Online Damage Monitoring of Joint Structures with Gaussian Mixture Model Under Load Conditions

In the damage detection of a joint structure based on guided waves, the diagnosis of invisible damage is a challenge. The complex multi-nail structure and the influence of complex load environment are the main factors that affect the reliability damage monitoring of the joint structure. Therefore, it is essential to effectively improve the reliability of damage monitoring of joints under load conditions. In this paper, an online damage monitoring method based on the Gaussian mixture model is proposed. In the monitoring process, the Gaussian mixture model modeling method was used to model the characteristics of guided wave signals at different times under the influence of load factors. Then, the probability difference damage index Diff was used to represent the change in the probability distribution between the models to achieve reliable damage monitoring. Fatigue test results of joint structures show that this method can realize reliable damage detection of joint structures under load environments.