Static and dynamic measurement of low-level strains with carbon fibers

A signal processing method used to treat the structural strain responses of long-gage carbon fiber (CF) sensors for static and dynamic strain measurements is developed in this paper. This method includes a high-confidence denoising method for continuous static monitoring and a wavelet transform (WT)-based signal processing method for dynamic strain measurements. The static denoising method is based on the measuring error range of CF sensors which defined by a long-term continuous loading and unloading experiments. For the experiment involving 1-measure-time at each strained point, the influence of noise on CF sensors follows a normal distribution with a standard deviation of 50 microstrain. For the experiment involving 25-measure-time at each strained point, the signals are concentrated in a smaller range, and the probability density clearly increases. For a high-confidence signal selection based on multiple-measure-time, the range of error measurement of CF sensors is concentrated from 50 microstrain to 10 microstrain. Furthermore, a study on the signal processing of the dynamic strain signal of CF sensors through WT-based signal processing method is presented. Due to the influence of the dynamic measurement of small strains, the baseline of the dynamic signal of CF sensors drifts. With the WT denoising method, the discrete degree of the dynamic signals was reduced. It is beneficial to find the peak position of each wave in the vibration response of CF sensors. With the peak positions, the baseline drift can be removed by a mathematical regression. Finally, the experimental results are presented to show the effectiveness of the processed signal of the CF sensors studied herein.