Evaluation of carbon fiber-embedded 3D printed structures for strengthening and structural-health monitoring

Accepted Manuscript Evaluation of carbon fiber-embedded 3D printed structures for strengthening and structural-health monitoring

Xinhua Yao, Congcong Luan, Deming Zhang, Liujian Lan, Jianzhong Fu PII: DOI: Reference:

S0264-1275(16)31387-9 doi: 10.1016/j.matdes.2016.10.078 JMADE 2435

To appear in:

Materials & Design

Received date: Revised date: Accepted date:

23 September 2016 28 October 2016 31 October 2016

Please cite this article as: Xinhua Yao, Congcong Luan, Deming Zhang, Liujian Lan, Jianzhong Fu , Evaluation of carbon fiber-embedded 3D printed structures for strengthening and structural-health monitoring. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Jmade(2016), doi: 10.1016/j.matdes.2016.10.078

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

CE

PT E

D

MA

NU

SC

RI

PT

Evaluation of carbon fiber-embedded 3D printed structures for strengthening and structural-health monitoring Xinhua Yao1,2, Congcong Luan1,2,*, Deming Zhang1,2, Liujian Lan1,2 and Jianzhong Fu1,2 1 The State Key Lab of Fluid Power Transmission and Control, College of Mechanical Engineering, Zhejiang University, Hangzhou 310027, China; 2 Key Laboratory of 3D Printing Process and Equipment of Zhejiang Province, College of Mechanical Engineering, Zhejiang University, Hangzhou 310027, China; *Author to whom correspondence should be addressed; E-Mail: [email protected]; Tel.: +86-15700078620; Fax: +86-571-8795-1145. Abstract: This paper presents a technique for both structural reinforcement and self-monitoring of thermoplastic parts manufactured by fused deposition modeling (FDM). Continuous carbon fiber tows were embedded into FDM printed structures during the printing process, and the strength and piezoresistive behavior of the printed structures were evaluated. The specimens reinforced with carbon fibers have a tensile strength increase of 70% and flexural strength increase of 18.7% compared to non-reinforced specimens. In addition, the slope of fractional change in electric resistance with strain became a good indicator of strain measurement within the elastic region and damage detection in the yield region. Furthermore, lightweight and print duration reductions were achieved by decreasing the fill density while maintaining the structural strength, where up to 26.01% weight reduction and 11.41% print time reductions were achieved without decreasing the tensile strength. Finally, an artificial hand printed by FDM with embedded carbon fibers is discussed as a demonstration of this approach.

AC

Key works: Carbon fiber, strengthening, monitoring, 3D printed structure, reinforcement 1. Introduction: Recently, fused deposition modeling (FDM) has become one of the most popular 3D printing technologies due to its simplicity, low-cost, and the potential applications for the method [1, 2]. However, FDM products still have deficiencies regarding poor mechanical strength due to the inherent nature of thermoplastic resins, which greatly limit industrial applications [3-5]. On the other hand, reductions in material, as a sustainability requirement for industrial applications, is also of significant importance

