Professional Documents
Culture Documents
Waste Management
journal homepage: www.elsevier.com/locate/wasman
a r t i c l e i n f o a b s t r a c t
Article history: With an increase in renewable wind energy via turbines, an underlying problem of the turbine blade dis-
Received 28 September 2017 posal is looming in many areas of the world. These wind turbine blades are predominately a mixture of
Revised 19 February 2018 glass fiber composites (GFCs) and wood and currently have not found an economically viable recycling
Accepted 24 February 2018
pathway. This work investigates a series of second generation composites fabricated using recycled wind
Available online 2 March 2018
turbine material and a polyurethane adhesive. The recycled material was first comminuted via a
hammer-mill through a range of varying screen sizes, resinated and compressed to a final thickness.
Keywords:
The refined particle size, moisture content and resin content were assessed for their influence on the
Recycling
Wind turbine blade
properties of recycled composites. Static bending, internal bond and water sorption properties were
Polymer-matrix composite obtained for all composites panels. Overall improvement of mechanical properties correlated with
Glass fiber increase in resin content, moisture content, and particle size. The current investigation demonstrates that
it is feasible and promising to recycle the wind turbine blade to fabricate value-added high-performance
composite.
Ó 2018 Elsevier Ltd. All rights reserved.
1. Introduction limitations that are quite challenging (Beauson et al., 2013; Jongert
et al., 2016) and costly. Options to incinerate or disposal has neg-
The demand for wind and other forms of clean energy is ative environmental contributions (Wackernagel and Rees, 1996;
increasing in the US and throughout the world. Wind energy is also Sikdar, 2003; World Commission on Environment and
expected to provide 14.9% of the global electricity demand by 2020 Development, 1987) and disregards the potential rWTBs have as
(Liu and Barlow, 2017). Under this scenario, a significant amount of a feedstock for second generation products. There is a potential
wind turbine blades will continue to burden our current landfills to recycle the glass fiber from the thermoset matrix via chemical
until a viable recycling strategy is found. Repurposing or recycling or thermal treatments (Phoenix Fiberglass Inc, 1994), however
of end-of-use wind turbine blade material will provide both eco- the resulting fiber is often lower in mechanical properties than vir-
nomic and environmental attributes. gin fibers and are very difficult to disperse into many matrices due
While some components of the wind turbine are recyclable to their crimped and entangled form. Mechanical techniques that
(such as the metal parts in the tower and gearbox) (Martínez employ shredders, hammer-mills, knife-mills, etc. provide a low-
et al., 2009; Schleisner, 2000), recycling the blades has been diffi- cost option to deliver a reliable feedstock (Phoenix Fiberglass Inc,
cult due primarily to the thermoset binder used in the synthetic 1994; Bledzki and Goracy, 1993; Petterson and Nilsson, 1994).
fiber (primary glass) composite (The Europen Commission for the However, the methods and classification procedures within the
Environment, 2016). In most WTBs, approximately two-thirds of mechanical refining process are vital parameters to address to
the total weight of the blade is made of GFC (Papadakis et al., achieve the maximum potential of the rWTB materials while main-
2010). Most GFC-based WTBs have a predicted life expectancy of taining minimal energy and costs.
20 to 25 years (Beauson et al., 2016). Estimations predict that the Much research with recycled synthetic fiber composites has
amount of end-of-life WTB materials will account for 100,000 tons been targeted to liberate the glass fiber from the composite matrix
per year in Europe in 2030 (Larsen, 2009; Marsh, 2017). (Kennerley et al., 1998; Pickering et al., 2000) or reuse of shredded
Furthermore, reusing the blades is not recommended due to design composites (SC) in new thermoset polymer composites in order to
reduce the amount of virgin glass fiber in existing composite
(Palmer, 2009; Jutte et al., 1991; Palmer et al., 2009; Derosa
⇑ Corresponding author. et al., 2004; Method of Determining and Expressing Fineness of
E-mail address: s.mamanpush@wsu.edu (S.H. Mamanpush).
https://doi.org/10.1016/j.wasman.2018.02.050
0956-053X/Ó 2018 Elsevier Ltd. All rights reserved.
