**1. Introduction**

Over the past decades, carbon-fiber-reinforced polymer (CFRP) composites have shown a constant increase in a variety of applications such as car or aircraft components, sports and medical equipment [1], and recently also additive manufacturing [2–4]. Their main benefit lies primarily in the relatively high strength/weight ratio or the ability to customize the material properties for dedicated purposes by changing the stacking sequence and related fiber orientation. A relative drawback of CRFP composites is the lack of ductile-like behavior and the corresponding absence of pre-warning phase before the structural collapse [5,6] leading to the brittle failure. CRFP composites are also characterized by the accumulation of damage inside the structure without any evidence on the structure surface [6] thus leading to a relatively challenging damage assessment. There are many non-destructive testing approaches, which can be applied on composite structures, namely infrared tomography [7], eddy current testing [8], ultrasonic testing [9], and X-ray tomography [10].

One of the most promising approaches, especially coupled with other methods [11] such as Scanning Electron Microscopy (SEM), is the acoustic emission (AE) method, which is also used in various applications as a real time monitoring tool [6]. The AE method exhibits great sensitivity including considerable reliability of active cracks detection [12], even in the case of initiation phase [13]. The AE technique is even capable of detecting the onset of plastic deformation [14], which has the character of white noise with low energy [15]. For gaining a more detailed insight into the damage monitoring process within

**Citation:** Šofer, M.; Cienciala, J.; Fusek, M.; Pavlíˇcek, P.; Moravec, R. Damage Analysis of Composite CFRP Tubes Using Acoustic Emission Monitoring and Pattern Recognition Approach. *Materials* **2021**, *14*, 786. https://doi.org/10.3390/ma14040786

Academic Editor: Michele Bacciocchi Received: 12 November 2020 Accepted: 28 January 2021 Published: 7 February 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

the meaning of AE source characterization, it is favorable to incorporate an adequate signal analysis tool. The supervised/unsupervised pattern recognition (UPR) approach [16] has become a very suitable and promising approach to tackle a wide variety of problems such as fatigue tests [17], structure health monitoring [18], and condition assessment of pressure vessels [19] and pressure components in operation [20]. Numerous studies [21–26] have been conducted in order to assess characteristic features of the AE transients originating from various failure mechanism in the CFRP composites such as matrix cracking, delamination, fiber break, and debonding (see Table 1 for further explanation).

**Table 1.** Basic characterization of damage mechanisms occurring in carbon-fiber-reinforced polymer (CFRP) composites.


Although Chou [21] points to a discrepancy concerning, in particular, the signal amplitude, duration as well as frequency spectra of the individual damage mechanisms, it was possible to compile a general overview, which is given in the following table (Table 2).

**Table 2.** Summary of the acoustic emission (AE) signal characteristics for given damage mechanisms in CFRP composites


\* A—amplitude, D—duration.

In the last decade, several/numerous studies utilizing advanced techniques for classification of failure modes, such as the use of statistical analysis of wavelet coefficients [27] or infrared thermography (IT) [28], have been conducted. Another interesting approach can be found in the work published by Munoz et al. [29], who identified and further characterized the damage mechanisms in the unidirectional CFRP composites subjected to axis and off-axis static tensile tests using the acoustic emission method and infrared thermography. Further utilization of unsupervised pattern recognition technique together with the IT method resulted in the identification of the failure mechanisms such as matrix cracking, fiber breakage, and interface failure, for which the characterization in terms of the signal amplitude or energy has been performed. In 2011, Gutkin et al. published an extensive research [30], in which the AE signal data from various test configurations were analyzed by three different pattern recognition approaches. The analysis resulted in characteristic frequency spectra for matrix cracking, delamination, debonding, fiber pull-out, and fiber failure. It has to be noted that the given findings in terms of the frequency spectra are to some extent similar to the results summarized in Table 2 and therefore confirms the factual accuracy of the study [30].

The main objective of this study is to investigate and comprehensively describe the AE signal characteristics of the damage mechanisms in three different types of CFRP composite tubes using a two-step method combining the unsupervised pattern recognition approach with the utilization of the boundary curve. The construction of the boundary curve has been conducted on the data from the preliminary carbon fiber sheaf tensile tests. The already identified boundary curve has then been used for further refinement of the data across individual clusters. Using the presented approach, it was possible to identify a total of four damage mechanisms presented in Table 1 with subsequent comparison of the obtained results with the already published research papers. The part of the study is also the comprehensive AE waveform analysis of the representative signals belonging to the individual clusters.

### **2. Experimental Procedure**

#### *2.1. Test Sample Characterization*

The experiments were carried out on three types of CFRP tubes with a different number of layers, their orientation, and woven fiber density of the used material (see Figure 1), where each type of CFRP tube has been represented by three test samples. The samples labeled "A" were manufactured using four layers of unidirectional carbon woven fabric with density of 200 g/m2 and one layer of aramid/carbon woven fabric (0–90◦) with density of 175 g/m<sup>2</sup> and average wall thickness of 1.45 mm. The production of samples labeled "B" included the use of two layers of unidirectional carbon woven fabric with density 300 g/m2 and one layer of carbon woven fabric (0–90◦) with density 280 g/m<sup>2</sup> with average wall thickness of 0.9 mm, while samples labeled "C" were manufactured using solely four layers of unidirectional carbon woven fabric with density of 300 g/m<sup>2</sup> with average wall thickness of 1.42 mm. Table 3 summarizes the specification of the tested CFRP tubes.

**Figure 1.** CFRP composite tubes under test; A—A series sample, B—B series sample, C—C series sample.

**Table 3.** Specification of the tested CFRP tubes

