Prediction of Vibration Characteristics of Braking System Based On Wavelet Packet Algorithm and Artificial Neural Network

  • Menwer F.M. Alenezi, Abdulhadi O. A. Alajmi

Abstract

This research study presents a novel method to classifying vibration signals generation on the braking system under different conditions. The concept of wavelet transform for feature extraction from the vibration signal combined with an artificial neural network to offer a powerful tool for classifying vibration signal of braking system is introduced. Vibration signals for disc brake system are de-noised using wavelet transform using db2 as mother wavelet and the same features extracted from raw signal and de-noised signal for four operating conditions namely; applied pressure on brake, adding water, using iron chips and adding sand particles into braking surface at two speeds and different loads. The results confirmed the important of the feasibility of the proposed algorithm. This method proposed herein classified and obtained 99.96% classification accuracy. It is also found that de-noising has its benefits on the optimization and performance of ANN classifier. It can be reported that the vibration signal with wavelet de-noising and ANN classifier could be used to determine the presence of strange objects affecting braking performance considerably.

Published
2020-01-28
How to Cite
Abdulhadi O. A. Alajmi, M. F. A. (2020). Prediction of Vibration Characteristics of Braking System Based On Wavelet Packet Algorithm and Artificial Neural Network. International Journal of Advanced Science and Technology, 29(3), 195 - 206. Retrieved from http://sersc.org/journals/index.php/IJAST/article/view/3838
Section
Articles