CORRELATION ANALYSIS OF VIBRATION AND SOUND SIGNALS OF A GASOLINE-ENGINE CAR

Abstract

Nowadays, the number of petroleum-fuel vehicles has been continuously increasing despite
of the arising of electric vehicles. Thanks to the environmental concerns, the EV is being developed in both of
its performance and cost reduction aspects. However the petroleum-fuel vehicles are still popular because of
their more economical sale prize, low-cost maintenance, and wider after-sale services. This paper presents the
correlation between the vibration and sound signals of a personal gasoline engine car, since the important
interfaces of the car users are the perceived vibration and sound signals. In the experimental design, three
vibration sensors and three condensed microphones at different positions have been allocated systematically
according to the most perception of car users. A quantity of the engine vibration and sound signals have been
measured and prepared as collections of signal databases. The signal feature of first pitch is considered due to
its indication of main engine frequency. The fast Fourier transform (FFT) is applied the frequency spectrum of
the signal in all databases. Subsequently, the first dominant peak occurring in the spectrum is selected and
recorded for both of its frequency and its magnitude. These attributes are then analyzed as comparison with
different engine speeds. Finally, the correlations of these attributes between the vibration and sound signals are
presented. From the proposed correlation function, some pairs of the signal features expose significant
relationships, for example, the sound signal at the exhaust pipe and the vibration signal at the seat when
considering the magnitudes at the emerging dominant frequencies.