Vibration Data Classification with Fast Fourier Transform and Neural Network

  • Sri Hartati Universitas Bhayangkara Surabaya
  • Richa Watiasih Universitas Bhayangkara Surabaya
  • Prihastono . Universitas Bhayangkara Surabaya
Keywords: Vibration, FFT, NN, Classification

Abstract

Implementation of vibration data classification method with Neural Network (NN) for electrical compressor mechanical looseness condition was proposed in this paper. With used NN was successfully trained for vibration data classification for three conditions that detected: normally condition, mechanical looseness conditions: screws buffer induction motor compressor were looseness and baseplate was weakness). The input matrix of vibration data from detection with two vibration detector were processed with Fast Fourier Transform (FFT). Vibration data from first vibration detector (LVDT1) 98% was successfully classification and for vibration data from second vibration detector (LVDT2) 99% was successfully classification.
 

Published
2017-08-29
Section
Articles