음향데이터 활용한 AI기반 전기차 인휠모터 상태 진단기술 개발

Abstract

Of late, the automotive industry has been rapidly transitioning from traditional internal combustion engines to hybrid and electric vehicles. Adoption of electric vehicles has already entered a growth phase and is evolving into various forms to suit different purposes. Particularly, to improve the utilization of interior space in electric vehicles, research is being conducted on weight reduction and structural changes of motor systems. The in-wheel motor mechanism is characterized by an independent motor drive system applied to each driving wheel. The advantage of this system is that it allows for the expansion of the vehicle's interior space and increase in the battery capacity by integrating the motor into the wheel. However, a drawback of independently driven in-wheel motors is the potential compromise in vehicle safety in the event of one of the motors failing. This paper presents a method to diagnose faults in independently driven in-wheel motors using drive acoustic data and an AI model with a multi-channel microphone array. The system achieved a 90% accuracy in fault diagnosis and localization. It can also be applied to preventive maintenance for future mobility solutions, such as purpose-built vehicles (PBVs).