A Testing Platform for Durability Studies of Polymers and Fiber-reinforced Polymer Composites under Concurrent Hygrothermo-mechanical Stimuli

Polymer and fiber-reinforced polymer (FRP) composites have been adopted in a variety of engineering structures, ranging from aircraft and spacecraft, naval vessels, civil infrastructure, (see for examples reviews of Katnam et al.¹, Hollaway², Mouritz et al.³), cars and trains, wind turbine blades, to prosthetics and biomaterials for sutures and implants. These materials’ durability is affected by complex service scenarios, which may include a combination of a) thermo-mechanical loading, e.g., freeze-thaw cycles in civil infrastructure⁴, subsonic/supersonic flight profiles⁵, wear in metal-backed polyethylene⁶); b) degradation due to environmental and chemical agents, e.g., sea water, de-icing, hydraulic fluid for aerospace and naval structures^7-10, degradation of polymethylmethacrylate dental composites due to saliva¹¹; c) complex interactions of materials in fastened or bonded joints, e.g., galvanic corrosion and debonding between dissimilar materials, whether in a carbon/fiber patch repair on an aircraft aluminum skin, or a carbon/PEEK bone plate fastened by stainless steel¹².

There is unfortunately limited knowledge of the impact of concurrent in-service stimuli on the long-term durability of these materials. Most polymers may be categorized as viscoelastic materials. Mechanical loadings and environmental conditions significantly influence the viscoelastic response of polymers. Hence, reliable models for these materials’ long-term behavior should be able to incorporate time-dependent responses to coupled hygrothermal, mechanical, chemical stimuli. This in turn will improve design predictions, safety and condition-based maintenance/replacement protocols.

There is a large literature body on experimental testing on hygrothermal effects, for example hygrothermal diffusion tests: if the scale of the samples allows it, the material samples may be positioned in a chamber at desired humidity and temperature levels. The samples are removed periodically to measure their mass and/or volume changes for a given amount of time, from weeks to years^10,13-17. The hygrothermal test may be followed by mechanical testing, i.e., residual static/fatigue strength/fracture mechanics testing^17-19, which only gives information on the effect of hygrothermal stimulus on the mechanical responses of materials. Test data may be fitted to diffusion models of varying complexity, from simple Fickian diffusion to models that include dependency on concentration, stress, temperature, reversible physical aging/plasticization and irreversible chemical reactions. This experimental output may be further incorporated in structural analyses.

Few authors have addressed the impact of simultaneous hygrothermal and mechanical stimuli. Among those researching FRP composites, Neumann and Garom²⁰ immersed stressed and unstressed specimens in distilled water. Stress was applied by positioning the specimens inside compressed stainless steel springs, tuning the load by using different spring stiffnesses and compressive loads. A similar procedure is reported by Wan et al.²¹. Helbling and Karbhari²² employed a bending fixture inside an environmental chamber for different relative humidity percentages (RH%) and temperature levels. The pre-conditioned specimens were subjected to a given bending strain level, corresponding to a percentage of the static ultimate tensile strain for that composite. Kasturiarachchi and Pritchard²³ prepared a stainless steel 4-point bending jig (one per specimen) that was positioned on a shelf in a large glass desiccator. The desiccator was partially filled with distilled water, had small leaks to prevent the buildup of pressure, and was placed in a humidity chamber at 95% RH. Gellert and Turley⁷ investigated marine-grade FRP composite specimens for their durability under combined creep loading and 100% RH. Their samples were loaded in 4-point bending at a constant load equal to 20% of the failure static flexure load, while fully immersed in sea water. The creep deflection was acquired periodically by using a thickness gauge between the outer surface of the beam in the central cross-section, and a glass plate (it is inferred that such measurement was performed outside the chamber). Abdel-Magid et al.²⁴ placed samples of glass/epoxy in an Invar environmental fixture which was provided by NASA Langley, as the specimens were loaded in tension along the fiber direction, at 20% of the ultimate axial load. Ellyin and Rohrbarcher²⁵ ran hygrothermal tests for up to 140 days, and then tested the specimens in fatigue on a hydraulic testing machine. The specimens were wrapped in a wet cheese cloth connected to a tube and a water supply. Earl et al.²⁶ positioned their loading fixture and the specimens in a large environmental chamber (5.5 m³).