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

SC

RI

PT

for lightweight and inexpensive printed structures even though it would seem to decrease the strength further. Thus, a novel method, which can increase the strength of the printed structures while reducing material consumption and even provide early warning of damage to the structure, will be valuable to FDM printing technology. Currently, there are mainly four strategies to improve the strength of FDM printed components: 1) structural optimization through the addition of ribs and internal printed supports [6]; 2) optimization of process parameters, such as the print extrusion, temperature, build orientation, raster angle, and contour width [7]; 3) development of new materials, including the feedstock filaments made of metal/polymer composite material [8], thermotropic liquid crystalline polymer (TLCP) fibers [5] and, especially, carbon fiber reinforced thermoplastic filament [9]; and 4) development of new methods/technologies, such as printing of continuous-fiber composites by in-nozzle impregnation [10], filling voids in the printed parts with high-strength resins [11] and adding short carbon fibers to a matrix of cellulose-modified gypsum powder [12]. However, all these approaches, including the new carbon fiber-based materials, only focus on reinforcing the printed structures, few of them aim to reduce the amount of material and none of them consider a self-monitoring function. Carbon fiber, as a potential material to be added to 3D printed structures, shows great promise not only because of its high strength, high modulus, and low density, but also for its piezoresistive behavior [13-15]. The piezoresistive behavior of carbon fiber refers to the change in its electrical resistivity with strain, where the resistivity decreases reversibly upon compression, increases reversibly upon tension, and irreversibly upon damage [16]. This feature is believed to be ideally suited to measure the deformation of structures, such that the application of carbon fibers into composites [17-20] and concretes [21-27] for structural-health monitoring has been reported in recent years. Haentzsche et al. investigated the exPAN carbon filament yarn integrated in a reinforced thermoplastic composite for structural-health monitoring [17, 19]. Additionally, the piezoresistive behavior of self-sensory carbon fiber textile-reinforced concrete beams under the effect of loading, unloading, and reloading in the un-cracked region was investigated by Goldfeld et al. [22]. Chung et al. have made great efforts in the research of carbon fiber reinforced cement and its ability to self-sense damage [26-29]. Thus, the application of carbon fibers in composites and concretes brings us to the idea of using carbon fibers in FDM-based 3D printed structures for structural reinforcement as well as self-monitoring.

ACCEPTED MANUSCRIPT

NU

SC

RI

PT

This paper examines several carbon fiber-based 3D printed structures, in which continuous carbon fibers are embedded into the polylactic acid (PLA) structure during the FDM printing process. In these structures, carbon fibers not only serve as the structural reinforcement, but also serve as sensory agents by monitoring changes in the electrical resistance of the fiber due to deformation. The capability of carbon fibers in 3D printed components to self-monitor and provide structural reinforcement has been well characterized under the uniaxial tension test and three-point bending test. Moreover, lightweight and reduction of print duration can be achieved by reducing the printed material (PLA) fill density without reducing the structural strength by the addition of carbon fibers into the printed structures. A carbon fiber-embedded 3D printing approach provides an effective method to resolve the intrinsic inadequate strength of FDM printed structures, and it also makes smart structure printing possible through the structure’s ability to self-monitor structural health. 2. Material and Instruments

PT E

D

MA

2.1 Material Polyacrylonitrile (PAN)-based continuous carbon fibers of three types TORAY Torayca (T300B-3000-40B, T700SC-6000-50C, and T700SC-12000-50C) were investigated as the reinforcement and structural health-monitoring elements in this study. The properties of the carbon fibers are detailed in Table 1. In the following sections, the number of filaments was used to indicate the types of carbon fibers.

strength

T300B-3000-40B

CE

Table1. Properties of the adopted carbon fibers provided by the TORAY Torayca Company [30].

3k

3.53GPa

230GPa

1.5%

7.0μm

1.76 g/cm3

1.7x10-3Ω.cm

T700SC-6000-50C

6k

4.9GPa

230GPa

2.1%

7.0μm

1.80 g/cm3

1.6x10-3Ω.cm

T700SC-12000-50C

12k

4.9GPa

230GPa

2.1%

7.0μm

1.80 g/cm3

1.6x10-3Ω.cm

Number of

Tensile

Tensile

Tensile

Filament

Electric Density

filaments

modulus

strain

diameter

AC

Types

Resistivity

A curing epoxy resin (DY-E44), used to align the filaments and stabilize the fibers, included two components: E-44 epoxy resin (epoxy value: 0.41~4.47/100g) and low molecular polyamide adhesive (amine value: 200±20, molecular weight: 600–1100). The proportion of the two components was 1:1 and the curing time was 7 hours at a temperature of 40~45°C. The polylactic acid (PLA) with a filament diameter of 1.75± 0.03mm was adopted as the printing material (tensile strength: ≥60MPa, elongation

ACCEPTED MANUSCRIPT rate: ≥3.0%, contraction rate: 0.197%, melting point: 150~160°C, and printing temperature: 190~210°C).

