Creep relaxation and fully reversible creep of foam core sandwich composites in seawater

1 Introduction

Ship hulls are composed of a complex array of primary, secondary and tertiary structures that are persistently subjected to a multitude of loading scenarios, such as, fatigue, creep and impact (hydrodynamic wave slamming, tool drop, repeated cargo loading-unloading, docking, etc.) resulting in lifelong accumulation of structural damage. Needless to say, the safe operational life is predicated upon the ability to predict damage, however, it is made inherently challenging due to an inability to detect the onset and propagation of damage until it reaches catastrophic dimensions in particular in layered structures, such as, foam core sandwich composite [1–3] – a situation that becomes far more complex under repeated loading [4, 5] and is further exacerbated in the presence of seawater [6, 7]. The common approach in damage characterization of complex material systems is to isolate damage signatures emanating from distinct modes of loading that can subsequently be implemented into cumulative damage models – such as, the Miner’s rule [3] – however, acquiring and discerning such signals is riddled with uncertainties especially when testing is performed in the seawater, as discussed elsewhere [7].

The cyclic creep reported in the literature is primarily limited to conventional fatigue testing whereby peak loads are held stationary for just long enough to elicit creep response (typically less than half an hour) [8–11]. Furthermore, creep has traditionally been associated with high temperature applications in conventional engineering materials and therefore, is ignored under ambient conditions. The atoms in metals above their recrystallization temperature become quite mobile, therefore, creep of metals is characterized by the fact that most of the deformation is irreversible – which is not the case in polymeric sandwich composites. Polymeric foam cores though certainly vulnerable to temperature hikes are especially susceptible to viscoelastic effects under ambient temperatures that can further accelerate in the presence of seawater due to volumetric expansion caused by water ingress, plasticization and ensuing breakdown of the cell walls [8, 12].

The literature is abundant on the various aspects that influence the lifetime characteristics of the foam core sandwich composites; for example, moisture absorption mechanisms [13–17], effects of moisture on the mechanical properties [18–21], influence of temperature on the mechanical properties [16, 22, 23], fatigue loading [4–6], as well as, the creep behavior of sandwich composites in air [8, 12, 24–27]. However, very little attention has been devoted to the creep of sandwich composites in the ambient marine conditions [28] – which is the actual operational environment of the ships. The current effort is devoted to understanding the damage and lifetime mechanisms under creep to failure and cyclic creep of foam core sandwich composites in seawater which is a unique undertaking as the literature is practically devoid of this highly relevant ship hull service life scenario.

It is common knowledge that cyclic loading can be more detrimental to a material than the quasi-static or sustained loading [4, 5, 8]. Fatigue typically refers to cyclic loading at a rate fast enough that viscoelasticity does not take effect. Viscoelastic effects, on the other hand, generally emanate under long term sustained loading and have been shown to be quite relevant to the service life of ship hulls. For example, ships are commonly loaded for extended periods of time followed by the unloaded phase, which leads to repeated creep (sagging) stress – (stress) relaxation scenario. The pioneering work by Figueroa et al. [28] has shown that cyclic creep-relaxation can lead to severe damage to the material as evidenced in the form of substantially shortened life.

Another scenario relevant to ship hulls is the fully reversible cyclic creep loading which also may come about under routine service operations. For example, ships are designed to carry evenly distributed loads in order to avoid undesirable stresses along the various components of the hull structure. However, it is not always possible to ensure such a scenario especially if partial cargo is loaded and unloaded at various docks along the route. Therefore, if there is more weight concentrated at the ends due to uneven cargo distribution causing excessive buoyancy in the middle, hogging stresses cause tensile stress in the deck area and compressive stress at the bottom in the amidships region. On the other hand, if there is more weight concentrated in the mid length of the vessel due to uneven cargo distribution causing excessive buoyancy at its ends, sagging stresses lead to tension at the bottom of the ship and compression develops in the deck area in the amidships region. The duration of the tension or compression periods in amidships depends on the length of the voyage – thus introducing fully reversible creep effects. This is a critical phenomenon that has never been studied before especially in sandwich composites operating in their natural seawater environment and is the main focus of this research.

2 Experimental setup

Sandwich composite panels of 6.35 mm thick polyurethane foam core with density of 64 kg/m³ and single glass fiber cross weaved facesheet and backsheet of 1.84 mm thickness were fabricated using the vaccum assisted resin transfer molding (VARTM) process. The specimens of dimensions 30 cm×5 cm×8.2 mm were cut out of large rectangular panels. Baseline mechanical testing results yielded average load to failure and stiffness of the sandwich composite to be 611 N and 246 kN/m, respectively. Six specimens were simultaneously tested in a 70×70×180 cm³ tank filled with the actual seawater, as shown in Figure 1.

