Acoustic emission analysis of adhesive joints

Andreas Wimmer, summer semester 2021

Acoustic Emission Testing (AET) can be used to detect and roughly localize damages in components. Particularly in the case of adhesive joints, only limited statements can be made about the damage mechanisms that occur except looking at the fracture surfaces. Therefore, the AET makes it possible to investigate the object during the damaging process. [1,5]

Acoustic emission testing (AET) in general

Function

Basically, AET is a non-destructive testing method that is suitable for detecting active damage by passively absorbing the resulting sound waves. If a component or an adhesive bond is loaded beyond the elastic range, permanent deformation or damage of the object occurs. During this process, a specific acoustic signal is emitted depending on the respective component and the failure mechanism occurring. For this purpose, sound sensors are attached to the surface. Depending on the type of examination, these can either detect the number of acoustic events or use several sensors in an array to localize the position and store it together with the frequency and absorbed energy. [1]

Usage

AET can be used either in the laboratory for individual tests on prepared test specimens or in machines and structures as a continuous monitoring tool. For example, structures made of concrete and steel can be monitored with this method and statements can be made about the remaining service life. Especially in the case of bonded joints, there is no monitoring method apart from AET that can detect the dynamic damage process inside the structure. Thus, early damage can be detected and macroscopic defects can be predicted at an early stage. In the following, it is shown how AET is performed on specimens, which damage mechanisms result from it and what follows from it for the application on real components. [1,4]

Specimens

To have defined test objects special test specimens are used in the laboratory environment. However, the geometry definition is only fixed within a test and differs depending on the performer and the materials used. Basically, the specimens are used to investigate the basic material behavior and the acoustic emissions that occur in the process. In the experiments, the basic feasibility of AET for bonded joints is to be demonstrated. These tests will then form the basis for using AET for large components at a later stage. This chapter deals with both the geometry and the materials on which AET can be carried out for bonded joints.

Design of the specimens

Basically, each specimen consists of at least two different materials, the adherends and the adhesive. The geometry of the specimens has a significant influence on the deformation behavior during the tests. Single layer joints (SLJ) and double layer joints (DLJ) are widely used, for which there exists also standardized geometry specifications. The schematic figures are shown in Figure 1 (SLJ) and Figure 2 (DJL). The testing machine can be a standard machine for tensile tests which records the load and crosshead displacement. The AET sensor(s) is/are mounted at the specimen’s surface to ensure in-situ monitoring. Furthermore, additional equipment can be used, which supplements the data of the AET with additional information, which can be useful for the evaluation. For example, this can be a 3D Digital Image Correlation or a Fiber Optic Sensor. [2,4]

The geometrically determined disadvantage of SLJ is that the distribution of the shear stress in the adhesive is not constant (Figure 3). This leads to strong deformations at the transitions to the joined parts due to the increase in stress. Furthermore, the crack nuclei are located there, which then lead to further failure. Nevertheless, this form of the specimen better represents the joints that occur in reality. [5]

Material

The fundamental advantage of bonded joints is that different materials can be joined together. In general, joints made of steel and carbon fiber reinforced polymers (CFRP) are of great interest in the context of AET. However, adhesive joints that include other metals or wood as a joining partner can also be studied using AET. The adhesive used can also vary greatly in its properties. In addition to properties relating to adhesion and cohesion, the elasticity of the adhesive is also very important, as this has a major influence on sound emission. A methacrylate-based adhesive is an example of a flexible adhesive, epoxy-based is an example of a brittle adhesive. [2]

AE-Measurements

The fundamental advantage of bonded joints is that different materials can be joined together. In general, joints made of steel and carbon fiber reinforced polymers (CFRP) are of great interest in the context of AET. However, adhesive joints that include other metals or wood as a joining partner can also be studied using AET. The adhesive used can also vary greatly in its properties. In addition to properties relating to adhesion and cohesion, the elasticity of the adhesive is also very important, as this has a major influence on sound emission. A methacrylate-based adhesive is an example of a flexible adhesive, epoxy-based is an example of a brittle adhesive. [2]