MA

NU

SC

RI

PT

2.2 Specimen fabrication Experimental investigations were performed on standard test PLA specimens fabricated using the commercial Kossel Rostock Delta D-force 3D printer, which was controlled by the “Repetier-Host v1.0.3” software. The 3D model was sliced using the “Cura 15.02.1” software. Two types of specimens were printed, a dog bone-shaped specimen (ISO 527-4:1997, test conditions for isotropic and orthotropic fiber-reinforced plastics composites, NEQ) for a uniaxial tension test and rectangular specimen (ISO 14125: 1998, fiber-reinforced plastic composites, determination of flexural properties, NEQ) for a three-point bending test. Figure 1 shows the dimensions of these two specimens.

D

Fig. 1. Dimensions of the specimens (mm): (a) dog bone-shaped specimen for the tensile test, and (b) rectangular specimen for the flexural test.

AC

CE

PT E

Figure 2 shows the specimen fabrication process. PLA filament was melted by the printer head at a temperature of 205°C and then deposited on a 50°C heated build platform (Fig. 2a). For the carbon fiber preparation, a 250mm length of fiber was impregnated with two-component (1:1) DY-E44 epoxy resin adhesive (Fig. 2b). Then, the impregnated fibers were manually placed on the specified printed layer (2mm above the bottom of the dog bone-shaped specimen and 1mm above the bottom of the rectangular specimen) and to make the carbon fibers achieve a uniform tension, a 2N constant tension force provided by a dynamometer was applied along the carbon fibers during the printing process (Fig. 2c). After that step, the PLA was printed continuously on the laid fiber layer until the end of printing (Fig. 2d). To ensure the quality of printing, the print speed was adjusted to 30% when printing on the laid fibers and then the print speed returned to 100% (35mm/s). Finally, the printed specimens were left to cure the resin for 7 hours at a temperature of 40°C in an incubator. Fig. 2e–h shows the four different PLA fill density interior structures, three types of carbon fibers, the distribution of carbon fibers in the PLA printed structures, and the final fabricated specimens, respectively. In the following sections, all the fill densities mentioned refer

ACCEPTED MANUSCRIPT

RI

PT

to the fill density of the PLA material.

MA

NU

SC

Fig. 2. Fabrication process: (a) PLA print: PLA printed on the 3D printer, (b) carbon fibers pre-preg: carbon fibers impregnated with DY-E44 resin, (c) embedding pre-preg carbon fibers: laying the pre-preg carbon fibers on the PLA printed surface with a 2N tension force, (d) ongoing PLA print: printed PLA on the laid fibers, (e) interior structures of the four different PLA fill density specimens, (f) three used TORAY Torayca carbon fibers, (g) distribution of carbon fibers in the PLA printed structures, and (h) the final fabricated specimens.

AC

CE

PT E

D

2.3 Instrumentation and testing In order to characterize the mechanical and piezoresistive properties of carbon fiber-based 3D printed structures, the uniaxial tension test and the three-point bending test were performed on the aforementioned specimens using the WDW-100 microcomputer control electron universal testing machine and the Electro Force mechanical test instrument, Electro Force 3300, respectively. Each test included four different fill density specimens (20%, 40%, 60%, and 100%) and three different types of carbon fibers at the 20% fill density specimen (3K, 6K, and 12K filaments). In the two tests, the loading speed was controlled at 2mm/min and the force were recorded during the whole load process. The electrical resistance of the specimens was measured by a TH2516 DC resistance instrument controlled by a PC with LabView. The test setups are depicted in Fig. 3. Four specimens (totaling 56) were tested for every type of specimen modification, shown in Table 2.

SC

RI

PT

ACCEPTED MANUSCRIPT

NU

Fig. 3. Illustration of the experimental setups: (a) tensile test setup, and (b) three-point bending

MA

test setup.

