The usefulness of wavelet transformation to reduce noise on vibrations of cable structures

Abstract

Vibration testing applications have been used to examine a structure's dynamic behavior, such as determining the frequency values in cable structures. The accelerometer is used to record cable vibration data. Cable vibration data that has been mixed with noise has a non-periodic signal. In analyzing non-periodic signals using Fast Fourier Transforms analysis, there are obstacles in determining the reading of the frequency value. The study proposes a Discrete Wavelet Transform (DWT) to overcome the existing obstacles in analyzing the non-periodic signal. It can minimize the noise in the cable vibration data recording, making it easier to determine the frequency value of the cable structure, especially at the first vibration mode. In minimizing noises, the use of a scale factor of = 0.1 becomes the most effective value with the highest Signal Noise Ratio (SNR) value and the smallest Root Mean Square Error (RMSE). Other results obtained are Signal Noise Ratio (SNR) in the range of 2 - 5 dB, the type of noise in the cable structure is white noise and, the ratio of the standard deviation of noise (s) to the amplitude (A) of recorded cable structure data with a range of 1 - 3.5.