Repeatability of glass fiber reinforced polymer laminate panels manufactured by hand lay-up and vacuum-assisted resin infusion

Abstract

The paper presents a comparison of the test results for the fiber volume fraction, static bending and Charpy impact strength of glass fiber reinforced polymer laminates reinforced by 0/90 fabric and chopped strand mat, produced by the hand lay-up and the vacuum-assisted resin infusion (VARI) method. The laminates were produced under equivalent conditions, with polyester matrix and lay-up areal mass 2100 g/m². In the comparison of the obtained measurement results, similar mechanical performance was stated in the case of the hand lay-up and the VARI production. However, significantly smaller scatter of results and better uniformity of the reinforcement in the matrix with smaller amount of local structure defects was observed for the samples obtained by VARI method. The laminates obtained by VARI method show a much more advantageous coefficient of variation than that of the hand lay-up method, especially in the case of the mat reinforcement.

Keywords

Glass fibers polymer-matrix composites statistical properties/methods mechanical testing

Introduction

Vacuum-assisted resin infusion (VARI) is currently one of the primary techniques in the manufacture of fiber reinforced polymer (FRP) composite products. It enables the production of big elements of a complicated shape, such as energetic windmill blades or yacht hulls.^1–3 In comparison to the classic technology – hand lay-up – VARI makes it possible to shorten the average process time, to improve the health and safety conditions in the process surroundings and to reduce the negative impact of the process on the environment.^1,2 The authors’ previous experience, also gained in the industry sector, showed that the quality of the laminates properly manufactured by the VARI and the hand lay-up methods is comparable. However, in the case of the hand manufacture, there is a very significant discrepancy in repeatability. This characteristic, resulting from the elimination of the human factor from the proper saturation process, is the main trump card of pressure technologies,⁴ including the VARI method. It should be noted that, as regards the product properties, the VARI technique is comparable with the resin transfer molding (RTM) methods,^5,6 and both methods are complimentary; there is a range of products impossible to be manufactured by RTM. Pressure technologies are necessary in the manufacture of products with advanced types of reinforcements, especially of the 3D type.^7–9 It was also stated that the VARI method is, by all means, the best solution for the natural fiber reinforcements.^10–13 VARI is also one of the alternatives in the production of innovative experimental constructions, such as: composite flexible structures,¹⁴ innovative building elevations,¹⁵ and braided pipelines.¹⁶ Another advantage of a fundamental importance is the possibility of running the process without (or with minimal) emission of solvent vapours, which makes it possible to fulfill all the announced implementations of the Industrial Safety Standard, as well as those of the environment protection.

One of the disadvantages of pressure technologies (especially VARI) is the necessity of designing the processes before their run. This requires applying numerical methods, characterized by a high degree of complication, especially at the pre-calculation level. Analyzing the fibrous structures, both for the purposes of mechanical property design^4,17,18 and the analysis of the resin flow in the preform,^4,19–22 is very complex in regard to the calculation apparatus^23,24 and requires entering numerous experimental data of the charge.¹⁹ Also problematic is the provision of the optimal time of resin cure; the resin has to fill the required area of the preform, which requires the least possible viscosity, and next – to harden as fast as possible.²¹ In order to “steer” the properties of the resin, various modifications are applied: ion crosslinking,²⁵ the addition of: nanotubes,^26,27 silica,^28,29 halloysite,³⁰ nanoclay,³¹ and the introduction of biodegradable components.³² The problem of process designing and the difficulties in acquiring a good repeatability also concerns metal matrix composites,^33–36 especially the technology applying ceramic framed reinforcements and continuous fibers.^37–39 In the manufacture of very small products and those of small sizes, the VARI technique is not the best solution, as it is connected with significant investments. However, it is indispensible in the case of bigger series, especially of products with a large surface area, such as yacht hulls. With the correctly designed process, the saturation of the hull of a 30-meter unit takes 40 min.^1,2,40 Even including the preparation time, this result is impossible to achieve by means of other methods.

The study presents a simple comparative analysis of laminated boards made by the hand lay-up and the VARI method. The standard percentage deviation of the measurement series results was assumed as the repeatability criterion: of the fiber volume fraction, of the flexural strength and of the Charpy impact strength. Two reinforcement-type laminates were produced and compared: plain weave fabric and chopped strand mat.

Materials and methods

Fibrous preforms made of two different structures: chopped strand mat and plain weave fabric, were impregnated with polyester resin, alternatively by hand lay-up and vacuum-assisted resin infusion (VARI). Fourteen layers of the chopped strand mat (150 g/m², vinylsilane finish, emulsion binding agent) by KROSGLASS, Poland, were applied for the first type of preform. Alternatively, 6 layers of the plain weave fabric (350 g/m², vinylsilane finish) by KROSGLASS, Poland, were applied for the second one. ESTROMAL 14 terephthalate polyester resin by ERG PUSTKOW, Poland, was applied as the matrix of the composites. LUPEROX (solution of methylethylketone peroxide in dibutyl phthalate) was used as the catalyst in the amount of 1.5 wt%.

