Fluid-Structure Interaction Between a Piezoelectric Microactuator and a Flexible Membrane in a Fluid Channel

Abstract

An intracochlear lead-zirconate-titanate (PZT) microactuator can complement a cochlear implant electrode array to rehabilitate hearing loss patients with enhanced speech recognition. The presence of the intracochlear microactuator can significantly alter the cochlear dynamics because the actuation now results from the microactuator instead of the stapes. To understand the effect of the microactuator on the cochlear dynamics, a test rig that mimics the box model of a human cochlea was designed. The test rig consists of two connected fluid channels, one aluminum membrane sandwiched between the channels, and a PZT thin-film microactuator. Since the actuators are manufactured using a very complex process, they are not identical. To account for variability, three test rigs employing three different actuators are created. For each one, frequency response functions of the microactuator and the aluminum membrane are measured using a laser Doppler vibrometer and a spectrum analyzer. Measurements are taken when the microactuator is in air, in a petri dish surrounded by oil, and in the fluid channel inside the test rig. In addition to frequency response measurements, the distance between the actuator and aluminum membrane inside the test rig is measured using the camera with fixed focal length and adjustable position. When microactuators are moved from air into the petri dish, their natural frequencies drop because of the added mass of surrounding oil. When microactuators are moved from the petri dish (i.e., an open environment) to the inside of the fluid channels (i.e., a closed environment), their natural frequencies drop further indicating additional increase in inertia. Two out of three actuator also experience the drop in static gain when they are moved from the petri dish into the channel, indicating an increase in stiffness. After the review of the squeeze film theory, the drop in static gain is attributed to experimental inaccuracy. The cause of the inaccuracy is discussed. The inertial effect is investigated using finite element method in commercial software Ansys Acoustics. Several models are proposed. The simplest model neglects the motion of the membrane. It is used to study the effect of model uncertainties and to capture the behavior of actuators in air and petri dish. Its inability to capture the behavior of the actuators in the channel indicates that membrane motion cannot be neglected, and the inertial effect is caused by the coupling of the actuator and membrane via fluid in between. A model that includes a very stiff membrane is able to capture natural frequencies in the channel, but unable to explain the phase difference between the actuator and membrane. The membrane mode shapes are extracted using principal component analysis to explain the deficiency of the model, and to justify the need for a full model of the test rig. The full model provides a valuable insight in the physic of interaction between the actuator, fluid, and membrane inside the test rig.