S.H. Mamanpush et al. / Waste Management 76 (2018) 708–714 709
Nomenclature
Feed Materials by Sieving; Derosa et al., 2005; Silva et al., 2012). posite and BW (genus Ochroma) were separated and run indepen-
However, just a few studies considered the SC as the predominate dently along with a sample of the unseparated rWTB. The
component (above 90%) within the composite (Kouparitsas et al., specimens were heated from ambient to 800 °C under nitrogen at
2002; Zheng et al., 2009; Dangtungee et al., 2012; Yazdanbakhsh the heating rate of 20 °C min 1.
et al., 2018).
The work within this paper investigated the development of a
2.3. Manufacturing of rWTB composites
composite panel derived from mechanically refined rWTB using a
thermoset adhesive as a binder. The second generation composite
The various size fractions of the rWTB materials were sprayed
was evaluated to determine the influence of mill screen size (MSS),
with resin and water (to obtain the targeted MC) within a drum
moisture content (MC), density, and resin level had on the mechan-
blender according to Table 1. The blended rWTB was then hand-
ical and physical properties.
formed and hot pressed to a size of 355:6 355:6 mm composites
panels (duplicate) with a thickness of 7.62 mm. The hot press tem-
2. Materials and method perature and time were set as 138 °C and 5 min accordingly, typi-
cal for pMDI composite processing (Gurke, June 2002) (Fig. 2(c)).
2.1. Materials Mechanical and physical tests (Flexural, IB, water sorption, and
thickness swell) were performed based on ASTM D1037-12 and
Recycled wind turbine blade (rWTB) material supplied by Glo- compared with ANSI 208.1-2009. One-way CoANOVA was per-
bal Fiberglass Solutions at an incoming MC of 1.25%. A polymeric formed to determine the differences of mechanical properties of
methyl-diisocyanate (Rubinate 1840) (pMDI) resin was kindly sup- fabricated panels. The actual density was used as a covariate to
plied by Huntsman and was used as the binder for the second gen- eliminate and effects from the final specimen density. The confi-
eration panel product. The rWTB material was then hammer- dence level was selected as 95%. Duncan grouping was used to
milled through 12.7, 6.35, 3.18, and 1.59 mm screen size, respec- identify statistically significance between the means.
tively. Particle size distribution of the hammer-milled was per-
formed, which is illustrate in Fig. 1.
3. Results and discussion
TGA was carried out to determine the thermal stability of the The thermal degradation profiles of BW, GFC, and rWTB (mix-
rWTB material and to also provide an estimate of the composition ture of BW and glass fiber composite) material by TGA reveal that
of materials within the rWTB. For this analysis the glass fiber com- most of the degradation events occur between 300 °C and 450 °C
40
35
30
25
Weight (%)
12.7 mm MSS
20
6.35 mm MSS
15
3.18 mm MSS
10
1.59 mm MSS
5
0
4 8 10 20 30 40 50 70 100 pan
Size (Mesh)
Fig. 1. Particle size distribution from varying MSS (Method of determining and expressing fineness of feed materials by sieving).
710 S.H. Mamanpush et al. / Waste Management 76 (2018) 708–714
Table 1
rWTB composites formulation (MDI, MC percent by weight, MSS and density).
Sample Resin content (%) Targeted MC (%) MSS (mm) Target density (g/cm3)
1 6 5 3.18 1.04
2 6 5 6.35 1.04
3 6 5 12.7 1.04
4 3 5 12.7 1.04
5 10 5 12.7 1.04
6 6 3 12.7 1.04
7 6 8 12.7 1.04
8 6 5 3.18 1.12
9 6 5 3.18 0.88
10 6 5 1.59 1.04
100 6 MOE and is equal to 5254 MPa and the rWTB composites with
90
5
1.59 mm MSS, 6MDI and 5 MC has the minimum amount of MOE
Weigth Loss (%)
80 equal to 2910 MPa. This is because with increasing the resin level,
DTG (%/ oC)
70 4
bonding capacity increased and it causes better mechanical prop-
60 3
50 erties, also because MOE is correlated to the fiber length
40 2 (Boquillon et al., 2004) by reducing the MSS, MOE reduced
30 1 significantly.
20 Similar to MOE, rWTB composites with 10MDI and 5MC has the
0
10
0 -1 maximum amount for both of the MOR and IB that are 41.6 MPa
0 200 400 600 800 and 2.35 MPa respectively. The rWTB composites with 1.59 mm
Temperature MSS has the minimum amount of MOR equal to 23.8 MPa and com-
rWTB Glass Fiber Composite wood posite with 3MDI, 5MC and 12.7 mm MSS has the minimum
amount of IB equal to 0.82 MPa.