The hand lay-up procedure was implemented with the use of a brush and a roller. The board laminating process was conducted by a person with a significant experience in the use of this method – he has produced about 1000 laminate panels in laboratory conditions, by about 10 years. The amount of the applied resin guaranteed a good saturation of the layers, and its excess was removed by means of the roller, with the avoidance of applying too strong pressure (local “squeeze-out”). The impregnation with the roller was performed for each layer in the “K1” and “K2” direction (Figure 1). After impregnation of the last layer, the laminates remained under ambient conditions (temperature 22℃, humidity 50%) for 20 h. Next, the cured panel was hardened at the temperature of 50℃ for 4 h and left again at room temperature for 20 h.

The VARI processes were conducted on the stand whose scheme is presented in Figure 1. The stand consisted of a flat rigid board-form, a partial vacuum resin transport system and an air-tight elastic shell (foil), which enclosed the working area.

A vacuum system with a TEPRO pump, Poland, power of 1100 W, was applied. The obtained pressure gradient equaled about 0.69 bar in the case of all the conducted processes. The inner diameter of the supplying tubes and the spiral tubes which formed the “resin inlet edge” and the “vent edge”, equaled 10 mm. Just before the preform was filled with the resin (when previously determined amount of the resin was used), the resin inlet was closed. It made possible to distribute the resin excess from the area of the inlet edge. The saturated preform remained under vacuum till its hardening was stated. After the vacuum’s disconnection, the pre-hardened preform was left under foil. The hardening process was conducted in the same way as in the case of the hand lay-up laminates.

All – the hand and the infusion – laminated panels were made during one day, in order to provide repeatability of the ambient conditions. In the case of both types of reinforcement, two panels were prepared for both applied technologies (a total of 8 panels).

The distribution of the samples on the panels (Figure 2) was mainly aimed at a “statistical” elimination of the effect of the resin flow direction on the properties of the infusion laminates.

The samples were cut out of the panels with a circular sawing machine in two stages. First, 8 squares were cut out of each panel, which were weighed in order to determine the volume fraction (16 samples for each technology, a total of 64 samples). Next, each square was cut into two specimens for the impact strength and flexural strength tests (32 samples for each laminate type, for each applied technology; a total of 128 samples for the impact strength tests and 128 for the static bending). The number of specimens for every experiment is given in Table 1.

Table 1.

Number of specimens from all laminate types for individual experiments.

Reinforcement – technology	Fiber volume fraction (specimens from cured laminate)	Bending tests	Impact tests
Fabric – hand lay-up (2 panels)	16	32	32
Fabric – VARI (2 panels)	16	32	32
Mat – hand lay-up (2 panels)	16	32	32
Mat – VARI (2 panels)	16	32	32

VARI: vacuum-assisted resin infusion.

The gravimetric method was used to determine the fiber volume fraction in the manufactured laminates. The measurements were conducted on the fabric (and mat) sheets which were simultaneously used for the manufacturing. They were cut from a central area of the beam (Figure 3).

Preforms were prepared by hemming (each of the layers for the hand lay-up case was hemmed individually and next to this they were stacked). Length and width of individual preforms (and stacks) was measured, and next to this they were weighted on RADWAG scales, accuracy of 0.0001 g. Areal mass was determined. In order to keep original areal mass all fiber strands fallen out of the sheet edges after cutting and before putting to the scales, were collected and weighted with the preforms (stacks). Determined areal mass of individual preforms is presented in Table 2. After saturation, curing and edge-cutting of the laminate panels, 8 samples from each panel (total of 64) were measured and weighted in order to determine areal mass. The average areal mass of individual laminates is given in Table 2 and average fiber volume fraction is presented in Table 4. The areal mass of the preforms was compared with the areal mass of the obtained laminates, according to the following equation

V_{F} = \frac{m_{F}}{m_{L}} \cdot 100

(1)

where V_F is the fiber volume fraction in the laminate, %, m_F, the areal mass of the preform (stack), g/m², and m_L, the areal mass of the cured laminate, g/m².

Table 2.

Areal mass of the preforms and cured laminates used for determination of fiber volume fraction.

Preform (stack)	Preform (stack) areal mass (g/m²)	Average areal mass of cured laminate (g/m²)
Mat – VARI 1	2107.4	3611.0 ± 38.4
Mat – VARI 2	2100.3	3626.2 ± 106.4
Mat – Hand lay-up 1	2108.0	5164.1 ± 479.5
Mat – Hand lay-up 2	2106.1	5549.7 ± 498.7
Fabric – VARI 1	2098.5	3060.4 ± 11.2
Fabric – VARI 2	2095.9	3084.5 ± 69.5
Fabric – Hand lay-up 1	2093.8	3106.5 ± 83.9
Fabric – Hand lay-up 2	2099.1	3120.9 ± 113.7

VARI: vacuum-assisted resin infusion.

Table 4.

Flexural strength of laminates produced by hand lay-up and VARI.

Reinforcement type, fabrication method	Flexural strength – average value (MPa)	σ (MPa)	Cv (%)
Plain weave fabric, hand lay-up	280 (scatter: 171 do 407)	64	23.0
Plain weave fabric, VARI	260 (scatter: 190 do 302)	30	11.7
Chopped strand mat, hand lay-up	171 (scatter: 97 do 248)	35	20.2
Chopped strand mat, VARI	225 (scatter: 191 do 258)	20	8.8

VARI: vacuum-assisted resin infusion.

The fiber volume fraction for individual specimens is presented in Figure 3.

If high accuracy of measurement tasks is kept, gravimetric method will ensure a precise determination of fiber volume fraction in the laminate. Occurrence of significant mistake caused by a local disturbance in areal mass is unlike, because of relatively low area of the preforms (stacks). It was assumed that the fluctuations of the volume fraction within the cured panels are the effect of – more or less uniform – saturation with the resin, which is largely related with the manufacturing technology.

The static three-point bending tests were performed according to the PN-EN-ISO 14125/2001 Standard. A constant ratio of support spacing–specimen thickness was established at L/h = 24. The deformation rate was selected according to the formula given in the Standard, depending on the sample thickness, with the accuracy of 0.5 mm/min.

The Charpy impact strength tests were conducted according to the PN-EN-ISO 179-1/2006 Standard. A constant ratio of support spacing-sample thickness was established at L/h = 17. The sample lengths were at a much approximated level within particular populations, and so they were not corrected. A pendulum of the striking energy of 15 J and the maximum speed of 3.7 m/s was used.

Both in the bending tests and those of impact strength, all the samples were placed on the supports, at their smooth side (reflecting the surface of the board – mold).

The structure of the produced laminates was observed with the use of light microscopes. A stereoscopic microscope Nikon SMZ 1000 and a metallographic microscope Nikon Epiphot 200 were used. A Nikon DS-FI 1 video camera, together with the dedicated software, was applied for the rerecording and acquisition of images.

Analysis of the results

The measurement results for the fiber volume fraction, the flexural strength, and the impact strength are presented in Tables 3 to 5.

Table 3.

Fiber volume fraction determined for laminates produced by hand lay-up and VARI.

Reinforcement type, fabrication method	Fiber volume fraction within laminate – average value (%)	σ (%)	Cv (%)
Plain weave fabric, hand lay-up	67.3 (scatter: 65.2 do 70.0)	1.5	2.2
Plain weave fabric, VARI	68.3 (scatter: 65.8 do 69.5)	1.1	1.6
Chopped strand mat, hand lay-up	39.4 (scatter: 32.1 do 46.5)	3.8	9.6
Chopped strand mat, VARI	58.1 (scatter: 55.8 do 59.8)	1.3	2.2

VARI: vacuum-assisted resin infusion.

Table 5.

Charpy impact strength of laminates produced by hand lay-up and VARI.

Reinforcement type, fabrication method	Impact strength – average value (kJ/m²)	σ (kJ/m²)	Cv (%)
Plain weave fabric, hand lay-up	72.9 (scatter: 56.1 do 94.7)	12.09	16.6
Plain weave fabric, VARI	73.3 (scatter: 59.0 do 89.5)	7.90	10.8
Chopped strand mat, hand lay-up	71.9 (scatter: 58.4 do 90.0)	8.54	11.9
Chopped strand mat, VARI	92.7 (scatter: 84.4 do 98.3)	4.64	5.0

VARI: vacuum-assisted resin infusion.

For the evaluation of the test variation degree, the classic measure of distribution variation was applied, which is the coefficient of variation. It is a relative measure which depends on the arithmetic mean value. The coefficient of variation Cv, expressed in percentage, was calculated according to the following equation

Cv = \frac{σ}{σ_{X}} \cdot 100 %

(2)

where σ is the standard variation of measurement series results (expressed in the physical value of the quantity), σ_X ≠ 0 represents the arithmetic mean value of measurement series results (expressed in the physical value of the quantity).

The fiber volume fraction of the laminates with a fabric base, produced by two different methods, are comparable. It was also stated that, in this case, the coefficient of variation differences are relatively slight (Table 3). However, the coefficient of variation is clearly lower for the VARI laminate, which proves an even distribution of the fraction in the panel. A much bigger difference in the coefficient of variation of the volume fraction is observed in the case of the mat reinforcement. For the hand lay-up laminate, it is over 4 times higher than for the VARI laminate (Table 3). In this case, the volume fraction itself demonstrates a very big difference – it is higher almost by half for the VARI laminate than that for the hand lay-up one.

The flexural strength of the laminates with the fabric base is comparable for the laminates produced by the alternative methods – it is even slightly higher for the hand lay-up laminate. However, the coefficient of variation is over 2 times higher for the hand lay-up laminate (Table 4). This is probably a result of the slightly more uniform distribution of the volume fraction in the board obtained by VARI. An almost 2.5 times higher percentage deviation of the flexural strength is demonstrated by the hand lay-up laminate with the mat base, as compared to the equivalent VARI laminate. The latter also demonstrates a flexural strength itself which is higher by half (Table 4). Undoubtedly, this is connected with the big difference in the volume fraction between the both laminates (Table 3).

The Charpy impact strength is comparable for the laminates with the fabric base, produced by the hand lay-up and VARI methods. The coefficient of variation is higher by over a half for the hand lay-up laminate than for that for the VARI one (Table 5). This is confirmed by the trend of the flexural strength results. An over 2 times higher coefficient of variation of the impact strength results is exhibited by the hand lay-up laminate with the mat base, as compared to the equivalent VARI laminate. The mean impact strength value is almost 1/3 higher in the case of the VARI laminate (Table 5). This means that the big difference in the volume fractions between both laminates (Table 3) has an essential effect both on the impact strength and the flexural strength.

Figures 4 to 6 present the distribution “maps” of the measured values on the schemes of the laminate panels.

In the case of the hand lay-up laminates, both with the fabric and the mat base, one can see a higher volume fraction in the central area of the board than that in the area closer to the edge (Figure 4). This may be connected with the removal of a larger excess of the resin from the central section – the resin is removed with a roller towards the edges and, as possible, outside the laminate. Probably, a consequence of such a “gradient” of the fiber volume fraction is a similar trend demonstrated by the flexural strength and impact strength (Figures 5 and 6).

Also in the case of the VARI laminates, a higher volume fraction, flexural strength and impact strength were observed in the central areas of the boards (Figures 4 to 6). In the case of the applied technique, such a trend is probably caused by the diversified conditions of the vacuum bag’s pressure in the central area and on the edges of the preform. It results from the difference in the resin flow rate in those areas, which has been observed in the VARI processes.⁴¹ However, a clearer, increasing, trend, both in the fiber volume fraction distribution and the flexural and impact strengths of the VARI laminates, is observed in the direction from the resin inlet edge to the vent edge (Figures 1, 4, 5, and 6). This is, undoubtedly, a result of the more thorough “suction” of the resin from the areas closer to the vent than in the case of the further areas. This effect occurs though application of metered amount of resin (only with very little surplus) and closure of the inlet respectively before the resin fulfilled the all preform. In the VARI process, the area filled with the resin, located a few tens of millimeters behind the resin front, is practically devoid of the vacuum bag’s pressure. This has a negative effect on the fiber volume fraction, especially in the areas significantly distant from the vent edge, where one can observe “swelling” of the flowing resin and accumulation of its excess. The difference in the resin’s local volume fraction and the way of filling the preform is significant for the laminate properties.⁴² The observed property gradient is similar for the laminates based on fabric and those based on mat (Figures 4 to 6).

The differences in volume fraction between the hand lay-up and the VARI laminates may arise from the different uniformity of pressure put on the preform during saturation. In hand lay-up process, the excess of resin was removed without applying a strong pressure, but only by local squeezing. While, in VARI process the pressure applied was probably uniform within all preform area (however, pressure changes in each point of the panel – it is different up to the passing of the resin front and after this moment). This difference of pressure uniformity may be the main cause determining the obtained volume fraction distribution.

The obtained results show that the mechanical properties of the laminates, independently on the structure and the manufacture method, strongly depend on the volume fraction of the fibers. It is its value which determines the flexural and impact strength values, whereas its variation in the laminate board determines the variation of those properties. Figure 7 presents linear approximations of the mechanical property dependences on the reinforcement volume fraction for the respective laminates.

The value of R² (correlation coefficient) of the presented approximations (Figure 7) confirm that the effect of the volume fraction on the properties is fundamental. The lowest value of R² (0.86) proves an 86% effect of the reinforcement volume fraction on the impact strength of the fabric laminate made by VARI.

The inaccuracy of the linear model, in the explanation of the effect of the volume fraction on the flexural and impact strengths, results – to some extent – from the stochastic factors, such as technological defects. However, it probably also results from the different than linear, shape of the function describing those dependences. Especially the values of parameters “b” in the approximating functions (Figure 7) point to the fact that, with a lower fiber volume fraction than that obtained within this work, the drop of the mechanical properties together with the drop of the volume fraction will be less intense. In the opposite case, for example, the impact strength of the fabric hand lay-up laminate (Figure 7(a)) would equal 0 kJ/m² by the volume fraction of 58.5%, which is absurd. The presented approximations should thus be treated as a comparative tool only within the range of the analyzed populations.

For the complementation of the image of repeatability, a comparative analysis was performed of the fiber volume fractions within the boards produced by the individual method. The mean values obtained within the particular boards were determined and the differences between those values were compared (Table 6).

Table 6.

Comparison of volume fractions of laminated boards obtained by hand lay-up and VARI.

Reinforcement type, fabrication method	Fiber volume fraction within laminate (panel 1) – average value, %	Fiber volume fraction within laminate (panel 2) – average value, %	Difference in fiber volume fraction between panel 1 and 2, %	Coefficient of variance of fiber volume fraction for panels 1 and 2 (in relation to an average value for panel 1), %
Plain weave fabric, hand lay-up	67.40 ± 1.82	67.26 ± 2.45	0.14	0.21
Plain weave fabric, VARI	68.57 ± 0.25	67.95 ± 1.53	0.62	0.90
Chopped strand mat, hand lay-up	40.82 ± 3.79	37.95 ± 3.41	2.86	7.01
Chopped strand mat, VARI	58.36 ± 0.62	57.92 ± 1.70	0.45	0.77

VARI: vacuum-assisted resin infusion.

A very good repeatability of the boards produced by the VARI method was achieved, both with the fabric and the mat base (coefficient of variation below 1%). The highest repeatability was obtained for the hand lay-up laminate with the fabric base. This value is not, however, significantly different from the results for the equivalent VARI laminate. Yet, this proves that it is possible to achieve a perfect repeatability of the products manufactured by the hand lay-up method, especially in the case of the use of fabric. The hand lay-up laminate with the mat base demonstrated a less advantageous coefficient of variation (over 7%).

Important characteristic concerning quality of the laminates – especially, when a manufacturing technique is under consideration – is their thickness.⁴³ Thickness results obtained for the laminates within the study is presented in Table 7.

Table 7.

Comparison of thickness of the laminates obtained by hand lay-up and VARI.

Reinforcement type, fabrication method	Thickness of the laminate – average value (mm)	Thickness of th laminate – the half of the panel far from the sucking edge (left half in the hand lay-up laminates) (mm)	Thickness of the laminate – the half of the panel close to the sucking edge (right half in the hand lay-up laminates) (mm)	Difference in thickness between the halves of the panels (%)
Plain weave fabric, hand lay-up	2.0 ± 0.1	2.0 ± 0.1	2.0 ± 0.1	1.1
Plain weave fabric, VARI	2.0 ± 0.1	2.0 ± 0.05	1.9 ± 0.05	6.9
Chopped strand mat, hand lay-up	3.9 ± 0.6	3.9 ± 0.6	4.0 ± 0.7	−2.7
Chopped strand mat, VARI	2.6 ± 0.3	2.8 ± 0.05	2.3 ± 0.05	18.0

VARI: vacuum-assisted resin infusion.

Average thickness of the laminates reinforced with the fabric, produced by hand lay-up and VARI, is comparable. On the other hand, the laminates reinforced with the mat showed very big difference – the hand lay-up laminate is on average by half thicker than the VARI one. Such results correspond well with the determined fiber volume fraction values. Performed analysis showed a “gradient” distribution of thickness in the VARI laminates – they are thickest in the area near the inlet edge and the thinnest near the sucking area. In Table 7 much higher percentage differences between thickness of the panel halves are visible for the VARI laminates than for the hand lay-up ones. It confirms the “gradient” trend. Analogous behavior shows the fiber volume fraction. The differences concerning impregnation of individual parts of the VARI laminates are explained in further part of the study. The percentage differences in thickness between the individual panel halves do not correspond well with the analogous differences in fiber volume fraction. It is inconsistent with predictions because thickness is strongly dependent on local resin content and therefore – it should be consequently inversely proportional to fiber volume fraction. Whereas on comparing the values in Figure 4 and in Table 7 the differences in the volume fraction are slight (less than 1%), visibly lower than for the thickness (several percent). Moreover, a correlation between fiber volume fraction and laminate thickness is really low for the VARI laminates (Table 8).

Table 8.

Comparison of Pearson corellation coefficient between thickness of the studied laminates and their fiber volume fraction.

Reinforcement type, fabrication method	Pearson correlation coefficient
Plain weave fabric, hand lay-up	0.816
Plain weave fabric, VARI	0.406
chopped strand mat, hand lay-up	0.785
Chopped strand mat, VARI	0.31

VARI: vacuum-assisted resin infusion.

The mat laminates and and the plain weave fabric laminates showed very similar, not very bad correlation between volume fraction and thickness. However, very low correlation present (in both cases) the VARI laminates. It is probably caused by the already mentioned “gradient” distribution of fiber volume fraction and thickness (Figure 4 and Table 7). In this case the Pearson correlation coefficient is determined from two “parallel” line-shaped sets of points, instead of one. It results in undervaluation of the coefficient.

In order to explain the differences in the tested properties between the laminates and the local variations in the observed trends, a microscopic observation of the structures was performed. Selected images are presented in Figures 8 to 11.

The laminate based on fabric, produced by the VARI method, characterizes in a good quality in the visual evaluation. It has a unifrom transparency, without visible opacities. In a more accurate microscopic analysis, one can clearly see partially closed “chambers” filled with air (Figure 8(a)). They occur in the area right next to the surface, at the crossing of the fiber bands. They are formed as a result of the “collapse” of the gelating resin under the effect of shrinkage. A film of resin remains glued to the bag adhering the surface of the preform and cures on it, thus sealing the “chamber”. Defects of this kind, undoubtedly, have an adverse effect on the laminate mechanical properties; they can constitute an area of crack initiation.^7,44,45 However, the “chambers’ are distributed uniformly on the whole surface of the board, and so they do not locally disturb the material’s properties or the global scatter of results (their negative effect is uniform for the whole panel surface). The cross-section of the structure of the fabric laminate produced by VARI demonstrates a good adhesion of the layers, without any visible undersaturated areas (Figure 10(a)). However, there is a significant difference between the structure of specimens cut from the section of the panel close and distant from the sucking edge – compare Figures 10(a) and 12. Evident reduction of the laminate thickness is visible with approaching to the sucking edge. Moreover, the specimens cut distant from the sucking edge (Figure 12) show “pools” of resin visible at the upper surface of the laminate, in pockets between the fiber strands.

Such “pools” are almost imperceptible in the specimen cut close to the sucking edge (Figure 10(a)). Moreover, the pockets themselves are hardly visible.

The observed differences in structure of the VARI laminates’ structure without the doubt affect the fiber volume fraction and thereby – mechanical properties in various parts of the laminate panels. The volume fraction is probably lower in the areas distant from the sucking edge – flexural strength and impact strength are evidently lower for the specimens cut from these areas (see Figures 5 and 6). The differences in structure have also influence on a distribution of fiber volume fraction through the thickness of the VARI laminate. In the section of the panel distant from the sucking edge, it is probably lower in the area close to the upper layers, where the excess of a resin – caused by under-pressure of the bag – collects (Figure 12). The through the thickness distribution of the volume fraction is rather uniform in the section close to the sucking edge (Figure 10(a)), as well as in the hand lay-up laminates (Figures 10(b) and 11(b)).

The cross-section of the mat-base laminate, produced by the VARI method also characterizes in a compactness of the chambers, and, with the magnification of 20×, reveals only few local undersaturations (Figure 11(a)).

The fabric laminate produced by the hand lay-up method is characterized by a good quality in the visual evaluation. However, contrary to the VARI laminate, it has randomly distributed opacities, proving a non-uniform resin saturation. The microscopic analysis showed no “chambers” with air, characteristic for the infusion laminate (Figure 8(b)). The lack of “chambers” does not mean the lack of surface flaws in the band crossing area; in the hand lay-up laminate, those areas are, simply, “open”. The defects are distributed evenly on the whole surface of the panel, and so their negative effect on the mechanical properties is uniform, analogically to the case of the VARI laminate. The cross-section, which reveals the laminate structure, shows a big number of local defects resulting from undersaturation (Figure 10(b)). The local undersaturation of the laminate is additionally confirmed by a certain number of fiber bands which have evidently come unstuck from the remaining structure, probably while the laminate was being cut with a rotating disk. Such bands are not observed in the case of the VARI laminate (Figure 10(a)). The structure of the hand lay-up laminate with the mat base demonstrates local flaws (Figure 11(b)), which are larger and more numerous than in the case of the infusion laminate (Figure 11(a)). Despite the fact that the difference in the volume fraction of the fibers between the mat-base laminates produced by the two respective methods is big (Table 3), the difference in the structure is not as significant as in the case of the fabric laminates.

The main reason for the differences in the properties of the laminates produced by the two alternative methods and in the scatter of measurement results are probably the structural defects introduced during the hand lay-up lamination. The main types of such defects are: local undersaturations, air bubbles, and surface damages (Figure 9). Such exemplary flaws are stochastic in character and their presence strongly affects the local distribution of the fiber volume fraction and the mechanical properties. It has also an effect on the global scatter of measurement results.

Conclusions

The volume fraction of the plain weave fabric in the laminates manufactured by the hand lay-up and VARI methods is comparable. In the case of the mat reinforcement, the VARI laminate showed a fiber volume fraction by half higher than that of the hand lay-up one.

The flexural and impact strengths of the plain weave fabric-based laminates produced by the hand lay-up and VARI methods are comparable. In the case of the mat reinforcement, the VARI laminate demonstrated a flexural and impact strength 1/3 higher than in the case of the hand lay-up laminate.

The hand lay-up laminated panels show a gradient of the fiber volume fraction and the flexural and impact strengths towards the center of the board. The panels produced by VARI demonstrate a similar gradient and, additionally, a distinct gradient of those properties in the direction of the “sucking edge”.

The coefficient of variation of the fiber volume fraction in the laminate reinforced by plain weave fabric is over 1/3 higher for the population of samples produced by the hand lay-up method than for that produced by VARI. In the case of the mat-reinforced laminates, this difference equals over 4 times.

The coefficient of variation of the flexural and impact strength of the laminates reinforced with the plain weave fabric is, respectively, 2 times and by half higher, for the hand lay-up laminates than for the VARI ones. In the case of the mat-reinforced laminates, the difference in both the flexural and the impact strength equals over 2 times.

The level of the flexural and impact strength is strongly dependent on the fiber volume fraction, both in the global and the local scale. The gradient or the disturbance of the volume fraction result in, approximately proportional, disturbances of the mechanical properties.

In the case of the fabric reinforcement, both for the VARI and the hand lay-up method makes it possible to achieve a good repeatability of the fiber volume fraction in the produced panels. In the case of the mat reinforcement, a much better repeatability is provided by the VARI method (9 time difference in the coefficients of variation).

The repeatable structural defects in the hand lay-up and VARI laminates are distributed uniformly and regularly. They have thus little effect on the repeatability of the mechanical test results.

The random local technological flaws much more often occur in the hand lay-up laminates. They are the main cause of the local disturbances in the mechanical properties of the laminate and negatively affect the scatter of test results.

The VARI laminates as well as the hand lay-up laminates show specific gradiented instability of the structure caused by the pressing forces acting during impregnation process.

The VARI laminates characterize in a distinctly better repeatability with respect to the volume fraction and the mechanical properties, as compared to the hand lay-up ones.

Footnotes

Funding

This study has been financed by the National Centre for Research and Development (NCBiR) of The Republic of Poland in the scope of the project LIDER/08/31/L-2/10/NCBiR/2011.

Conflict of interest

None declared.

References

Reuterlov

. Cost effective infusion of sandwich composites for marine applications. Reinf Plast 2002; 12: 30–34.

Marsh

. Resin film infusion - composites cost reducer. Reinf Plast 2002; 2: 44–49.

Brouwer

Van Herpt

ECFC

Labordus

. Vacuum injection moulding for large structural applications. Compos Part A: Appl Sci Manuf 2003; 34: 551–558.

Lightfoot

Wisnom

Potter

. Defects in woven preforms: Formation mechanisms and the effects of laminate design and layup protocol. Compos Part A: Appl Sci Manuf 2013; 51: 99–107.

Poodts

Minak

Dolcini

. Fe analysis and production experience of a sandwich structure component manufactured by means of vaccum assisted resin infusion process. Compos Part B: Eng 2013. DOI: http://dx.doi.org/10.1016/j.compositesb.2013.04.064.

Seyhan

. A statistical study of fatigue life prediction of fibre reinforced polymer composites. Polym Polym Compos 2011; 9: 717–724.

Koziol

Sleziona

. Failure progress of stitched reinforced polyester-glass laminate at static bending. Polimery 2008; 11-12: 502–508.

Koziol

. Experimental study on the effect of stitch arrangement on mechanical performance of GFRP laminates manufactured on a basis of stitched preforms. J Compos Mater 2012; 9: 1067–1078.

Koziol

. Effect of thread tension on mechanical performance of stitched glass fibre-reinforced polymer laminates – experimental study. J Compos Mater 2013; 16: 1919–1930.

10.

Zimniewska

Myalski

Koziol

. Natural fiber textile structures suitable for composite materials. J Natural Fibers 2012; 4: 229–239.

11.

Bogdan-Wlodek

Koziol

Myalski

. Influence of surface treatment on the wetting process of jute fibres with thermosetting polyester resin. Polish J Chem Technol 2012; 1: 21–27.

12.

Koziol

Bogdan-Wlodek

Myalski

. Influence of wet chemistry treatment on the mechanical performance of natural fibres. Polish J Chem Technol 2011; 4: 21–27.

13.

Koziol

Wieczorek

Bogdan-Wlodek

. Evaluation of jute fiber preforms absorbability using optical profilographometer. J Compos Mater 2013; 47: 2309–2319.

14.

Hufenbach

Gude

Adam

. Experimental investigation of composite-based compliant structures. Kompozyty 2011; 3: 187–191.

15.

Press article: Scottsdale library is first building designed with PPG’s new VARI Cool coatings. Focus on Powder Coatings, Issue 6, 2010.

16.

Hufenbach

Bohm

Blazejewski

. Manufacture, design and testing of piping elements from braided composites. Chem Eng Technol 2005; 7: 808–813.

17.

Ferreira

AJM

Roque

CMC

Neves

AMA

. Buckling and vibration analysis of isotropic and laminated plates by radial basis functions. Compos Part B: Eng 2011; 3: 592–606.

18.

Wen

Aliabadi

. Damage mechanics analysis of plain woven fabric composite micromechanical model for mesh-free simulations. J Compos Mater 2012; 18: 2239–2253.

19.

Steller

. Free volume effect on behavior of polymer liquids in shear flows. Polimery 2012; 7-8: 562–570.