Table 2. Specimens for the uniaxial tension test and the three-point bending test. Uniaxial tension test Non

3K

20%

4

4

40%

4

60%

4

100%

4

Total

12K

Non

3K

6K

12K

numbers

4

4

4

4

4

4

32

0

0

4

0

0

0

8

0

0

0

4

0

0

0

8

0

0

0

4

0

0

0

8

CE

0

AC

3. Results

Three-point bending test

6K

PT E

density

D

Fill

3.1 Uniaxial tension test The results of the uniaxial tension test for four 20% fill density specimens (without fibers, with 3K fibers, with 6K fibers, and with 12K fibers) are presented in Fig. 4. By defining 1 

l l0

(1)

where l0 is the gauge length of the specimen, l is the breaking elongation of the specimen (gauge length elongation, measured by using a vernier caliper). The

ACCEPTED MANUSCRIPT

PT E

D

MA

NU

SC

RI

PT

variations of both fractional change in electrical resistance and tensile stress with strain 1 (Fig. 4, except Fig. 4a) illustrate the effect of carbon fibers on reinforcement and self-sensing of 3D printed structures. The results of the uniaxial tension test are also summarized in Table 3 based on 28 specimens. It can be seen that both the tensile strength and the elongation rate at breaking of the specimens are improved when integrated with carbon fibers.

Fig. 4. Fractional change in resistance and stress versus the tensile strain with different types of

AC

CE

carbon fibers at 20% fill density: (a) without fibers, (b) with 3K filaments, (c) with 6K filaments, and (d) with 12K filaments.

ACCEPTED MANUSCRIPT As shown in Table 3, specimens integrated with carbon fibers led to 38.68%, 58.86%, and 70.02% improvements in average tensile strength for the 20% fill density with 3K, 6K, and 12K fibers compared with no fiber specimens, respectively. It is observed that the average tensile strength of the specimen at 20% fill density with 3K fibers is 8.99% higher than the specimen at 60% fill density without carbon fibers, with a 26.01% weight reduction and a 11.41% print time reduction. Combining 3D

Tensile strength

Weight

Print time

20%

19.156±0.710MPa

5.372±0.049g

33min

40%

21.852±0.687MPa

6.443±0.052g

37min

60%

24.375±0.785MPa

7.461±0.051g

41min

100%

46.370±0.895MPa

9.655±0.052g

20%-3K

26.566±0.195MPa

5.923±0.045g

20%-6K

30.432±0.583MPa

5.971±0.057g

20%-12K

32.570±1.797MPa

6.082±0.050g

Elongation rate

Initial resistance

1.88%

--

RI

Specimens

PT

Table 3. Tensile strength, weight, print time, and initial resistance of the specimens for a uniaxial tension test.

--

2.96%

--

60min

3.53%

--

36.8min

2.07%

34.363±0.109Ω

36.8min

2.24%

17.567±0.176Ω

36.8min

2.52%

9.776±0.172Ω

MA

NU

SC

2.61%

AC

CE

PT E

D

printing technology with carbon fiber materials provides a potential solution for lightweight products and print duration reduction. 3.2 Three-point bending test Figure 5 presents the three-point bending results of four 20% fill density specimens (without fibers, with 3K fibers, with 6K fibers, and with 12K fibers). Before specimen buckling, the fractional change in resistance increases almost linearly with flexural strain  2 (Eq. 2), while the growth of fractional change in resistance slows down after buckling. The flexural strain is defined as 2 

6S  h l2

(2)

where S is the deflection at the center of the specimen, h is the thickness of the specimen, and l is the span of the test setup. It can be seen from Table 4 that the specimens integrated with carbon fibers led to 4.66%, 15.61%, and 18.71% improvements in flexural strength for the 20% fill density with 3K, 6K, and 12K fibers, respectively. It is observed that the specimen at 20% fill density with 6K fibers has a similar flexural strength to the 60% fill density without carbon fibers specimen, with a 16.53% weight reduction and 11.11% print time reduction.

