Mechanical and thermal properties of glass fiber–vinyl ester resin composite for pipeline repair exposed to hot-wet conditioning

Abstract

Fiber-reinforced composites are a well-recognized option for repair and rehabilitation of the pipelines for the oil and gas industry. Infilled composite sleeve system provides an effective rehabilitation solution, where the sleeve acts as prime reinforcement without any direct contact with steel. However, the long-term performance of the repair is dependent, in part, on the effect of hygrothermal ageing of the composites. In this publication, glass transition temperature and mechanical properties are compared for glass-fiber reinforced vinyl ester composite, both as-manufactured and after hot-wet conditioning at 80℃. The tensile and shear strength reduced substantially during conditioning, whilst the elastic modulus was relatively stable. The average glass transition temperature of the composite dropped from the as-manufactured value of 110℃ to 97℃ and 101℃, after 1000 and 3000 h of conditioning, respectively, indicating that it is stable and that the composite is suitable for use as a pipeline repair material operating at 80℃. The results indicate that a 1000 h conditioning period, specified as a minimum period in ISO/TS 24817 is suitable for representing long-term properties for stiffness-based designs for the composite material and conditioning temperature investigated.

Keywords

Glass fibers vinyl ester resin polymer-matrix composite mechanical properties thermal properties glass transition temperature pipeline

Introduction

Fiber composites have become one of the well-accepted options for repairing steel pipelines. In the oil and gas (O&G) industry, prompt action is required to repair damage, to minimize financial losses.¹ Composite repair systems are now being designed for both in-air and underwater application conditions, for room temperature (RT) and elevated temperature service. However, polymeric composites are vulnerable to property degradation due to moisture uptake.^2–4 The moisture absorption during environmental exposuremay be the result of absorption of water within the composite constituents themselves or can be an indication of damage from matrix cracks and debonding at the fiber–matrix interfaces.^5–7 The ingress of moisture may cause the composite to have reduced damage tolerance and structural durability.^8–11 Hence, the performance and structural integrity of repaired pipelines depend on the long-term properties of the fiber-reinforced composites, and so this requires detailed investigation.

Traditionally, there are two types of repair systems – flexible “wet lay-up system” and pre-cured “layered system” that are applied in the repair of defective pipelines in the O&G industry.¹² However, the leak containment of the failed pipe is not fully successful through the previous options and the necessity of easily applicable solution is still sought using composites. Stand-off sleeve systems provide higher structural integrity than both flexible lay-up system and pre-cured layered systems. Most of the heavy duty repair technologies are based on this principle. The infilled stand-off half sleeve concept is a new technology in composite rehabilitation arena, which is similar to that of metal split-sleeve concept used by PLIDCO®.¹³ This type of repair system uses an infill layer, which acts as a load transfer medium and primarily experiences compression under internal pressure. Extensive study on the properties of various infills was already conducted.¹⁴ The main reinforcement is provided by the composite sleeve without any direct contact with steel. Internal pressure in the pipe produces primarily circumferential stress on the sleeve. There is a lack of sufficient understanding on the performance of the composite when used for infilled sleeve rehabilitation in underwater service conditions, demanded by current guidelines. Moreover, high pressure oil and gas pipelines operating in offshore production facilities are often susceptible to high temperature, which should be considered in designing the pipelines.^15,16 Thus, the effect of hot-wet condition, which is the extreme in-service condition, needs to be studied for the repair system to be applied in these pipelines. This study is thus focused on the performance of the composite sleeve of an infilled sleeve system when subjected to internal pressure and long-term hygrothermal ageing.

This paper investigates the effect of hot-wet conditioning on the mechanical and thermal properties of glass fiber reinforced composites resembling a composite repair system for steel pipelines subjected to elevated temperature and in underwater condition. Glass fiber–vinyl ester composite laminates were manufactured and tested for tensile, interlaminar shear and thermal properties,in both the as-manufactured (AM) (or unconditioned) and hot-wet conditioned states. The experiments were chosen based on the requirements of ISO/TS 24817,¹⁷ which is a standard for qualification of composite repairs for the oil and gas pipelines.

The study was conducted under two different objectives. Firstly, there was a requirement to understand the effect of hot-wet conditioning on the composite in question, in order to allow it to be assessed for repair applications in the O&G industry. The second was to assess whether the minimum “long-term” conditioning time of 1000 h presented in ISO/TS 24817¹⁷ is justified for representing the actual long-term condition of a glass fiber–vinyl ester resin composite laminate, since long-term measurements are typically performed with laminates conditioned until saturation, rather than for a specified periods of time. Besides, there is hardly any study that addresses hygrothermal conditioning for glass fiber–vinyl ester resin intended as infilled pipeline rehabilitation satisfying current standard. The orientation and distribution of fibers were also intended to represent a cylindrical sleeve as reinforcement in the repair system. This assessment is made with the understanding that the standard needs to provide a balance between scientific outcomes and commercial product deployment times; hence, test times should be minimized wherever possible.

