Development of a Noise Filtering Algorithm for Strain Signals Using the Fast Fourier Transform

The purpose of this study is to filter the noise in strain signals based on the fast Fourier transform. The strain signals were measured at an automotive lower arm made from the SAE 1045 carbon steel driven on paving block and asphalt. This technique removed lower amplitude cycles as much as possible due to their contribution to the minimum fatigue damage and simultaneous maintenance of cumulative fatigue damage. The filtering algorithm was able to remove up to 36.2 % of the lower amplitude cycles and maintain more than 90 % of the fatigue damage. This proved that the algorithm developed was successfully identified and eliminated low amplitude cycles.