Figure 1:

Creep testing apparatus.

Creep testing was performed at R=1 (load-relaxation) and at R=-1 (fully reversible loading) for a wide range of loading and unloading and/or reverse loading times. R=1 tests were performed at loading/relaxation times of 24/24, 24/12, 24/6, 12/12, and 6/6 h whereby loads were kept constant for a fixed period of time followed by a predetermined unloaded phase. The fully reversible creep tests were performed for loading/reverse loading times of 36/36, 24/24, 12/12, 6/6, and 3/3 h. This setup was designed to mimic various loading/unloading scenarios of actual ship during service. Creep tests were performed at 50% of the average ultimate static load in order to ensure a substantial creep life, the details of which are elaborated in an earlier publication by the authors [28]. The lifetime and damage during primary (instantaneous response), secondary (slow, long term response) and tertiary (nearing catastrophic failure) phases along with associated damage were monitored visually and recorded for analysis. Three to eight specimens were tested under each loading condition.

3 Results

The results shown in Figure 2 indicate nearly a 50% drop in the overall lifetime and a dramatic shift in compliance in specimens creep tested to failure in seawater when compared with creep life in air. Moreover, in comparison with baseline creep to failure tests in seawater, the results indicate a further 70% lifetime reduction and substantial increase in cyclic compliance in specimens subjected to cyclic R=1 (load relaxation) as well as R=-1 (fully reversible) loading, as also presented in Figure 2. The damage progression (measured primarily as deflection or change in compliance) was intermittent with numerous transitory stalling in the secondary phase that consumed over 95% of the creep life, which agrees with some earlier observations [8, 25].

Figure 2:

Typical cyclic R=1 and R=-1 test results compared with creep to failure tests performed in seawater and air.

For the constant loading and variable relaxation times (R=1 testing), the results point to an increasing deflection and life to failure with decreasing relaxation times, thus suggesting a dependence on the relaxation time, as can be appreciated from Figure 3A and Table 1. Load relaxation tests performed at equal loading and relaxation times (Figure 3B and Table 1) however, failed to provide a consistent trend as the life was seen to be the shortest for 24/24 h tests while 12/12 h loading produced the longest life and 6/6 h providing intermediate lifespan. In other words, if the loading time was kept constant (in R=1 tests), the relaxation time dominated; on the contrary, the loading cycles appear to affect the service life behavior (Table 1); which would suggest an optimization of stress/relaxation times for optimal performance.

Figure 3:

Typical R=1 test results showing the dependence on the relaxation times.

Table 1

Analysis of cyclic creep.

Under fully reversible loading (R=-1), Figures 4 and 5 and Table 1 indicated a decreasing lifetime and correspondingly increasing cyclic compliances as the loading cycles shortened. Hence, suggesting a dependence on the loading/unloading periods and/or number of reversals that is consistent with expected behavior under fatigue loading condition. As compared with R=1 tests, the fully reversible tests revealed a substantial reduction followed by a gradual increase in lifetime with increasing loading periods (Figures 6 and 7, and Table 1). It is well known that a tension/compression (R=-1) scenario is far more detrimental to a material than a tension or compression (R=1) alone, thus the results are consistent with the expected behavior, see e.g. Sharma et al. [4]. The outcomes outlined in this section are quite significant being most relevant to the ship hull service lifetime scenario especially when the literature on this topic is almost non-existent.

Figure 4:

Typical R=-1 test results showing the dependence on the number of reversals.

Figure 5:

R=-1 creep test results. Average deflection to failure similar in all cases at ~23 mm.

Figure 6:

A comparison of R=1 at fixed loading and various unloading times with R=-1 test results at equal loading and reverse loading times.

Figure 7:

A comparison of R=1 and R=-1 creep testing results at equal loading and unloading/reverse loading times.

The modes of failure in all cases were local buckling and indentation in the vicinity of the loading site (as shown in Figure 8). There is a substantial lack of discussion in the literature regarding modes of failure in sandwich composites under creep, however, failure modes under quasi-static loading are well established and appear to support the current observations [1, 3, 29–31].

Figure 8:

Modes of failure.

4 Discussion

5 Conclusions

Foam core sandwich composites were subjected to cyclic load relaxation and fully reversible creep for a broad range of loading, relaxation and reverse loading periods. Load relaxation tests suggest the lifetime dependence on the relaxation times for constant loading periods and variable relaxation periods; whereas, the dependence appears to be upon the number of loading cycles under equal loading/relaxation periods. For the fully reversible tests, the dependence of lifetime appears to be on the number of reversals, i.e. increasing number of reversals lead to shortening of cyclic creep lifetime. Furthermore, in comparison with creep to failure testing, lifetime was considerably reduced for both load relaxation and fully reversible cyclic creep.