Experimental methodology

Materials

Glass fiber-reinforced vinyl ester resin composite offers economical advantage to carbon fiber-reinforced composites under certain installation conditions. For instance, glass is and less susceptible to galvanic corrosion compared to carbon when in direct contact with metal pipes. A comprehensive study on the mechanical performance of glass fiber-reinforced epoxy resin, used for overwrap repair of submerged pipelines, has already been performed.¹⁸ Vinyl ester resin, which has excellent chemical resistance and long-term performance characteristics, is identified as a viable alternative to epoxy resin in the O&G industry.Vinyl ester has better chemical resistance than cheaper polyester resins, especially hydrolytic stability, and at the same time offers greater control over cure rate and reaction conditions than epoxy resins.¹⁹ Moreover, an industry-specific evaluation of this resin for application in pipeline repair is identified as a gap in the open literature. For this reason, fiber-reinforced composites made up of vinyl ester resin and E-glass fibers were used in this study.

The resin system was Derakane 411-350 Epoxy Vinyl Ester manufactured by Ashland. This resin has a viscosity of 350 mPa s. The fiber reinforcement was a 1216 gsm 0/90 biaxial glass fiber non-crimp fabric (NCF) supplied by Colan Fabrics consisting of 567 gsm warp (0° direction) yarns and 638 gsm weft (90° direction) yarns stitched together by 11 gsm polyester yarns.

Specimen preparation

All panels were manufactured via a vacuum bag resin infusion (VBRI) process. VBRI is a process similar to vacuum-assisted resin transfer molding (VARTM) as available in literature.²⁰ A dry fibrous perform is laid up on a single-sided mold, and a permeable resin distribution medium is placed on top. This layup is vacuum bagged, and resin is injected at atmospheric pressure, whilst vacuum is maintained at the vent. The infused panels were post-cured in an oven at a temperature of 80℃ for 12 h. This manufacture method was selected, as it was seen as a likely candidate for repair systems in the O&G industry.

Tensile specimens

The longitudinal tension (LT) and transverse tension (TT) coupons were prepared in accordance with ASTM D3039.²¹ A panel of [0/90]₄ configuration was laid up; the four plies are stacked together with no reflection about the centerline with respect to the thickness. The LT specimens were cut along the 0° direction of the panel, while the TT specimens were cut along the 90° direction. The tensile coupons were 250 mm long by 25 mm wide, with an average thickness of 3.6 mm.

Interlaminar shear specimens

The interlaminar shear specimens were prepared in accordance with ASTM D5379.²² A composite block with a laminate configuration of [0/90]₉₀ was manufactured with a thickness of approximately 80 mm. The block was then cut into a thickness of 4.5 mm and machined into 76 mm long and 20 mm wide specimens. A V-notch was machined centrally in both sides. The specimens were prepared such that the intended shear plane between the V-notches would result in a shear failure through the thickness of the laminate. A pair of bonded strain gauges oriented in +45° and −45° directions was attached to the specimen to measure the shear strain. The strain gauges were sealed using flexible silicon glue protecting the circuit from the ingress of moisture during the hot-wet conditioning. This sealant was only used to cover the area directly surrounding the gauges, leaving the other side of the specimen exposed to the water.

Glass transition temperature specimens

The glass transition temperature coupons were prepared in accordance with ASTM E1640.²³ These specimens were cut from the same panel prepared for the tensile coupons. The coupons were 60 mm long by 12 mm wide, with an average thickness of 3.6 mm (0.9 mm per ply).

Hot-wet conditioning

The specimens were hot-wet conditioned (HWC) for two different durations in a temperature-controlled water bath, as shown in Figure 1. As the specimens were held in the water bath, they absorbed moisture, causing their mass to increase. The first duration after which testing was performed was 1000 h, selected in accordance with the time defined in ISO/TS 24817¹⁷ as being representative of long-term conditioning. The second duration after which testing was performed was 3000 h, which is shown to be in excess of the time necessary to reach saturation later in this paper. The specimens were assessed for saturation as per ASTM D5229.²⁴ Mass measurements were taken using the Adam PW254 scales with an accuracy of 0.1 mg at every seven days using one of the glass transition test specimens.

The water temperature during HWC was chosen to be 80℃, which was selected based on the maximum temperature allowed for leaking defects according to ISO/TS 24817¹⁷ and ASME PCC-2.²⁵ These standards state that the maximum service temperature of a repair component for non-leaking (Type A) and leaking (Type B) pipes should be 20℃ and 30℃ lower than the T_g, respectively. The T_g for the AM composite was found to be 110℃, following the procedure described in ASTM E1640.²³ Hence, the HWC temperature was chosen to be 80℃ for two reasons. Firstly, this is the highest service temperature to which the composite material would be exposed if used to repair leaking defects. Secondly, moisture uptake is faster at a higher temperature compared to lower.

