Three-dimensional finite-time Lyapunov-exponent analysis of fluid dusty plasmas under weightlessness using a machine-learning particle reconstruction technique.

We have performed experiments on dusty plasmas under the weightlessness conditions of parabolic flights where the dust particles form an extended homogeneous dust cloud. The three-dimensional (3D) dynamic state of the dust cloud is characterized. Therefore, the particle trajectories have been recorded using a four-camera stereoscopic camera system. From that, the 3D particle trajectories have been determined using both a machine-learning particle reconstruction technique and the deterministic shake-the-box algorithm. From the trajectories, characteristic fluid parameters, such as flow fields and finite-time Lyapunov exponent (FTLE)-based fluid structures have been calculated and analyzed. The FTLE analysis indicates that the fluid is characterized by an incompressible flow with small-scale behavior. Furthermore, it is demonstrated that the machine-learning based approach allows to reliably characterize the dynamic states by comparison with the shake-the-box algorithm.