to closed-loop settings where either the beam tip displacement or the absolute motion of the cylinder was controlled, instead of the relative motion. These control settings were not possible in the physical experiments, due to limitations in the test setup. However, the model results were promising. Both of these settings could give large beam tip displacements, a high effectiveness, in combination with a high predictability. More research with physical experiments on these settings is recommended.