Test details

The mechanical properties of the AM and HWC composites were determined at RT and at 80℃. The elevated temperature of 80℃ was also used as the upper temperature limit in the mechanical testing, since this is the highest temperature that the composite would be loaded at in service for a leaking defect repairs. The LT specimens were hot-wet conditioned for 1000 and 3000 h, whereas the TT specimens were not. The reason for testing only LT specimens is because there is less fiber in the LT direction, and hence, this direction is expected to be more susceptible to the effect of moisture uptake, since the properties are closer to that of the matrix resin, in comparison to the TT. The test matrix and test conditions are presented in Table 1.

Table 1.

Test matrix summary.

Specimen type	ID	Test temperature(℃)	Teststandards	No. of tests	Loading rate
Tensile, TT	AM	RT	ASTM D 3039²¹	5	2 mm/min
Tensile, TT	AM	80
Tensile, LT	AM	RT
	AM	80
	HWC – 1000 h	RT
	HWC – 1000 h	80
	HWC – 3000 h	RT
	HWC – 3000 h	80
Shear, S	AM	RT	ASTM D5379²²	5	1.3 mm/min
	AM	80
	HWC – 1000 h	RT
	HWC – 1000 h	80
	HWC –3000 h	RT
	HWC –3000 h	80
Glass transition temperature, T_g	AM	RT to 150	ASTM E1640²³	3	1℃/min
	HWC – 1000 h	RT to 150
	HWC – 3000 h	RT to 150

Mechanical tests

All the mechanical tests were carried out using a 100 kN MTS hydraulic testing machine. Figure 2 shows the set-up and equipment used for testing at 80℃. The AM specimens were preheated in an oven at 80℃ for at least 30 min prior to testing. The HWC specimens were preheated in a waterbath at 80℃, for at least 30 min prior to being placed into temperature chamber, so that they would acclimatize quickly to minimize the possibility of them drying out. Once the chamber temperature reached 80℃ and was maintained for 10 min, the tests on HWC specimens were started.

During tensile testing (LT and TT), a biaxial extensometer was placed centrally across the specimen to measure the strain in both the longitudinal and transverse directions. The strain in the through-thickness direction of the specimens was not measured. Once the strain reached 3000 μɛ, the tests were halted and the extensometer removed. The test was then resumed without any strain measurement. The tensile modulus and Poisson’s ratio were calculated from the stress–strain curve as specified in the standard. The V-notched (Iosipescu) shear specimens (S) were tested using a Wyoming shear testing fixture, and the shear modulus calculated as specified in the standard.

Thermal analysis

Dynamic mechanical analysis (DMA) was used to determine the T_g of the composite. The DMA test was carried out using a calibrated DMA Q800 with Universal Analysis 2000 V5.1 Build 92 manufactured by TA Instruments. Figure 3 shows the specimen mounted on the DMA apparatus for testing. The specimen was clamped in the three-point bending fixture with a span of 35 mm and was tested in accordance with ASTM E1640.²³ The heating rate was 1℃/min from RT to 150℃, whilst applying a strain of 1% at a frequency of 1 Hz. The T_g was determined based on the onset of rapid storage modulus reduction with temperature increase (T_g). An alternative form of glass transition measurement, the peak of the tan δ signal (T_t), was also measured. T_g was used as the basis for setting the temperatures during HWC and elevated temperature testing.

Experimental results and observations

Moisture absorption

A color change from greenish white (AM) to brownish (HWC) on the specimen surface was observed due to HWC, which can be seen later in the Results section, although there was no apparent physical change. Change of appearance was also reported in GFRP plates due to corrosive environment.²⁶ The water in the water bath produced a mild odor. These phenomena were also reported in previous literature when glass fiber-reinforced composites were subjected to hygrothermal ageing.^27–30

Figure 4 shows the moisture absorption of a T_g specimen over 3000 h of conditioning. The percentage moisture content and mass change are presented against immersion time. The highest moisture content is found to be about 0.22%, which was reached within 672 h (28 days). In comparison, other studies^31–33 conducted on glass fiber–vinyl ester composite conditioned a range of temperature from 15 to 80℃ showed that the composite absorbed moisture by about 0.5–1.1%, indicating that the moisture uptake of the composite presented in this study is relatively low. ASTM D5229²⁴ suggests an average moisture content of the material changes by less than 0.01% within the span of the reference time period, one week, as an indication of equilibrium. The moisture content was found to reach equilibrium at 1008 h.

