PSI - Issue 52

Yingwu Li et al. / Procedia Structural Integrity 52 (2024) 709–718

717

Author name / Structural Integrity Procedia 00 (2023) 000–000

9

Fig. 9. The type A evaluation of standard uncertainty of all test scenarios mentioned in Section 2.1 and Section 3.1

4. Measurement uncertainty of distributed fibre optic sensors with OFDR system

According to the type A evaluation of standard uncertainty, as defined in the ”Guide to the Expression of Uncer tainty in Measurement” (GUM) Iso and OIML (1995) published by the Joint Committee for Guides in Metrology (JCGM), the uncertainty of measurement can be mathematically expressed in Equation 2. s (¯ q ) = n i = 1 ( q i − ¯ q ) 2 n ( n − 1) (2) In this equation, the type A evaluation of standard uncertainty is determined based on the mean value ( q ) of the measurement data, the quantity of data points ( n ), and the individual measurement points ( q i ). Utilizing this definition, the uncertainty degree of measurements in all test scenarios mentioned in Section 2.1 and Section 3.1 is calculated and visualized in Figure 9. The x-axis of the histogram represents di ff erent test scenarios, with labels mentioned in Section 2.1 and Section 3.1(same as Figure 5, Table 1, and Table 3), comprising a total of fifty-six experiments (corresponding to fifty-six bars). The y-axis displays the values of the type A evaluation of standard uncertainty. The results in this figure are obtained from 200 measurement points in each test scenario. According to the findings depicted in Figure 9, the type A evaluation of standard uncertainty in various test scenarios is consistently less than 0.1 GHz ( n = 200). This suggests that the uncertainty degree of measurement is low and remains stable when distributed fiber optic sensors are employed to obtain strain or temperature data in composite structures. This research focused on assessing the measurement consistency of distributed fiber optic sensing under various test scenarios. The measurement data from fifty-six independent test scenarios were categorized into six groups, in cluding tensile experiments (thirteen tests), fatigue experiments (twelve tests), three-point bending experiments (nine tests), and three groups of temperature experiments (twenty-two tests). The strain-frequency shift coe ffi cient and temperature-frequency shift coe ffi cient were represented using Cumming plots. The normality of these coe ffi cients was determined using the Kolmogorov-Smirnov (K-S) test, and their consistency was further assessed through Cron bach’s alpha, McDonald’s omega, and Split-half reliability based on the normal distribution conclusion. In addition, the standard uncertainty (type A) of frequency shift measurement in these fifty-six independent test scenarios are calculated and presented in this research. The main findings of this study are as follows: 5. Conclusions