Prażnowski, K. and Mamala, J., "Problems in Assessing Pneumatic Wheel Unbalance of a Passenger Car Determined with Test Road in Normal Conditions," SAE Technical Paper 2017-01-1805, 2017.
The vibrations of the sprung mass of a passenger car, traveling along a road surface, are random. They also form its main source but there are besides other factors to consider. The resulting force ratio is overlapped by other phenomena occurring at the interface of the pneumatic tire with the road surface, such as non-uniformity of tires, shape deformations and imbalances. The resulting additional inertia force acts on the kinematic force that was previously induced on the car body. The vibrations of the sprung mass of the car body at the time can be considered as a potential source of diagnostic information, but getting insight their direct identification is difficult. Moreover, the basic identification is complicated because of the forces induced due to the random interference from road roughness. In such a case, the ratio defined as SNR assumes negative values. Due to the lack of comprehensive studies involving identification of the unbalance of the whole pneumatic wheel in real conditions, these authors decided to undertake experimental research in this area. The main objective of the study was to identify the unbalance of a pneumatic wheel based on the vibration of the sprung mass in a car body. The registered acceleration sensor signal was analyzed in the frequency domain, thus obtaining information about its specific components such as the value of the amplitude and frequency range prevailing in its spectrum. The analysis of the recorded signals applied selected statistical methods (autocorrelation, probability density distribution), as well as the frequency domain (PSD, FFT, STFT). To this purpose, this paper presents the results of both bench and road implemented at selected speeds ranging from 50 to 100 km / h. The results of the study were compared with the results of simulation calculations using the model of quarter car. The application process was based on a one-dimensional Bayes classifier and a designated distribution function for the strengthening of a classifier representing the state of the road surface.