The moisture absorption behavior in Figure 4 suggests that most of the moisture is absorbed within the first 168 h of HWC. The diffusion pattern can be correlated using the Fickian diffusion pattern^34,35 which is discussed and adopted in previous studies.^3,6 It can also be seen that the moisture absorption as the function of time has a maximum value of 0.22%, following which it decreases. Hence, considering this moisture content of saturation as the maximum equilibrium amount of absorption, M_∞, the moisture diffusion coefficient, D can be determined as 2.22 × 10⁻⁶mm²/s which are within the range of polymer composite absorption characteristics.^9,28 The resulted prediction plot is also shown in Figure 4. This relation of moisture uptake is similar to the diffusion pattern shown by Aniskevich et al.⁹ at an elevated temperature close to the glass transition temperature.

As stated previously, the HWC glass composite appears to reach saturation after about 1008 h of immersion. Similar findings were observed in literature,³¹ wherein the glass fiber–vinyl ester composite conditioned in saline water at 30℃ reached saturation at around 2000 h (90 days). Another study suggested that the saturation of glass fiber–vinyl ester composite conditioned under saline water at a temperature range of 4–38℃ occurred at around 3000 h.³³ Thus, the results suggest that the moisture uptake and saturation of composites are accelerated at elevated temperature.

Table 2.

Summary of mechanical and thermal properties of composite.

Properties		AM		HWC – 1000		HWC – 3000
Properties		RT	80℃	RT	80℃	RT	80℃
LT	Strength (MPa)	427.1 (16.4)	395.9 (11.2)	186.7 (7.9)	147.5 (11.2)	153.2 (3.38)	127.7 (8.92)
	modulus (GPa)	24.6 (0.1)	23.0 (0.5)	23.2 (0.7)	23.2 (0.8)	20.7 (0.30)	22.3 (0.53)
	Poisson’s ratio	0.126 (0.010)	0.238 (0.018)	0.113 (0.010)	0.089 (0.020)	0.076 (0.007)	0.104 (0.012)
TT	Strength (MPa)	501.6 (15.3)	438.6 (23.3)	–	–	–	–
	modulus (GPa)	28.0 (0.2)	26.7 (0.3)	–	–	–	–
	Poisson’s ratio	0.133 (0.013)	0.203 (0.059)	–	–	–	–
S	Shear strength(MPa)	30.0 (2.6)	18.8 (4.1)	10.2 (1.8)	7.4 (2.9)	9.2 (1.9)	6.7 (2.7)
	Shear modulus(GPa)	4.15 (0.10)	3.13 (0.19)	3.27^a (0.16)	5.58^b (N/A)	4.55^b (N/A)	^c
	Failure strain (μɛ)	8008 (515)	7120 (1457)	2919^a (468)	1373^b (N/A)	1823^b (N/A)	^c
T_g (℃)		110 (1.2)		97 (1.7)		101 (0.6)
T_t (℃)		128 (0.3)		113 (0.5)		115 (0.4)

Only two readings used to calculate average, where failure occurred along notched shear plane.

Only one reading used, where failure occurred along notched shear plane.

Results neglected since failure occurred near the test fixture supports.

Mechanical properties

The mechanical properties of both the AM and HWC composites at RT and 80℃ are summarized in Table 2 with the bracketed values indicating the standard deviations. The results are described in the following section.

Laminate tensile properties

When tested at RT, respective strengths and elastic moduli of the LT specimens were found to be as follows: 427 MPa and 24.6 GPa for AM, 187 MPa and 23.2 GPa for the 1000 h HWC specimens, and 153 MPa and 20.7 GPa for the 3000 h HWC specimens. Hence, the 3000 h HWC specimens have the lowest strength and stiffness, and the AM specimens have the highest strength and stiffness, with the 1000 h HWC specimens in between.

When tested at 80℃, the respective strengths and elastic moduli of the LT specimens were found to be as follows: 396 MPa and 23.0 GPa for AM, 148 MPa and 23.3 GPa for 1000 h HWC specimens, and 128 MPa and 22.3 GPa for 3000 HWC specimens. These results indicate a gradual reduction in the tensile strength. The tensile modulus at 80℃ peaks at 1000 h conditioning, with lower values AM and after 3000 h HWC specimens: this difference is likely the result of experimental scatter.

For a given specimen condition, if the LT strength and modulus are compared based on temperature change, it is clear that the strength is reduced, but the elastic modulus is relatively unaffected. Furthermore, if the LT and TT specimens are compared in the AM condition, both the strength and modulus are higher in the TT direction for each test temperature; this is the result of higher fiber content in the TT direction.

Figure 5 shows the typical load-extension behavior of the AM specimens, whilst Figure 6 shows that of the LT specimens after 1000 and 3000 h HWC. The behavior is linear to failure with small disturbances seen at approximately 2 mm cross-head extension due to the disengagement of the extensometer. However, the AM specimens show abrupt change near the peak load as indicated by the sudden drop in the load-extension line due to the initiation of failure in the specimens.