1
Fig. 3. TGA curve recorded in nitrogen at heating rate of 20 °C min . According to the obtained results MSS also has statistically sig-
nificant influence on the MOE and IB but it had no statistically
influence on the MOR. The results for IB test shows that by increas-
for all specimen types. At 800 °C the TGA results of GFC shows the
ing the MSS, IB statistically reduced considerably. The best results
residue of 60.61% that remained, this can be assigned to the fiber-
for IB is for 1.59 mm MSS.
glass content of the sample. Also TGA of BW shows 5.24% remain-
Decreasing the MOE by using smaller MSS is likely due to the
der, attributed to ash in the wood. TGA result of rWTB material
reduction in fiber or particle length because MOE is more linked
shows that 47.19% total loss which was contributed to thermoset
to the particle shape (Boquillon et al., 2004). On the other hand,
resin, non-ash wood material, and paints and other coatings on
by using smaller MSS the particle size is decreased which allows
the wind turbine blade. The resulting 52.81% residue was assigned
for more or improved packing efficiency that creates more bonding
as the content of glass fiber in the rWTB mixture (Fig. 3).
sites within the composite. This is shown within the IB results that
assess the composite bonding performance.
3.2. Mechanical properties of rWTB composites Comparing the results of rWTB composites with the grading
requirements of wood-base particleboard (Composite panel
For evaluating the mechanical properties of rWTB composites, association, ANSI A208.1-2009, 2009) indicate that rWTB compos-
flexure and IB tests were used. The results of these tests are shown ites have higher mechanical properties as shown in Fig. 5. The MOE
in Figs. 4–6. Fig. 4 shows the results of MOE, MOR and IB with (MPa) of rWTB composites with MSS of 3.18, 6.35 and 12.7 mm are
changing MDI (%), MC (%) and MSS (mm). almost twice the wood-base particleboard ANSI H-2 requirements.
According to the obtained results, resin content has a statisti- The composite stiffness increased when the size of the particles
cally significant influence on MOE, MOR as well as IB. This is an increased (Fig. 5), however the IB properties decreased with
expected result as the more resin within the system creates a more increasing particle size (Fig. 6).
bonding capacity and improved properties. The results also indi- Considering the effect of density on the mechanical properties
cated that MC (%) had no significant influence on the mechanical of composites indicates that increasing density resulted in higher
properties, however a slight increase of MOE, MOR and IB with mechanical properties. However, once a density of 1.04 g/cm3
increasing the MC was observed. was obtained little or no improvement in mechanical properties
By reviewing the obtained results, the rWTB composites with was observed.
10MDI, 5MC and 12.7 mm MSS has the maximum amount of
S.H. Mamanpush et al. / Waste Management 76 (2018) 708–714 711
60 6
50 A 5
B A
40 4
MOE (GPa)
MOR (MPa)
30 C 3
B
20 C 2 MOR vs MDI
MOE vs MDI
10 1
0 0
3 6 10
MDI (%) (MC=5%)
(a)
60 6
50 5
40 4
MOE (GPa)
MOR (MPa)
30 3
MOR vs MC
20 2
MOE vs MC
10 1
0 0
3 5 8
MC (%) (MDI= 6%)
(b)
MC (%)
3 5 8
3
2.5
A
2
IB (MPa)
IB vs MDI
1.5 B MC= 5%,
MDI=6%,
MSS=12.7mm
IB vs MC
1
C
0.5
0
3 6 10
MDI (%)
(c)
Fig. 4. Mechanical properties of rWTB composites considering (a) MDI (%), (b) MC (%) and (c) internal bond results with varying MDI (%) and MC (%). Note: Letters indicates
Duncan grouping.
3.3. Physical properties composite and is equal to 3.69% and the thickness swelling of
wood-base particleboard after 24 h of immersion is 26.95% main-
3.3.1. Thickness swell taining a similar improvement over traditional particleboard.
Results of thickness swelling are given in Fig. 7. Comparing the The lowest amount of thickness swelling after 2 h of immersion
results of thickness swelling for rWTB composites with traditional was found for the composite with 3.18 mm MSS and density of
wood-base particleboard (reference: 4% PMDI EW (genus Eucalyptus 1.12 g/cm3 (MDI = 6% and MC = 5%) and is equal to 0.11%, also after
(Pan et al., 2007) shows considerably better characteristics of rWTB 24 h of immersion the rWTB composites with 1.59 mm MSS and
composites, where the maximum amount of thickness swelling of density of 1.04 g/cm3 has the minimum amount of swelling equal
rWTB composites (MSS = 12.7 mm, MDI = 3%, MC = 5%) after 2 h of to 0.71%. Comparing the results of these two rWTB composites
immersion is only 1.96% and thickness swelling for the control sam- with the wood-base particleboard shows that thickness swelling
ple (Pan et al., 2007) after 2 h is approximately12.5% This is an 85% of these two composites are 0.88% and 2.63% of wood-base parti-
reduction in dimension change in a water immersion environment cleboard respectively.
