PSI - Issue 38

Frédéric Kihm et al. / Procedia Structural Integrity 38 (2022) 12–29

25

14

Kihm, Miu, Bonato / Structural Integrity Procedia 00 (2021) 000 – 000

4.5. Linear Models

We begin by investigating a linear model fit between strain damage and Rear_axle_inclination damage. This is shown in Fig. 7.

Fig. 7. comparison of damage from Rear_axle_inclination versus strain-induced damage

We can observe that the linear model interpolates the data well throughout the entire range. As it appears in Fig. 8, the distribution of residuals appears to resemble a normal distribution. The relative absolute error of the predictions on the training set is 39.8%.

Fig. 8. distribution of residuals

UNRESTRICTED