The typical failure modes of unconditioned tensile specimens are shown in Figure 7. The failure of the unconditioned AM specimens for both RT and ET are similar. The LT failures are progressive indicating partial tear of fibers near the grip and finally total collapse of the laminate releasing a cluster of fiber fragments as seen from Figure 7(a). This type of failure was also reported for glass fiber reinforced composite previously.^36,37 The crack covers the entire gauge area where matrix cracking and fiber pull out are also observed. However, the damage area in the TT specimens was more limited than that of the LT specimens, which can be seen through comparison of Figure 7(b) to Figure 7(a).

The typical failure of the LT 1000 h and 3000 h HWC specimens is shown in Figure 8. The failure pattern of the specimens tested at RT and elevated temperature is similar. Small amounts of fiber–matrix debonding are observed for both specimens. The dominant failure of the unconditioned specimens occurred at or within 5 mm of the grip which can be defined as failure “Type A” according to ASTM D3039.²¹

Interlaminar shear properties

Table 2 summarizes the shear properties of the composite. The shear strength and modulus of the unconditioned AM specimens is around 30 MPa and 4.2 GPa, respectively, when tested at RT and is reduced by about 37% and 25%, respectively when tested at 80℃. The strength of the specimens with 1000 h HWC applied is 10 MPa at RT and is reduced by 28% when tested at 80℃, while the stiffness is at 3.3 GPa when at RT but is increased by 71% when tested at 80℃. The strength of the specimens with 3000 hHWC applied is 9 MPa when tested at RT but is reduced to approximately 7 MPa when tested at 80℃ with the stiffness almost the same.

The typical shear stress–strain behavior of the AM specimens is shown in Figure 9, whilst that of the HWC specimens is shown in Figure 10. The stress–strain relationships of the AM and HWC specimens is almost linearly elastic up to failure. The reduction of strain with the increment of stress at the end of the stress–strain plots near the failure of the specimens indicates that either one or two of the strain gauges is no longer functioning, which generally occurs at failure of the specimens. The peak shear stress of specimens tested at RT is higher than that when tested at elevated temperature for each condition tested.

Figure 11(a) shows failed shear specimens in the test fixture. Figure 11(b) presents the typical failure pattern of the shear specimens at RT and 80℃. It can be seen that the cracks start at the root of the top notch in the AM specimens and progress along the shear plane down to the root on the bottom notch. After the peak load, the failure is sudden and results in complete separation of the specimen. Figure 12 shows the typical failure pattern of the HWC shear specimens. The cracks primarily occur near the notch. However, in some instances, there are two cracks in the HWC specimens located adjacent to the edge of the test rig used to apply the load; instances of this are noted in Table 2. Failure of the HWC specimens tested at both RT and 80℃ is similar.

The instances where cracks and failures initiated on a shear plane away from that located between the V-notches may be the result of the flexible sealant applied to the strain gauges to protect them from moisture-induced damage across the conditioning period. For all specimens, shear strength was calculated using the area across the notch, even when the failure occurred near the test fixture supports. In these cases, the shear plane located between the V-notches can be considered to have strength in excess of the stated individual values. However, the average shear modulus and failure strain presented in Table 2 are calculated only from specimens that failed along the shear plane between the V-notches; no modulus values are presented for specimens where failure occurred along the shear plane below the loading points, due to them being considered unreliable.

Glass transition temperatures

Figure 13 shows typical plots of storage modulus and tan δ signals against the temperature for the composite. T_g and T_t of the AM specimens are found to be 110℃ and 128℃, respectively. In case of the 1000 h HWC specimens, T_g and T_t were measured to be 97℃ and 113℃, respectively, whereas for the 3000 h HWC specimens, T_g and T_t were measured to be 101℃ and 115℃, respectively.

The storage modulus plots suggest that the modulus of the AM and HWC specimens is similar; however, the transition range is shifted to the left in the case of the HWC specimens by approximately 13℃. This behavior is similar to the tan δ signals, where there is approximately 15℃ shift to the left for the HWC specimens.

Discussion

Specimens were tested AM, after 1000 h of HWC and after 3000 h of HWC. At 3000 h of conditioning, the specimens were considered to be saturated as discussed in the previous section. This information will now be used to assess the 1000 h HWC time for suitability to represent long-term performance of this composite from two angles. Firstly, the upper service temperature limits, based on the reduction in T_g. Secondly, the change in mechanical performance over time.

The first angle is the upper service limit. T_g was measured in the research presented in this paper, both because it is a requirement of ISO/TS 24817¹⁷ and ASME PCC-2,²⁵ but also to understand the temperature at which the composite structure is no longer able to bear load. It is generally understood that at the amorphous T_g, the matrix of the composite will change form; above T_g, the matrix will be of a rubbery form whilst below T_g the matrix will be in glassy form. Furthermore, it is understood that the composites undergo a significant reduction in their ability to bear loads at temperature above T_g; thus, using T_g as a reference point for determining a suitable upper temperature limit is a reasonable approach.