Types of breakage

As mentioned, the type of adhesive has a significant influence on the fracture behavior and the emitted sound waves. Depending on the classification, the fracture types can be summarized in five groups. There are:

• Cohesive failure (failure within the adhesive layer)
• Steel deformation
• Adhesive failure (failure at the adherends/adhesive interface)
• Delamination in the CFRP skin
• Fibre breakage

The following figures are intended to illustrate these relationships and to address the influence of adhesive elasticity. As shown in Figure 4, delamination and adhesive failure is typical for brittle adhesives. The corresponding frequency curve is plotted in Figure 5. The comparatively low AE activities indicate that there is an abrupt failure of the bonded joint with little deformation. [2]

Figure 6, on the other hand, shows the fracture behavior for an elastic adhesive. The failure modes here have predominantly cohesive causes. As can be seen in Figure 7, the AE activity is significantly higher and the deformation before failure is significantly larger at around 11 mm. [2]

Automated determination of the failure mode

As shown in Figure 8, the different failure modes have different specific curves for the peak amplitude over time. Special attention should be paid to the vertical axis. The amplitude values are plotted in different value ranges depending on the type of failure. The desire is to evaluate the failure modes that have occurred in such a way that the failure mechanism can be determined without looking at the fracture surface. This has the purpose of being able to determine the type and size of damage in the case of partial damage. For this purpose, it is necessary to divide the failure modes into clusters with specific properties. [3]

To obtain information on the individual components of the bonded joints, tensile tests are performed on each material and the AE data is recorded. Thus, a material-specific displacement/load/peak frequency curve is gained. This is then used to train a decision tree classifier. This can assign the information of the individual materials when performing a DLJ test in such a way that the failure mode of the overall composite can be determined. After appropriate training, the failure mode can be determined with an agreement of over 99 %. [2]

Kaiser effect

The Kaiser effect theory assumes that damage only emits further acoustic signals as soon as the load is higher than the load causing it. If the load decreases in the meantime and then returns to the previous value, no acoustic emission is to be expected. This theory can also be observed in the AET of bonded joints. However, the expression varies depending on the adhesive and the load. This results in the condition for AET to be used as a continuous monitoring method, since otherwise it cannot be ruled out that existing damage is not detected due to the Kaiser effect. [4]

Conclusion

AET is a good supplement to already established non-destructive testing methods for quality assurance and long-term monitoring of bonded joints. By continuously recording data, damage to the component can thus be detected at an early stage and countermeasures can be initiated. With the help of modern techniques such as machine learning, the detection of failure modes can be optimized further and further and the significance of the AET can be continuously increased. [3,4,5]

Literature

1. Grosse, C. (Hrsg.): Grundlagen der zerstörungsfreien Prüfung. Lehrstuhl für zerstörungsfreie Prüfung, TUM - Technische Universität München, München (2020), p. 182.

2. Saeedifar, M., Saleh, M. N., De Freitas, S. T., Zarouchas, D.: Damage characterization of adhesively-bonded Bi-material joints using acoustic emission, in Composites Part B 176 (2019) 107356, p. 1 - 21 https://www.sciencedirect.com/science/article/abs/pii/S1359836819322061

3. Xu, D., Liu, P. F., chen, Z. P., Leng, J.X., Jiao, L.: Achieving robust damage mode identification of adhesive composite joints for wind turbine blade using acousic emission and machine learning, in Composite Sturctures 236 (2020) 111840, p. 1 - 17 https://www.sciencedirect.com/science/article/abs/pii/S0263822319334245 
4. Hill, R.: The use of acoustic emission for characterising adhesive joint failure in NDT International, Volume 10, Issue 2, (April 1977)) p. 63 – 72
5. Crostack, H.-A., Hahn, O., Kötting, G., Nolte, F.: Untersuchungen zur Ermittlung und Bewertung von Schädigungsvorgängen in Klebeverbindungen mittels Schallemissionsanalyse, Verlag Chemie GmbH, Weinheim (1984), Z. Werkstofftech. 15, p. 288 - 297
6. Grabmann, S.: Fügetechnik Übung – Kleben, Technische Universität München (2021), p. 15