over wood-base particleboard. Also, after 24 h of immersion, the With resin levels, we see that after 24 h of immersion, by
maximum thickness swelling of rWTB composites is for the same increasing the amount of MDI from 3% to 10% thickness swelling
712 S.H. Mamanpush et al. / Waste Management 76 (2018) 708–714
40 12
35
10
30
A
A A 8
25
MOR (MPa)
MOE (GPa)
A
20 6
MOR vs MSS
15
AB A 4 MOE vs MSS
C BC
10
2
5
0 0
1.59 3.18 6.35 12.7
MSS (mm)
Fig. 5. Comparing MOE (MPa) and MOR (MPa) of various MSS rWTB Composites with wood-base particleboards. Note: Letters indicate Duncan groupings and horizontal line
indicate the requirement of H2 grade (Composite panel association, ANSI A208.1-, 2009).
60 5
4.5
50
4
3.5
40
A A
MOR (Mpa)
MOE (GPa)
3
30 2.5
A MOR vs Density
B A 2
MOE vs Density
20
1.5
B
1
10
0.5
0 0
0.88 1.04 1.12
Density (g/cm3)
MSS (mm)
1.59 3.18 6.35 12.7
3.5
3 A
BA A
2.5 A
BC
IB (MPa)
2 C IB vs Density
1.5 IB vs MSS
B
1
0.5
0
0.88 1.04 1.12
Density (g/cm3)
Fig. 6. Mechanical properties of rWTB composites considering MSS and density. Note: Letters indicates Duncan grouping.
reduced from 3.7% to 1.9%, indicating the role MDI (%) plays on the swelled more compared to the rWTB composites with densities
thickness swelling behavior of rWTB composites. Evaluating the of 1.12 and 0.88 g/cm3. but after 24 h of immersion this compos-
influence of MC on the thickness swelling shows that after 24 h ite swell less.
of immersion there is no considerable difference between the The influences of MSS showed that after 24 h of immersion,
thickness swelling of rWTB composites with MC of 3% and 5%, thickness swelling reduced when the smallest MSS material was
but by increasing the MC to 8% the thickness swelling increased used. Thickness swelling is 1.9% for rWTB composites with 12.7
slightly in the 2 h immersion data. mm MSS, but for the rWTB composites with 1.59 mm MSS thick-
Investigating the influence of density on the thickness swel- ness swelling is reduced to 0.71%. These results correlate quite well
ling of rWTB composites showed no consistent trend. After 2 h with the IB data, where smaller particles generated a higher pro-
of immersion rWTB composites with a density of 1.04 g/cm3 duct integrity.
S.H. Mamanpush et al. / Waste Management 76 (2018) 708–714 713
3 3
2 2
1
1
0
0 2H 24 H
2H 24 H
3 5 8
3 6 10 MC (%)
MDI (%)
(MDI=6%, MSS=12.7 mm, Density=1.04 g/cm3)
(MC=5%, MSS=12.7 mm, Density=1.04 g/cm3)
(a) (b)
Thickness Swelling (%)
2.5
0 0
2H 24 H 2H 24 H
9 9
Water Sorption (%)
8 8
7 7
6 6
5 5
4 4
3 3
2 2
1 1
0 0
2H 24 H 2H 24 H
3 6 10 3 5 8
6 6
4 4
2 2
0 0
2H 24 H 2H 24 H
12.7 6,35 3.18 1.59 0.88 1.04 1.12
MSS (mm) Density (g/cm3)
(MDI=6%, MC=5%, Density=1.04 g/cm3) (MDI=6%, MC=5%, MSS= 3.18mm)
(c) (d)
Fig. 8. Water sorption of rWTB composites considering (a) MDI (%) (b) MC (%) (c) MSS and (d) density influences.
714 S.H. Mamanpush et al. / Waste Management 76 (2018) 708–714