PSI - Issue 72

V.P. Matveenko et al. / Procedia Structural Integrity 72 (2025) 229–234

233

1

1 M  

,

(4)

y

x



, m n

, m j n 

M

0

j



where y is the output signal; x is the input signal; M is the number of averaged samples. This version of the relationship is presented for a two-dimensional signal with averaging over one of the dimensions. Fig. 3 shows results representation similar to Fig. 2 for filtered data using a moving average filter on the spatial coordinate over 10 samples.

Fig. 3. Visualization of noise in the readings of distributed fiber-optic sensors based on Rayleigh scattering after applying a moving average filter over ten measurements.

The application of th e moving average filter reduced the maximum strain range to 4.5 µε along the spatial coordinate and 6 µε along the time coordinate. Additionally, the average standard deviation decreased to 0.7 µε for the spatial coordinate and 0.85 µε for the time coordin ate. 4. Conclusions This study conducted a statistical analysis of noise in strain measurements by distributed fiber-optic sensors based on Rayleigh scattering, emphasizing the necessity of understanding measurement uncertainty in strain sensing applications, which is crucial for providing measurements with high precision. The investigation focused on characterizing the inherent noise levels in the absence of external mechanical loads and temperature fluctuations on the studied section of the optical fiber. The results indicated that the maximum observed strain amplitude range was 11.4 µε, with a corresponding standard deviation of 2.1 µε along the fiber length. Similarly, strain variations over time exhibited a maximum amplitude range of 17.6 µε, with a standard deviation of 2.6 µε. These findings highlight the effect of noise on DFOS measurements and the importance of applying filtering techniques or other methods to enhance signal quality in applications where high precision measurements are demanded.