The findings presented in this paper indicate that the T_g changes from an average AM value of 110℃, to 97℃ after 1000 h HWC, then back up to 101℃ after 3000 h. The apparent increase in T_g is unexpected, since T_g generally reduces with HWC time, and may either be attributed to experimental scatter or a post-cure type effect brought on by the increased conditioning time at 80℃. Since the upper temperature limit for leaking-defects manufactured from this composite is 80℃ (30℃ less than the AM T_g), and the T_g is above 80℃ at saturation, it is a reasonable assumption that the T_g of the composite will be acceptable for the duration of the repair life. Therefore, the structure can be considered solid and capable of bearing load up to the upper temperature limit of 80℃ for the duration of its lifetime. Should the only measurements been on the 1000 h HWC samples, the same conclusions would have been arrived at, so from this perspective, the 1000 h HWC time is suitable for determining long-term performance.

The second angle considered is mechanical performance. Mechanical performance was measured using the requirements outlined in ISO/TS 24817.¹⁷ There are no strict requirements on the performance of the composite AM or after HWC presented in the standard. The tensile stiffness and strength in the TT direction were found to be superior to those measured in the LT direction, as can be seen from the results presented in Table 2, due to higher fiber content in the TT direction. The remainder of this discussion will focus on the LT and S properties, since these were tested AM and after being subjected to 1000 and 3000 h HWC, whereas the TT specimens were not.

The average LT-strength was found to reduce with conditioning time when the AM results are compared to the HWC specimens. The results of Table 2 show that there is a significant reduction in the LT strength for 1000 h (reductions of 56.3% at RT and 62.7% at ET) and 3000 h (reductions of 64.1%at RT and 67.7% at ET), indicating that 1000 h HWC is not suitable for assessing long-term tensile strength of this composite. In the case of the shear strength, the AM specimens also exhibited higher strength than the 1000 h (reductions of 66.0% at RT and 60.9% at ET) and 3000 h (reductions of 69.2% at RT and 64.5% at ET) conditioned specimens. GFRP bars with vinyl ester resin also experienced reduction in mechanical properties over longer period of time.³⁸ It is noted that the strength reduction was higher in the case of the 1000 h conditioning time, and so the minimum 1000 hlong-term requirement can be considered as acceptable in this instance.

The elastic modulus of the LT HWC specimens was reduced compared to the AM specimens in all but one instance (1000 h HWC tested at ET, exhibiting 1.2% gain in modulus). However, in general, the reduction in LT elastic moduli was much lower compared to the reductions in strength. Hence, the difference in the values when comparing the 1000 h (reductions of 5.8% at RT and −1.2% at ET) and 3000 h (reductions of 15.9% at RT and 2.9% at ET) HWC results are reduced. The reductions in shear moduli are difficult to comment on based on these findings, due to the limited number of specimens, which were successfully tested until failure on the shear plane containing the strain gauges, as discussed earlier in shear properties section. Therefore, using the values presented for the elastic modulus, and noting that most of the differences are within expected experiment scatter (less than 10%), the 1000 h conditioning time is considered as suitable for representing the long-term values of this composite.

The previous two paragraphs present an argument for and against using the 1000 h conditioning time to represent long-term properties of the composite. If strength is a driving factor in the design, with elastic modulus being considered less important, then the specimens should be conditioned until saturation in order to represent long-term properties. Conversely, if elastic modulus or stiffness is the driving factor in the design, with strength being less important because the structure is operating at much lower stresses than are allowed, then the 1000 h conditioning time is considered suitable. In the case at hand, assessment of the current composite system against ISO/TS 24817¹⁷ would indicate that the 1000 h conditioning requirement is suitable, since the standard specifies determination of stiffness (and not strength), and these are the parameters used in most design equations within the standard.

It is important to note that an HWC temperature of 80℃ was used in this case, as it is the highest service temperature that would be acceptable for a leaking defect repair constructed using the glass fiber–vinyl ester composite in question. This temperature therefore represents the time to saturation under HWC for this particular composite and application in its highest temperature exposure condition. Should HWC have been performed at a lower temperature, it is possible that the time to saturation would have increased; hence, the specimens conditioned until saturation would have higher relative moisture content than those HWC for 1000 h, compared to the relative moisture content difference of the saturated and 1000 h HWC specimens conditioned at 80℃.

The authors of this paper recognize that ISO/TS 24817¹⁷ is an industry standard, and it is difficult to specify long-term conditioning as needing to be performed until saturation of the composite as this would delay deployment of products to the market. Whilst it is true that conditioning to saturation presents the most complete understanding, it is possible that long-term performance can be determined at 1000 h HWC using a suitably high conditioning temperature to allow rapid moisture uptake, provided that the design is stiffness based and is operating well below the strength of the composite. This needs to be addressed on a case-by-case basis. However, where possible, conditioning should be performed until saturation of the composite is achieved.

