Analysis and prediction of automotive running status based on vibration detection system

Abstract

In order to detect and diagnose the operating status without disassembling automobile parts, engine speed and suspension status were tested and analyzed based on vibration detection technology. Based on a DSP processor and an acceleration sensor, the speed of a single cylinder engine was tested under steady-state and transient conditions, respectively. Using the cepstrum method, the vibration signals were collected at different speeds are processed, the fundamental frequency was extracted, and the engine speed information was obtained. According to the speed verification using the Otsu EWT method, the maximum relative error of engine speed was less than 1.5 %, and the error decreases as the speed increases. In order to ensure the reliability of suspension vibration signals, the overall structure and safety of the vibration characteristic detection scheme were designed. Due to the limited sampling frequency of suspension load status sensor data, the EMD method was adopted to process dynamic load signals, which had a good separation effect on suspension vibration signals.