Aeroelastic simulation

In order to assist the design of the OVMI and speed up its development, an aeroelastic framework of the insect flight is currently in development. The aim is to provide appropriate performance items to the designer and avoid thus time-consuming trial and error iterations in his task. Several design concepts are thus evaluated and settled quickly using the aeroelastic framework as a test bench.

Principles

To be efficient and useful for designer, the framework has to be efficient, robust, modular while carrying out its task quickly and modeling the main structural and aerodynamic phenomena of the insect flight. No CFD is here done due to the high CPU requirement and load usually needed by this method and come against the goals of a preliminary design tasks of this framework.

To model the wing structure and compute its reaction to aerodynamic load, the Finite Element Method is chosen using beam and shell elements for the discretization of the venation and membrane pattern of the wing. The aerodynamic forces are computed using the kinematics data available on every nodes at each time-step.

Two aerodynamic models are available to compute depending on the required analysis and how the flexibility of the wing is accounted, as highlighted by the picture below. The first one averages the chordwise deformation, in green, and is used to quickly evaluate the aerodynamic forces generated by a specific kinematics without any aeroelastic coupling. The second one, in red, takes into account both the chordwise and the spanwise deformation enabling more accurate computation and also the aeroelastic coupling.

The framework is implemented so that every structural or aerodynamic computation is handled within the Finite Element solver. In addition to the computation itself, a Python script is integrated into the framework to manage and oversee each analysis properly, autonomously and efficiently. Thus the pre- and the post-processing are automatic reducing drastically the user interaction time and load. Thus various wing geometry can be quickly evaluated as seen below and gives performance items such the mean lift or power consumption.

Assisting design choice

Among the wide range of applications envisioned for our design tool, a striking application of its capability is to assist the designer in choosing the appropriate actuation strategy for its MAV or NAV requirements. The actuation strategy is one of the core design choice as it affects the entire vehicle through the number and the size of each actuator, the number of mechanical link needed and also its energy consumption.

In the case of the OVMI, a heaving actuation was originally used and thanks to the simulation made on a arbitrary wing, a flapping actuation is yet implemented favoring higher wing kinematics and aerodynamic forces, as highlighted in the video below.

Optimization

Given the thousands of insects flying, another application of our aeroelastic framework is to drive the designer through the broad but unknown design space meeting his requirements. A genetic algorithm is here used so as to scan easily this design space while avoiding the expected large number of local extrema. Considering the modularity of our implementation, an additional Python layer is used through the Pyevolve module. Several optimization problems are now addressed quite satisfactorily as seen below for an mean lift optimized wing.

For more details, please see our papers below.

Reference papers: