The Impact of Wood Fibers in Composite Panels Made from Recycled Fiberglass Wind Turbine Blades

The utilization of wind power has increased almost exponentially during the last decade, leading to the growth in both the number and size of wind turbines. Modern turbine blades are typically made of fiberglass sandwiched around a balsa wood (BW) core. With an increase in the use of wind energy a collateral issue of what to do with the large and voluminous wind turbine blades (WTB) that reached end-of-life has arisen. Through previous research, we developed a recycling process that will keep these turbine blades out of the landfill by considering recycled wind turbine blades as a feedstock for second generation composite panels. The objective of this research is to further the research by understanding the effect that different compositions of wood and fiberglass have on the properties of these panels. Through testing, it was found that the addition of balsa wood increases the modulus of rupture (MOR) and the modulus of elasticity (MOE). The internal bonding (IB) strength values are well above that of typical particle board. All panels tested were visibly more resistant to water sorption than typical particle board. The thermal conductivity for all samples was within the typical range for particleboard. Fire test result indicates that the addition of balsa wood increases the heat and visible smoke release rate of composites.


Introduction
Over the past decade, there has been a concerted effort to increase the amount of energy generated by renewable resources [1]. Wind power is rapidly becoming one of the most efficient methods for this renewable energy generation [2] and it is estimated that 50% of the energy consumed in Europe will come from wind power by the year 2050 [1]. The increase in generation is possible due to the addition of more wind farms, and technological advances that allow for the production of larger wind turbines [3]. The largest turbines today have blades up to 85 m long [1] and an expected life span of 20-25 years [3].
While many components of the wind turbines can be easily recycled, the turbine blades are not as easy to recycle [1]. Typical wind turbine blades are made up mostly of E-glass fibers bound around a core of BW with a strong thermoset matrix [1]. The turbine blades used in this research all have a balsa wood core. Currently, the most common disposal method when the blades wear out is to dump them into a landfill [3]. The increased dependence on wind energy means that estimated worldwide wind turbine waste will reach 225,000 tons annually within the next 20 years [3] making it vital to develop a robust recycling solution before the waste levels become unmanageable.
Options to incinerate or disposal have negative environmental contributions and disregards the potential rWTBs have as a feedstock for second generation products [4]. Thermal and chemical treatments for recycling glass fiber from the thermoset matrix are two common options, however the resulting fiber has significantly lower mechanical properties than the virgin fibers [5] and are not economically viable recycling pathways for damaged or endof-life WTBs.
Through previous research, some researchers have settled on a mechanical breakdown method, where the fiberglass is mechanically ground into small pieces called shredded composites (SC) that consist of fibers with varying lengths, original matrix material, and fiber clusters [6]. Clustering occurs when residual matrix material causes small fibers to stick to each other and ball up [1,6]. Composite panels are made by combining SCs with resin and water and hot pressing them into composite panels that are very similar to particle board.
Through previous research, the development of a composite panel derived from mechanically refined rWTB was investigated using a thermoset adhesive as a binder [4,7,8] and the development of an extruded composite derived from mechanically hammer-milled recycled WTB using a thermoplastic matrix [9][10][11]. The objective of this research is to help further the development of the wind turbine blade recycling process by understanding what impact different compositions of balsa wood and fiberglass SCs have on the properties of the composite panels. It takes extra work to separate out the wood from the fiberglass so to streamline the recycling process it would be ideal to add both to the mill at the same time. For this to happen it is vital to know what impact, if any, the wood SCs will have on the properties of the material.

Materials
Recycled wind turbine blade material was supplied by Global Fiberglass Solutions Inc. All the wind turbine blades sent by GFS for this research have a balsa wood core. The wood was manually separated from the GFC. The moisture content (MC) for the BW was determined to be 4.90% and the glass fiber composites (GFC) was 1.47%. A polymeric methyl-diisocyanate (Rubinate 1840) (pMDI) resin was supplied by Huntsman and was used as the binder for the second-generation panel product. Both the BW and GFC were then hammer milled through 3.2 mm mill screen size (MSS) (Fig. 1). Particle size distribution of the hammermilled materials was performed, which is illustrate in Fig. 2. The GFC did have a larger amount of tiny (100 mesh and smaller) particles. These particles are so small that they are detrimental to the strength of the panels, so a method for screening these should be researched in the future. It is also important to note that a large number of fiber aggregates were observed in the fiberglass sections, particularly in the 20 and 40 sieve screen sizes (SSS).

Thermal Analysis
TGA was carried out to determine the thermal stability of the materials and to also provide an estimate of the composition of materials within the GFC. For this analysis the GFC and BW (genus Ochroma) were separated and run independently. The specimens were heated from ambient to 800 °C under nitrogen at the heating rate of 20 °C min −1 .

Manufacturing of Second-Generation Composites
The various combination of wood/GFC materials were sprayed with pMDI resin and water (to obtain the targeted MC and resin content [4]) within a drum blender. The blended material was then hand-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 temperature and time were set as 138 °C and 5 min accordingly, typical for pMDI composite processing [4].
Mechanical and physical tests (Flexural, IB, water sorption, and thickness swell) were performed based on ASTM D1037-12. One panel was cut into 25.4 × 25.4 mm squares for thermal conductivity testing based on ASTM E 1225-4. The thermal conductivity testing was done on a Laser-Comp Fox 304 Heat Flow Meter. The meter was connected to a Julabo F32 Refrigerated/Heating Circulator set to 4.0 °C to help get the Fox 304 to the low temperatures required. The system was set to run through eight steps, with the bottom plate running from − 20 to 50 °C in increasing increments of 10 °C. The upper plate was set to run from 5 to 75 °C, also in increments of 10 °C. [12].

Fire Resistance Analysis
Fire test was performed in accordance with ASTM E 1354-17, which is the standard test method for heat and visible smoke release rates for materials and products using an oxygen consumption calorimeter [13]. Specimens in the test are burned in ambient air conditions, while being subjected to a predetermined external heat flux. Testing was done at 50 kW/m 2 with spark ignition [13]. Samples from composite panels with 0, 10 and 20% of wood content are considered for fire resistancy test.

Results and Discussion
Density is a vital characteristic that must be analyzed because it can be seen that density is an important factor which can affect the physical and mechanical properties of composite panels therefore the average density has to remain constant [4]. The density of each of the samples was determined and the results indicate that all eight panels are very close to the target density of 1.04 g/cm 3 . Sieve size (mm) 1 3

Thermogravimetric Analysis (TGA)
The thermal degradation profiles of BW and GFC material by TGA reveal that most of the degradation events occur between 300 and 450 °C for all specimen types. At 800 °C the TGA results of GFC shows that a residue of 60.61% remained, which can be attributed to the fiber glass content of the sample. Also, TGA of BW shows a 5.24% remainder, attributed to ash from the wood [4] (Fig. 3).

Thermal Conductivity
The results of the thermal conductivity test can be seen in Fig. 4. There was very little change in thermal conductivity across the temperature range, with all samples having values around 0.10 W/mK. According to the U.S wood handbook the standard thermal conductivity of particle board is between 0.10 and 0.14 W/mK, depending on the thickness of the board [14]. The panels that were created are at the lower end of this range but not nearly as low as current fiberglass insulation. Figure 5 shows the average maximum IB strength for each of the eight different compositions. The samples between 0 and 15% wood all tested as expected, with the sample breaking through the center. The remaining samples (20% thru 50% wood) all experienced surface fractures. On almost all of these samples a thin layer of material was stuck to the surface of one side of the testing block. Two possible reasons for this are that the glue failed before the internal bond strength of the material, or that the glue used to bond the specimens to the testing block was unable to penetrate far enough into the samples due to their composition and high density. Additional testing is needed with a stronger adhesive to determine the true IB of these materials. As such, the values provided cannot be taken at their face value, they might actually be much higher than what is seen. Even so, the IB strength values recorded for these samples are all above 2.06 MPa, which is about 4x higher than the average IB strength for typical particle board [14]. The average MOR and MOE across all eight samples is shown in Fig. 6. Figure 6a shows that the modulus of rupture increases with increasing wood content. The MOR of typical particle board is around 18 MPa [14]. The samples with 0-15% wood have a MOR around this typical value and the remaining samples (20-50% wood) have a higher MOR, with the largest being the 50% wood sample that has a MOR of almost 27.57 MPa. This indicates that increasing the wood content gives the material a higher overall strength, but even the samples will little to no wood are comparable with typical particle board panels [14]. The MOE also shows an increase as the amount of wood goes up, but the change isn't as significant as in the MOR. Typical particle board has a MOE around 2.75 GPa [14]. Again, we see the panels with 0-15% wood have values comparable with this value, and the panels with higher wood content have higher MOE values. When taken together, the MOR and MOE show that adding wood increases the strength and elasticity of the material. Figure 7a shows the change in thickness by percentage. This measurement shows an obvious correlation between the amount of wood present and the amount of swelling in each. The samples with low wood content experienced less than 1.0% swelling even after being submerged for 4 weeks. Standard particle board swells as much as 15% over the first 24 h [14], so even the 50% wood content samples experience notably less swelling than the standard. One possible concern for using these materials in a wet environment would be mildew and mold. The pure fiberglass samples aren't noticeably impacted but the samples with high wood content were The results for the water sorption test can be seen in Fig. 7b. Standard particle board sorbs as much as 17.0% water over the first 2 h [15], the maximum amount of water sorption of samples after 2 h of immersion is 1.4%, resulting in a 91% water sorption reduction compared to wood-base particleboard. By increasing the wood content from 40 to 50%, the swelling and water sorption increased significantly due to high cellulose content in the particleboard that absorb the moisture owing to its hydrophilic nature [15].

Fire Resistance Results
The primary results from the cone calorimeter test are the curves for mass loss, heat release rate and smoke obscuration as a function of time. For reporting purposes, these curves were reduced to single numbers via individual results such as the recorded peaks or calculated averages (Table 1) [16]. Heat flux and surface area for this test was 50 KW/m 2 and 88.4 cm 2 respectively. All specimens burned with bright orange flames and black smoke. There are no pass/fail criteria with ASTM E1354-2017 standard. Comparing the second-generation results with wood-base particleboard [16] confirms that the flame and smoke test results is in the acceptable range.
Fire test result indicates that the addition of balsa wood increases the heat and visible smoke release rate of composites. Also increasing the BW (%) causes to the t(ig) and MLR to increase.

Conclusion
The addition of balsa wood to recycled fiberglass has a considerable impact on the mechanical properties of the material. The higher the wood content, the higher the MOR and MOE, which indicates that the wood adds some strength and elasticity to the panels. The IB strength of all panels is considerably higher than that of normal particle board, and it is possible that the higher wood content increases the IB strength, but due to the types of fractures seen in testing, this cannot be proven without additional testing using a higher bond strength adhesive. The panels were also notably more resistant to water sorption with less shrinkage and weight gain when compared to typical particle board. The results of the thermal conductivity test show that this material is not visibly more insulative than typical particle board.