PSI - Issue 52

Yingwu Li et al. / Procedia Structural Integrity 52 (2024) 709–718 / Structural Integrity Procedia 00 (2023) 000–000

713 5

Author name

Fig. 4. The Cumming plot of strain-frequency shift coe ffi cients in three-point bending experiment

(a)

(b)

Fig. 5. The K-S test results for the di ff erent test scenarios in both strain and temperature measurement

test scenarios are displayed in Figure 5(a). In this histogram, the x-axis representing various test scenarios, while the y-axis shows the p-value of the K-S test (ranging from 0 to 1). The p-value means the probability of observing a test statistic as extreme as, or more extreme than, the observed value under the hypothesis of normal distribution. The p values for test scenarios in the tensile experiment are indicated by blue bars, for the fatigue experiment by green bars, and for the three-point bending experiment by red bars. The red dashed line represents the significance level of the hypothesis test, conventionally set at 0.05. Based on the figure, it can be inferred that the measured strain-frequency shift coe ffi cients under most test scenarios conform to the normal distribution. The measurement errors in di ff erent test scenarios are clearly independent, and all the measurements in these tests exhibit the same relationship with the underlying construct, as indicated by Equation 1 (tau-equivalent). Leveraging the normal distribution conclusion obtained from the K-S test, various coe ffi cients, such as Cronbach’s alpha ( α ), McDonald’s omega ( ω ), and Split-half reliability coe ffi cient, can be employed to assess the internal consistency of