Determination of the influence of thermo-mechanical factors on the residual stresses of cylindrical composite tubes: Experimental and computational analyses

doi:10.1016/j.ijpvp.2020.104098

International Journal of Pressure Vessels and Piping

Volume 183, June 2020, 104098

https://doi.org/10.1016/j.ijpvp.2020.104098 Get rights and content

Highlights

•
Process-Induced residual stresses are experimentally measured accomplishing semi-destructive ‘slitting’ method.
•
Finite element (FE) analysis is performed obtaining compliance coefficients.
•
The influence of various thermo-mechanical factors on the residual stress response of hybrid FRP tubes is characterized and discussed in detail.
•
Outcomes of previous research by authors are compared and verified.
•
Thermal residual stresses decrease dramatically as discussed parameters are optimized.

Abstract

The well-established benefits of compressive residual stresses over the fatigue lifetime constantly motivate engineers to scrutinize these trapped stresses in mechanical/civil structures. This research deals with characterization of thermal residual stresses in fiber reinforced polymer (FRP) tubes. Three important thermo-mechanical parameters, namely “cooling temperature”, “CNT reinforcement” and “tube size” are taken into account with the aim of achieving optimum design in manufacturing of FRP tubes. The work includes experimental evaluation of residual stresses coupled with developing a MATLAB programming code and performing a thorough finite element (FE) analysis. After fabricating several hybrid FRP tubes with respect to the mentioned parameters, we measured residual stresses conducting “slitting” semi-destructive method. The impact of either of parameters is then examined and some key points are suggested regarding controlling residual stress in FRP structures. The outcomes of this work suggest that slitting method can practically provide access to the in-depth distribution and through thickness of trapped stresses in the FRP laminates with a good accuracy. It is also underscored that studied parameters could play a crucial role in creation of thermal residual stresses.

Introduction

Mechanical performance of engineering structures can be significantly influenced by residual stresses. Warpage and matrix microcracking are prime examples regarding destructive effects caused by these stresses in composite structures, which can lead to decrease in structural strength [1,2]. Thermal residual stresses (TRS) are created due to a mismatch in the coefficients of thermal expansion (CTE) between matrix and fiber [3,4]. Residual stresses can also be created in welding process of tubular structures and pipe lines [[5], [6], [7], [8]]. Repair welding is a type of modification process that may affect the performance and health quality of the whole structure by formation of the undesirable residual stress and distortions in the repaired area [[9], [10], [11], [12]].

Tsai and Chi [13] investigated TRS effects on the mechanical performance of fiber reinforced composites with different fiber arrays. Fibers were assumed to be linear elastic; meanwhile, matrix was a nonlinear material. Numerical results indicated that for composites with square edge packing, geometrical and mechanical behaviors are appreciably affected by residual stresses. Sglavo et al. [14] presented a procedure for measurement of residual stress profile in regular geometry bodies such as plates and disks based on the sole measurement of sample curvature deriving from progressively etching one of its surfaces. Then, a relationship was obtained to correlate curvature data as a function of etching depth to the original residual stress profiles.

Of all practical methods regarding assessing residual stresses, hole drilling [[15], [16], [17]] and slitting [18] are two common ones for evaluating TRS in laminated composite structures. Ghasemi et al. [15,19] employed integral hole drilling method to calculate non-uniform residual stresses trapped within various composite laminates. They also proposed a simulated hole drilling method which could be used instead of experimental techniques for calibration factor calculations. In other research by Ghasemi and Mohammadi [20] residual stresses of fiber metal laminates (FMLs) were measured using incremental hole drilling (IHD) technique. Experimental and theoretical results were compared and maximum difference between them in the first step of drilling was seen to be 3.1%. Mahmoudi and Hasani [21] proposed a modification to the cross slitting method to enhance accuracy of the technique. The proposed modification was examined both numerically and experimentally. The modified cross slitting (MCS) proved to provide better results than the cross-slitting method (CSM). However, ability of measuring two components of residual stresses was restored.

A vast number of studies have been conducted on determination of residual stresses in flat structures so far. When it comes to curved structures, however, it is seen that limited analyses are available and investigating these structures is of great importance. Akbari et al. [22,23] presented experimental investigation of residual stresses in filament wound composite rings. They applied both hole drilling and slitting methods and concluded that slitting method has the potential for measurement of residual stress in components manufactured by methods, including layer deposition, such as laser cladding and filament winding. Results showed that there is a reasonable agreement between two methods.