Finally, using these findings, the use of the 1000 h conditioning time to represent long-term composite properties, specified as the minimum period in ISO/TS 24817,¹⁷ was assessed for suitability with this composite. Since the tensile elastic modulus is relatively stable between the 1000 h and 3000 h conditioning times, coupled with the finding that the glass transition temperature does not reduce between the 1000 and 3000 h conditioning times, the 1000 h conditioning time is suitable for assessing long-term properties of materials used in stiffness-driven designs that are operating at stresses well below expected failure of the composite. Conversely, the finding that the tensile and shear strengths reduce considerably from the AM condition to the 1000 h condition, and there is further reduction for 3000 h conditioning, indicates that the 1000 h conditioning requirement is unsuitable for assessing long-term material properties used in construction of structures operating close to their failure stress.

Conclusions

In this study, the tensile and shear mechanical properties, and glass transition temperature of a glass fiber–vinyl ester composites were investigated. The composite specimens were tested AM and after being subjected to hot-wet conditioning for 1000 and 3000 h. Mechanical tests were performed at RT and 80℃. The following conclusions can be drawn from this study:

The composites increased in mass by 0.22% due to moisture uptake and reached saturation after 1000 h, indicating that they were well saturated by 3000 h conditioning.

The tensile strength reduced substantially over the hot-wet conditioning period, and the decrease was considerable between the 1000 h and 3000 h conditioning times, which is attributed primarily to softening of the composite matrix. In contrast, the stiffness was relatively stable between the AM and the two hot-wet conditioned states.

The glass transition temperature of the AM composite was measured to be 110℃. This is reduced to about 97℃ after 1000 h conditioning and 101℃ after 3000 h conditioning, indicating that the composites are suitable for pipeline repair in continuous service at 80℃.

The finding that the 1000 h long-term conditioning time is acceptable for stiffness-based designs indicates that use of the duration in materials characterization as per current ISO standard is justified.

Footnotes

Acknowledgement

The support and technical feedback from Dr. Paul Falzon of Cooperative Research Centre for Advanced Composite Structures (CRC-ACS), Australia during this study are acknowledged.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was undertaken as part of P1.3 Deepwater Composites within the CRC-ACS research program, established and supported under the Australian Government's Cooperative Research Centre’s Program.

References

Shamsuddoha

Islam

Aravinthan

. Effectiveness of using fibre-reinforced polymer composites for underwater steel pipeline repairs. Compos Struct 2013; 100: 40–54.

Ray

. Temperature effect during humid ageing on interfaces of glass and carbon fibres reinforced epoxy composites. J Colloid Interface Sci 2006; 298: 111–117.

Jiang

Kolstein

Bijlaard

FSK

. Moisture diffusion in glass–fiber-reinforced polymer composite bridge under hot/wet environment. Compos Part B: Eng 2013; 45: 407–116.

Alessi

Pitarresi

Spadaro

. Effect of hydrothermal ageing on the thermal and delamination fracture behaviour of CFRP composites. Compos Part B: Eng 2014; 67: 145–153.

Chin

Nguyen

Aouadi

. Sorption and diffusion of water, salt water, and concrete pore solution in composite matrices. J Appl Polym Sci 1999; 71: 483–492.

Ellyin

Maser

. Environmental effects on the mechanical properties of glass-fiber epoxy composite tubular specimens. Compos Sci Technol 2004; 64: 1863–1874.

Soykok

Sayman

Pasinli

. Effects of hot water aging on failure behavior of mechanically fastened glass fiber/epoxy composite joints. Compos Part B: Eng 2013; 54: 59–70.

McBagonluri

Garcia

Hayes

. Characterization of fatigue and combined environment on durability performance of glass/vinyl ester composite for infrastructure applications. Int J Fatigue 2000; 22: 53–64.

Aniskevich

Arnautov

. Mechanical properties of pultruded glass fiber-reinforced plastic after moistening. Compos Struct 2012; 94: 2914–2919.

10.

Engindeniz

Zureick

. Deflection response of glass fiber-reinforced pultruded components in hot weather climates. J Compos Constr 2008; 12: 355–363.

11.

Atas

Dogan

. An experimental investigation on the repeated impact response of glass/epoxy composites subjected to thermal ageing. Compos Part B: Eng 2015; 75: 127–134.

12.

Alexander C and Francini B. State of the art assessment of composite systems used to repair transmission pipelines. In: 6th international pipeline conference, Calgary, 25–29 September 2006, pp. 823–830. Alberta: ASME.

13.

