The mainstream theories of aeroacoustics are all based on the advective wave equations for thermodynamic variables. However, these equations do not cover the fluid flow that generates and propagates sound waves. which leads to the the main theoretical obstacle to identifying sound source and achieving effective noise control. This project is to study the physical diagnosis of aeroacoustic noise of complex flow via theoretical analysis and direct numerical simulations, aiming to improve the theoretical basis of viscous gas dynamics and achieve a breakthrough in the existing dilemma of aeroacoustics.