The residual stress quality and structural integrity of composite profiles can be affected by a number of parameters associated with design and manufacturing stages such as curing and cooling conditions, materials, lay-up arrangement and geometry of the structures [[24], [25], [26], [27], [28]]. In many cases, the influence of mentioned parameters cannot be taken into account in theoretical analyses due to simplifying calculations. Therefore, an experimental characterization is evidently needed. In this research, non-uniform residual stresses of hybrid FRP composite tubes being manufactured by filament winding are determined using slitting method. ANSYS commercial software is used to simulate slitting process and compliance coefficients are determined for each layer with consideration of such influential factors as material, fiber orientation and position of different layers. Next to this, residual strains in each layer are precisely measured in every single step of slitting process. These are then correlated to residual stresses using compliance coefficients. These trapped stresses are non-uniform and vary along the thickness. This study aims to analyze the effects of thermo-mechanical parameters such as: cooling temperature, addition of multi-walled carbon nanotubes (MWCNTs) as reinforcement material and diameter of the cylinder on the TRS of laminated composite tubes. The optimum design of FRP tube in terms of mentioned parameters is discussed and ultimately reported in conclusions.

Section snippets

Materials and geometrical features

In present study, fiber reinforced polymer (FRP) composite tubes have been produced using carbon fiber T300 [22] and ML 506 epoxy resin [29]. The filament winding method was applied to fabricate the samples in which the fibers were impregnated with resin. Then, the impregnated fibers were wound around the mandrel. The specimens were cured for 4 h at 100 °C. After curing, two different cooling temperatures were considered such that some cylinders were cooled at room temperature (25 °C); while,

Experimental procedure

In this study, slitting method is applied to determine the TRS of carbon fiber reinforced polymer composite cylinders. In this method, a strain gauge is mounted on the sample and a narrow slit is cut along the thickness incrementally. In each step of slitting process, released strains are measured and then, strains are correlated to residual stresses with application of calibration coefficients. According to the available apparatuses, there are a number of methods to cut a slit. In present

Theoretical background of slitting method

The unknown residual stresses in slitting experiment are correlated to measured strains as [23]: $ε_{y y} (a_{i}) = \int_{0}^{a_{i}} G (x, a_{i}) σ_{y y} (x) dx$ in which $ε_{y y} (a_{i})$ and $σ_{y y} (x)$ are measured strains and unknown residual stresses, respectively. Also, $a_{i}$ is the slit depth and $G (x, a_{i})$ is called Kernel function and should be determined by finite element modeling which is presented in next section. To solve Eq. (1), a proper distribution of residual stress along the thickness should be considered. There are several applicable

Finite element modeling

The compliance coefficients in slitting method are determined using ANSYS commercial software. According to lay-up arrangement, each uni-directional layer is considered and defined as a separate material in ANSYS. It should be noted that the slitting section of each layer has effects on the other layers. Mechanical and thermal properties of a uni-directional ply are presented in Table 1:

A 3-D model is simulated for the laminated composite tube using eight-node solid185 elements. Also, the

Results and discussion

In this section, results are discussed. The results are represented in three distinct parts. In part one, analyses are presented in terms of “cooling temperature”. In part two, results are presented for “diameter size of the tube”. Then, optimal results of these two parts are taken into account and, ultimately, the effects of addition of MWCNTs are studied in part three. It should be mentioned that all specimens in sections 6.1 Cooling temperature, 6.2 Diameter size of the tube have been

Conclusions

The motivation of present paper was to characterize the effects of thermo-mechanical factors on the residual stress response of FRP composite tubes. Slitting method was conducted to assess the non-uniform residual stresses over the thickness of composite tube. Results revealed that all studied specimens underwent tensile residual stresses in outer layers; while, the inner layers experienced compressive thermal residual stresses. This phenomenon is in line with bending behavior which consists of

Author statement

1. Ahmad Reza Ghasemi: Ahmadreza is Associate Professor of University of Kashan. The role of Ahmad Reza is the corresponding author and advisor of the thesis that manage and coordinate the research activity. He is a developer and designer of the methodology and prepared the samples and equipment of the test. 2. Behzad Asghari: Behzad is M. Sc student of University of Kashan. The role of Behzad is a developer and designer of the methodology and creation of models and performance of the test.3.