הנחיות כלליות

SCHOOL OF MECHANICAL ENGINEERING SEMINAR
Wednesday December 20.12.2023 at 14:00

ZOOM SEMINAR

Towards Active Hydrodynamic Sensing

Almog Greenberg

M.Sc. student under the supervision of Prof. Alexander Liberzon

Turbulence Structure Laboratory

School of Mechanical Engineering, Tel Aviv University, Israel

This study explores the feasibility of active hydrodynamic sensing, i.e., the ability to detect and identify the upstream object based on hydrodynamic disturbances using controlled motion inspired by the fish's mechano-sensory organ called the Lateral Line system.

We use sensitive differential pressure sensors (0.1-20 Pa) distributed along a NACA 0012 hydrofoil located in the wake of an upstream cylinder that produces a Karman Vortex Street (KVS). Additionally, we add periodic pitching at the prescribed frequency and angle of attack amplitude to enhance our ability to detect vortices from the surrounding environment. We have developed an experimental system of an open flume, enabling variable Reynolds numbers in different laminar regimes and offering optical access for Particle Image Velocimetry (PIV) measurements in the flow field.

This is a feasibility study supported by the theoretical analysis of Prof. Gregory Zilman and computational fluid dynamics (CFD) in a similar range of Reynolds numbers. Specifically, we focus on the interplay between the hydrofoil and the stable KVS. We discern contributions from the uniform stream and hydrofoil pitching motion by analyzing the hydrofoil's differential pressure signal. CFD simulations help to reduce the number of experiments and explain interplay features such as differential pressure changes with angle of attack, phase shifts between vortices, and hydrofoil pitching amplitude. Results are compared with particle image velocimetry (PIV) flow fields, validating adherence to theory and CFD assumptions. PIV also guides the selection of an optimal cylinder size for generating KVS, considering open flume wall effects within specified parameters.

The results indicate the feasibility of sensing the flow state of a stagnant fluid versus uniform flow at different angles of attack versus the wake of an upstream cylinder, which is a non-trivial hydrodynamic sensing task. Furthermore, we demonstrate that even in noisy signals, we can distinguish between the pressure due to the pitching foil only and the pressure due to the KVS approaching the hydrofoil.  We have yet to explore the different pitching amplitudes and frequencies. We are encouraged by the theoretical analysis that shows the direction of future research by controlling the pitching oscillating frequency relative to the periodicity of incoming KVS.

This research opens the way for detecting the surrounding environment using active hydrodynamic sensing based on differential pressure sensors. It has important implications for developing new technologies for environmental monitoring, underwater exploration, and oceanography.

Join Zoom Meeting

https://tau-ac-il.zoom.us/j/86497933118

School of Mechanical Engineering Seminar
Wednesday December 06.12.2023 at 14:00

ZOOM SEMINAR

Autonomous Finite Element Analyses of Femurs with Metastatic Tumors

Oren Rachmil

M.Sc. research under the supervision of Prof. Zohar Yosibash

Computational Mechanics & Experimental Biomechanics Lab

School of Mechanical Engineering, Tel Aviv University, Israel

Metastatic femoral tumors may lead to pathological fractures during daily activities. A CT-based finite element (FE) analysis of a patient’s femurs was shown to assist orthopedic surgeons in making informed decisions about the risk of fracture and the need for a prophylactic fixation. Improving the accuracy of such analyses requires an automatic and accurate segmentation of the tumors and their automatic inclusion in the FE model.

We present herein a deep learning (DL) algorithm (nnU-Net) to automatically segment lytic tumors within the femur that are then integrated into Simfini, an autonomous finite element (AFE) framework that is used in clinical practice to assist orthopedic oncologists in determining the risk of pathological femoral fractures due to metastatic tumors. A comprehensive evaluation of the DL model’s segmentation performance was assessed against two experienced musculoskeletal radiologists from Tel-Aviv Sourasky Medical Center.

The segmented lytic lesions were then integrated into Simfini that autonomously generated finite element models.  Rigorous analyses were conducted using three methodologies: the original approach without tumor segmentation, manual segmentation of lytic tumors, and the innovative automatic DL algorithm.

Results indicate that the DL algorithm may segment lytic femoral tumors in CT scans as accurately as experienced radiologists with similar DSC scores. Additionally, the incorporation of segmented tumors into an AFE model led to elevated principal strains, offering potential enhancements in predicting the risk of pathological femoral fractures. The clinical significance of tumor inclusion in autonomous FE algorithms should be further investigated in follow-on clinical research.

Join Zoom Meeting

https://tau-ac-il.zoom.us/j/86497933118