32th Congress of the International Council of the Aeronautical Sciences

05 - Propulsion

APPLICATION OF REINFORCEMENT LEARNING IN 1-D AERODYNAMIC DESIGN OF AXIAL COMPRESSOR

Y. Liu╣, X. Hang╣, C. Jiang╣; ╣Beihang University, China

The reinforcement learning algorithm extracts and saves the design experience into the artificial neural network through a large number of interactions with the one-dimensional compressor aerodynamic calculation program (HARIKA) to complete the one-dimensional design. we used the DDPG to accelerate the aerodynamic design process of axial compressors. The adiabatic efficiency is improved by 8.7% and the surge margin reached 29%, which verifies the effectiveness of the design method.


View Paper