Mathematics-informed neural network for 2x2 matrix factorisation and a new Wiener-Hopf model for aeroengine noise radiation

Add to your list(s) Download to your calendar using vCal

• Xun Huang (Peking University)
• Tuesday 02 July 2024, 11:45-12:15
• Seminar Room 1, Newton Institute.

If you have a question about this talk, please contact nobody.

WHTW02 - WHT Follow on: the applications, generalisation and implementation of the Wiener-Hopf Method

In this presentation, I will discuss two recent advancements in my group’s research on Wiener-Hopf modelling and applications. First, we embed Cauchy-Riemann equations into machine learning architectures using so-called mathematics-informed neural networks (MINN). By integrating Cauchy integrals and boundary conditions, the neural network effectively learns to numerically generate 2×2 matrix kernel factorization for Wiener-Hopf analytical models. We validate this approach by comparing machine learning results with existing analytical solutions using a benchmark case of wave scattering from parallel hard-soft plates.  Second, we develop a Wiener-Hopf model to sound radiation from an elliptic duct, which has garnered interest in the context of modern blended-wing-body designs for eco-friendly aviation. We represent the incident and scattered sound fields using Mathieu functions in elliptic cylindrical coordinates and validate the achieved Wiener-Hopf analytic model by comparing with FEM results.  Lastly, I will provide a brief introduction (travel, venue, Visa matters, etc.) of Hangzhou, China, the scenic city Marco Polo once described as “greater than any in the world”, for the upcoming Wiener-Hopf workshop in November. My co-authors, Mr. Sicong Liang and Mr. Ruichen Wang, will attend INI ’s workshop and be available to share more information about these topics.

This talk is part of the Isaac Newton Institute Seminar Series series.

This talk is included in these lists:

This talk is not included in any other list

Note that ex-directory lists are not shown.