应机械结构强度与振动国家重点实验室的邀请，新加坡国立大学(National University of Singapore) 机械工程系Rajeev K. Jaiman 教授来访我院并作学术报告。

报告人： Rajeev K. Jaiman 教授

时间：2017年12月4日上午9:00- 10:00

地点：航天航空学院教一楼第四会议室

报告题目： Deep Learning and Control Strategies for Unsteady Aerodynamics

个人介绍：

Rajeev K. Jaiman is currently an Assistant Professor in the Department of Mechanical Engineering at the National University of Singapore (NUS). Prior to joining NUS in 2012, Dr. Jaiman was a Director of CFD Development at Altair Engineering, Inc., Mountain View, California. He was responsible for the development of commercial finite element based CFD solver, AcuSolve. His contributions on coupled fluid-structure interaction modeling have become known as a standard for an excellence in both industry and academia. Prior to Altair, he was a Lead CFD Developer at ACUSIM Software, Inc from 2007-2010. The technologies that Dr. Jaiman has developed are being used in wind turbine, offshore oil/gas, nuclear reactors, automotive and aerospace industries. Dr. Jaiman earned his Bachelor of Technology degree in Aerospace Engineering from the Indian Institute of Technology, Mumbai. Dr. Jaiman received his master's and doctorate degrees in Aerospace Engineering from the University of Illinois at Urbana-Champaign. In 2005, he was awarded by the Strehlow Memorial “Outstanding Researcher” recognition. In 2002 and 2004, he was nominated as a fellow of Computational Science and Engineering (CSE) department.

报告摘要：

Efficient model reduction and control of unsteady aerodynamics and flow-induced vibration is of a practical importance in aerospace engineering. Successful prediction and control of unsteady forces and aeroelastic vibrations can lead to safer, efficient and cost-effective structures. In particular, the need for such prediction and control becomes crucial for high-performance light structures with increased flexibility. For example, asymmetric vortex shedding shed from a bluff body causes a large unsteady transverse load, which in turn may lead to structural vibrations when the structure is free to vibrate. These large vortex-induced vibrations can lead to damage and potential risk to the aerospace structures. The coupled fluid-structure analysis based on high-fidelity computational fluid dynamics (CFD) can reveal a vast amount of physical insight in terms of vorticity distribution, force dynamics, frequency characteristics and phase relations. Despite improved algorithms and powerful supercomputers, the state-of-art CFD-based fluid-structure simulation is less attractive with regard to parametric optimization and the development of control strategies. Model reduction techniques naturally provide the development of adaptive control strategies for fluid-structure interaction and they are aligned with the aerospace industry needs on structural life prediction and monitoring via digital computing (e.g., digital twining of jet engines, aircraft control surfaces). The primary focus of this talk is: (i) to present an efficient low-order model for unsteady aerodynamics and flow-induced vibration, (ii) to demonstrate a feedback active control algorithm based on the dimensionality reduction for unsteady wake flow and fluid-structure interaction. Finally, I will highlight our recent success on an efficient deep learning technique for the real-time predictions of aerodynamics forces.