20.

Rieber

Jiang

Deter

. Influence of textile parameters on the in-plane permeability. Compos Part A: Appl Sci Manuf 2013; 52: 89–98.

21.

DeValve

Pitchumani

. Simulation of void formation in liquid composite molding processes. Compos Part A: Appl Sci Manuf 2013; 51: 22–32.

22.

Trochu

Ruiz

Achim

. Advanced numerical simulation of liquid composite molding for process analysis and optimization. Compos Part A: Appl Sci Manuf 2006; 37: 890–902.

23.

Nurzynska

Kubo

Muramoto

. Texture operator for snow particle classification into snowflake and graupel. Atmos Res 2012; 118: 121–132.

24.

Sorrentino

Bellini

Gerevini

. New methodology to determine the compressibility curve in a RIFT process. J Compos Mater 2014; 48(10): 1233–1240.

25.

Maka

Spychaj

. Epoxy resin crosslinked with conventional and deep eutectic ionic liquids. Polimery 2012; 6: 456–462.

26.

Gojny

Nastalczyk

Roslaniec

. Surface modified multi-walled carbon nanotubes in CNT/epoxy-composites. Chem Phys Lett 2003; 370: 820–824.

27.

Gojny

Schulte

. Functionalisation effect on the thermo mechanical behaviour of multi-wall carbon nanotube/epoxy-composites. Compos Sci Technol 2004; 64: 2303–2308.

28.

Nowak

. Experimental and numerical analysis of viscoelastic behavior of epoxy resin with silica filler based on master curve concept. Kompozyty 2012; 1: 3–8.

29.

Faghihian

Nourmoradi

Shokouhi

. Performance of silica aerogels modified with amino functional groups in Pb(II) and Cd(II) removal from aqueous solutions. Polish J Chem Technol 2012; 1: 50–56.

30.

Legocka

Wierzbicka

Al-Zahari

TJM

. Modified halloysite as a filler for epoxy resins. Polish J Chem Technol 2011; 3: 47–52.

31.

Zhu

H-G

Leung

CKY

Kim

J-K

. Tensile properties degradation of GFRP composites containing nanoclay in three different environments. J Compos Mater 2012; 18: 2179–2192.

32.

Grabowska

Bulwan

Zapotoczny

. Biodegradation of new polymer foundry binders composition of poly(acrylic acid)/dextrin. Polimery 2012; 7–8: 529–534.

33.

Dolata-Grosz

Sleziona

Formanek

. Structure and properties of aluminium cast composites strengthened by dispersion phases. J Mater Process Technol 2006; 175(1–3): 192–197.

34.

Dolata-Grosz

Formanek

Sleziona

. Al-FeAl-TiAl-Al2O3 composite with hybrid reinforcement. J Mater Process Technol 2005; 162: 33–38.

35.

Formanek

Jozwiak

Szczucka-Lasota

. Intermetallic alloys with ceramic particles and technological concept for high loaded materials. J Mater Process Technol 2005; 162: 46–51.

36.

Dolata AJ and Dyzia M. Aspects of fabrication aluminium matrix heterophase composites by suspension method. IOP Conference Series-Materials Science and Engineering, Article Number: 012020.

37.

Boczkowska

Chabera

Dolata

. Porous ceramic-metal composites obtained by infiltration methods. Metalurgija 2013; 52(3): 345–348.

38.

Dyzia

Dolata

Sleziona

. Preliminary analysis of aluminum matrix compositions for composites reinforcement with carbon fibers. Steel Res Int 2012; 83(10): 981–987.

39.

Sleziona

Dyzia

Myalski

. The structure and properties of sinters produced from composite powders Al-Al2O3-Al3Fe-Al3Ti. J Mater Process Technol 2005; 162: 127–130.

40.

Press article: Simulation software developed for VARI processing. Reinforced Plastics, September 2004.

41.

Koziol

Rydarowski

Myalski

. Evaluation of vacuum resin infusion process on selected glass fibre preforms. Kompozyty 2012; 4: 285–290.

42.

Brauner

Block

Purol

. Microlevel manufacturing process simulation of carbon fiber/epoxy composites to analyze the effect of chemical and thermal induced residual stresses. J Compos Mater 2012; 17: 2123–2143.

43.

Feichtinger K, Ma W and Elkin R. Properties of structural sandwich core materials: Hand lay-up vs. vacuum-infusion processing. In: Proceedings of “composites 2006 convention and trade show”, American Composites Manufacturers Association, St. Louis, MO, USA, 18–20 October 2006.

44.

Krasnikovs

Varna

. Transverse cracks in cross-ply laminates. 1. Stress analysis. Mech Compos Mater 1997; 6: 565–582.

45.

Praveen

Reddy

. Transverse matrix cracks in cross-ply laminates: Stress transfer, stiffness reduction and crack opening profiles. Acta Mechanica 1998; 3–4: 227–248.