Optimization of the automotive air conditioning system using radial basis function neural network

Abstract

The defrosting performance of automotive air conditioners plays an important role in driving safety. This paper uses computational fluid dynamics (CFD) to simulate the internal flow field of the automobile numerically. Simulation results show that the flow distribution is unreasonable. The horizontal grilles are added at the outlets to improve the defrosting performance of the automobile. Airflow jet angle and the length of the air conditioning outlets (L1, L2) are selected as design variables based on the radial basis neural network to find the optimal combination scheme. The area of the defrosting dead corner has been reduced from 20% to 5% after optimization, and the frost layer of the front windshield has been completely melted in 25 min. The experiment test is conducted to verify the improvement of the defrosting performance of automotive air conditioners. The design methodology can be applied to the development of the air conditioner.