As discussed in many experimental studies, the environmental conditions affect the polymers’ mechanical properties and responses. Some limited experiments also show that the existence of mechanical stress/strain influences the diffusion process in the polymers. Hence, to enhance understanding on the overall performance of polymer-based materials under both mechanical and non-mechanical effects, there is a need for concurrent testing.

There were several objectives behind the design of the testing platform discussed in this paper. First, the platform is part of the experimental setup in a multi-year investigation on the hygrothermo-mechanical behavior of different types of FRP sandwich composites for wind turbine and naval engineering applications. The test data are used to calibrate the parameters in the viscoelastic constitutive equations for the polymeric composites. The constitutive models are based on the work developed over the years by Muliana and collaborators^27-30. The second objective was to have a low-cost and user-friendly testing platform, for example one that could be easily relocated in a laboratory (e.g., to a scale for mass measurements, or to the source of the fluid, e.g., one coming from a faucet, a fumehood or a flammable cabinet). The third goal was to create a testing platform that is resistant to a number of chemicals commonly used in service (particularly hydraulic fluid, de-icing, cleaning solvents for aerospace applications^8-10), thus specimens could be immersed in such chemicals, and their durability could be assessed.

The chamber (Figure 1) was constructed with high-density polyethylene, which has high chemical resistance. As mentioned above, it is expected that future work will include hygrothermo-mechanical investigation of composites immersed in hydraulic fluid, de-icing, cleaning solvents. Since thermal regulation is an integral aspect of testing, expanded polystyrene foam was fit around the sides of the tank and secured in place by tape and the steel frame itself, to prevent heat exchange with the environment.

The lid of the chamber (Figure 2) was manufactured from transparent, 9.525 mm-thick polycarbonate, allowing the users to observe the specimens during testing without disturbing the test. The lid is secured in place by aluminum T-bars, which were machined to slide under overhanging brackets on the sides of the tank.

Bending in the specimens is enacted by three aluminum blocks, which hang down from the lid, and are fastened through slots in the lid. The three blocks allow up to four specimens to be tested at one time, while the lid slots allow the block spacing to be adjusted depending on the length of the specimens. Each block is rounded at the contact edge to a 12.7 mm diameter, in adherence to ASTM standard D790-10. The specimens are positioned beneath two of the three blocks, with an upward force applied at its center to induce bending (Figures 1-2).

The apparatus was designed with maximum versatility and ease-of-use in mind. Casters with 41.275 mm diameter are fastened beneath the chamber for mobility purposes. Above them, the tank is supported by a welded steel frame with a wire mesh bottom and cross beams for support. Angle stock spacers for the outside tank corners were manufactured to keep the insulation from being crushed by the overhead weight and displacement gauges (string pot apparatus, discussed later). Around the top, angle stock was used again for framing. Pulley and string potentiometer systems to measure mid-span deflection are mounted on four steel, square-tubing arches (Figure 3). The center two arches out of these four carry the string potentiometers and are adjustable to account for specimen versatility. The string potentiometers were constructed using a torsional spring (as can be found in retractable key lanyards) and potentiometers with three-pronged electronic outputs. The pulleys are aligned and mounted for use with a steel cable running from a rigid connection by the specimen to a hanging rod over the side of the chamber for adjustable weight application.

The load is applied to the specimen using a series of cables, pulleys, linkages and bolts. First, the specimen is placed into the U-bolt so that the 10 mm cross bar is contacting the middle of the span. A 9.525 mm diameter steel rod with eye bolts at each end is then connected to the U-bolt. This steel connection passes through the lid of the chamber. A steel cable and Kevlar thread are attached to the eyebolt opposite the U-bolt. This allows the Kevlar thread from the string potentiometer to read data from a rigid point. The steel cable continues upwards and passes over two pulleys that allow the load to be applied at the periphery of the tank. The cable is then attached to a 9.525 mm diameter steel rod that serves as a slotted weight hanger. This hanger provides a place where the slotted weights can be set in order to apply the desired load.