Robust Design Optimization of Labyrinth Seal Using Machine Learning -Improvement of Leakage Performance and Aeroelastic Stability-

In aero engines, rotating labyrinth seals are widely used to suppress the flow leakage between static and rotating components. These seals are usually designed with light weight to further enhance the engine performance. Hence, they are susceptible to flutter vibration due to their low stiffness. In this study, the ways to optimize both seal leakage performance and flutter margin are investigated using fluid -structure interaction (FSI) simulation and machine learning. Through this study, it is found that by optimizing sub cavity geometries between seal fins, both the leakage performance and flutter margin can be effectively improved. NOMENCLATURE A Area, m2 a Speed of sound, m/s D Depth of seal cavity, m E Energy, J f Frequency, Hz G Mass flow rate, kg/s L Axial length of seal cavity, m N Rotational speed, rpm P Pressure, Pa T Period, s t Time, s V Velocity, m/s Greek π Seal pressure ratio δ Logarithmic decrement ζ Damping ratio θ Angle Subscripts ac Acoustic aero Aerodynamic B Backward mode b Base case cyc One vibration cycle F Forward mode ftip Seal fin tip in Inlet mech Mechanical s Sub cavity su Blade surface sl Slip t Total th Circumferential component Abbreviations CFD Computational fluid dynamics FSI Fluid structure interaction ND Nodal diameter