PLIDCO Split+Sleeve. Permanently repairs a variety of pipelines, Ohio, USA: The Pipe Line Development Company, 2012.

14.

Shamsuddoha

Islam

Aravinthan

. Characterisation of mechanical and thermal properties of epoxy grouts for composite repair of steel pipelines. Mater Des 2013; 52: 315–327.

15.

Klever

Palmer

Kyriakides

Limit-state design of high-temperature pipelines. In: Murray

Fernandez

Thygesen

(eds). International conference on offshore mechanics and arctic engineering (OMAE-13), Houston, TX: American Society of Mechanical Engineers, 1994, pp. 77–92.

16.

Mason JF. Pipe liners for corrosive high temperature oil and gas production applications. In: Corrosion 97, New Orleans, 9–14 March 1997, pp. 80–84. Louisiana: NACE International.

17.

ISO/TS 24817. Petroleum, petrochemical and natural gas industries – composite repairs of pipework – qualification and design, installation, testing and inspection, London: International Organization for Standardization (ISO), 2006.

18.

Djukic

Leong

Falzon

. Qualification of a composite system for pipeline repairs under dry, wet, and water-submerged conditions. J Reinf Plast Compos 2014; 33: 566–578.

19.

Li H. Synthesis, characterization and properties of vinyl ester matrix resins. PhD Thesis, Department of Chemistry, Virginia, USA: Virginia Polytechnic Institute and State University, 1998.

20.

Baker

Paton

Kruckenburg

Component form and manufacture. In: Baker

Dutton

Kelly

(eds). Composite materials for aircraft structures, 2nd ed. Reston, Virginia: American Institute of Aeronautics and Astronautics, 2004, pp. 113–169.

21.

ASTM D 3039. Standard test method for tensile properties of polymer matrix composite materials. ASTM Committee D30. West Conshohocken, PA, USA: American Society for Testing and Materials (ASTM), 2008.

22.

ASTM D 5379. Standard test method for shear properties of composite materials by the v-notched beam method. ASTM Committee D30. West Conshohocken, PA, USA: American Society for Testing and Materials, 1998.

23.

ASTM E 1640. Standard test method for assignment of the glass transition temperature by dynamic mechanical analysis. ASTM Committee E37. West Conshohocken, PA, USA: American Society for Testing and Materials, 2009.

24.

ASTM D 5229. Standard test method for moisture absorption properties and equilibrium conditioning of polymer matrix composite materials. ASTM Committee D30. West Conshohocken, PA, USA: American Society for Testing and Materials, 2010.

25.

ASME PCC-2. Repair of pressure equipment and piping, New York: The American Society of Mechanical Engineers, 2006.

26.

Feng

Wang

. Effects of corrosive environments on properties of pultruded GFRP plates. Compos Part B: Eng 2014; 67: 427–433.

27.

Turvey

Wang

. Failure of pultruded GRP single-bolt tension joints under hot–wet conditions. Compos Struct 2007; 77: 514–520.

28.

Adams

Singh

. The dynamic properties of fibre-reinforced polymers exposed to hot, wet conditions. Compos Sci Technol 1996; 56: 977–997.

29.

Nishizaki

Meiarashi

. Long-term deterioration of GFRP in water and moist environment. J Compos Constr 2002; 6: 21–27.

30.

Carra

Carvelli

. Ageing of pultruded glass fibre reinforced polymer composites exposed to combined environmental agents. Compos Struct 2014; 108: 1019–1026.

31.

Kootsookos

Mouritz

. Seawater durability of glass- and carbon-polymer composites. Compos Sci Technol 2004; 64: 1503–1511.

32.

Fraga

Alvarez

Vazquez

. Relationship between dynamic mechanical properties and water absorption of unsaturated polyester and vinyl ester glass fiber composites. J Compos Mater 2003; 37: 1553–1574.

33.

Zhang

Karbhari

Mai

L-Y

. Evaluation of property retention in e-glass/vinylester composites after exposure to salt solution and natural weathering. J Reinf Plast Compos 2000; 19: 704–731.

34.

Crank

. The mathematics of diffusion, Oxford: Clarendon Press, 1975.

35.

Springer

. Environmental effects on composite materials, Lancaster, PA: Technomic Publishing Company, 1981.

36.

Wonderly

Grenestedt

Fernlund

. Comparison of mechanical properties of glass fibre/vinyl ester and carbon fibre/vinyl ester composites. Compos Part B: Eng 2005; 36: 417–426.

37.

Shokrieh MM, Torabizadeh MA and Fereidoon A. Progressive failure analysis of glass/epoxy composites at low temperatures. Strength Mater 2012; 44: 314–324.

38.

Al-Salloum

El-Gamal

Almusallam

. Effect of harsh environmental conditions on the tensile properties of GFRP bars. Compos Part B: Eng 2013; 45: 835–844.