Controlling deformations of a 3D-printed vibrating beam using stiffness modulation

Finished: 2020-11-26

MSc assignment

Recently it has become possible to 3D-print complex flexible structures. An example that can be printed in principle is a robotic bird wing. Some of the materials which are available for 3D-printing are Flexible materials, fiber-reinforced filaments, (flexible) conductive materials and multi-material prints. Using fused deposition modeling (FDM) sensors can be embedded in flexible structures. At the RaM-chair capacitive and resistive sensors are already investigated. Using these new technologies, new challenges arises. The main challenges of these sensors are the non-linearity, creep, drift and hysteresis. Therefore the 3D-printed sensors are hard to characterize. Still, the sensors can be very useful. The sensors can be embedded in the wing, so no sensors have to be assembled to the wing, post-fabrication. This will influence the stiffness and aerodynamic shape of the wing less than using external sensors.

Though hard to characterize, these embedded sensors can be used as input to a control structure. Looking at examples from nature, birds can use proprioception to sense the position of their wings. Those sensors can also be imperfect| e.g. due to injuries or aging and are highly nonlinear by nature. If many imperfect sensors are distributed over the wings, they are capable of determining spatio-temporal patterns of mechanical loading. A beam with multiple embedded sensors can be seen as a simplified model system to explore the mechanics and control of a bird wing.  

This project aims to optimize a flapping motion of a flexible FDM-manufactured beam by an adaptive control system. When a change of external effects takes place, the control system can adapt, so the desired behavior is restored. The control system will use multiple, sensibly positioned embedded 3D-printed sensors as input. Two actuators will be implemented in the system. One will actuate the flapping motion, while the other one can vary the stiffness of the beam. Because there are only two actuators, which are non-collocated with the sensors. A deformation model of the beam should be designed and implemented in the control loop. This project will result in a set-up of the complete controlled system. The test results will be analyzed, validated and discussed.