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

MA

Fig. 5. Fractional change in resistance and force versus the flexural strain with different types of carbon fibers at 20% fill density: (a) without fibers, (b) with 3K filaments, (c) with 6K filaments,

D

and (d) with 12K filaments.

PT E

Table 4. Flexural strength, weight, print time, and initial resistance of the specimens for the three-point bending test. Flexural strength

Weight

Print time

Initial resistance

20%

57.462±0.437MPa

2.975±0.050g

16 min

--

40%

60.336±1.125MPa

3.438±0.048g

18 min

--

66.427±0.938MPa

3.933±0.057g

20 min

--

82.802±1.285MPa

4.967±0.057g

29 min

--

20%-3K

60.140±1.322MPa

3.225±0.050g

18 min

22.795±0.262Ω

20%-6K

66.432±0.826MPa

3.375±0.050g

18 min

9.989±0.295Ω

20%-12K

68.211±1.643MPa

3.560±0.055g

18 min

4.598±0.116Ω

60%

CE

Specimens

AC

100%

4. Discussion The sensitivity of carbon fibers that are embedded in specimens as deformation sensors is determined by their piezoresistive effects. Figure 6 shows the linear fit of

ACCEPTED MANUSCRIPT the fractional change in resistance versus the tensile strain for the uniaxial tension test and the flexural strain for the three-point bending test. By defining the gauge factor (GF, the slope of fitting line) of these two fitting lines as GF 

R / R0



(3)

AC

CE

PT E

D

MA

NU

SC

RI

PT

where R0 is the initial resistance, R is the relative change in the resistance,  is the tensile strain 1 (Eq. 1) for the uniaxial tension test and the flexural strain  2 (Eq. 2) for the three-point bending test, respectively, it can be found that there are obvious changes of the slope (GF) of each fitting line (Fig. 6). The printed structures were tested within the elastic range when the GF of the fitting lines was smaller for uniaxial tension and bigger for the three-point bending tests. The sudden increase in GF values in the uniaxial tensile tests indicates that the specimens were elongated out of the elastic stage and became yielding; while in the three-point bending tests, the change in GF values means that the deformation specimens came into the buckling stage. This bilinear relationship between the fractional change in electrical resistance and the tensile strain (or the flexural strain) has also been reported by Yeh et al. in their recent work on cementitious composite. [30] Thus, the sensitivity of carbon fibers on GF values can be used as a self-sensing function for strain measurement within the elastic stage (before the proportional limit point A) and damage detection in the yield stage (from the proportional limit point A to yield point B). Figure 7 shows the variation of fractional change in resistance and flexural strain versus time for five load cycles (sine-wave load, ƒ=0.5Hz, A=1.5mm, test setup as shown in Fig. 3b). The correlation between fractional change in resistance and the strain is generally repeatable and the reversibility of the resistance change is mostly reversible though the existence of some irreversible component. The irreversible component is not considered to significantly impair the application of the continuous carbon fiber as strain sensors as reported in Goldfeld et al. [22].

PT

ACCEPTED MANUSCRIPT

RI

Fig. 6. Linear fit of fractional change in resistance in two segments: (a) versus the tensile strain,

PT E

D

MA

NU

SC

and (b) versus the flexural strain.

Fig. 7. Variation of fractional change in resistance and flexural strain versus time for five load cycles.

AC

CE

Table 5 summarizes the gauge factors of three 20% fill density specimens under the uniaxial tension test and the three-bending test in two segments (as shown in Fig. 6) and all values of R-squared (R2), which refer to the goodness of fit and describe how well the linear model here fits the set of testing data, are greater than 0.944 (R2 = 1 means completely fitting). An approximate tenfold mutation increase between the two gauge factors for the uniaxial tension test and a tenfold mutation decrease for the three-point bending test except the 20%-3K1 specimen are observed. The dramatic change in gauge factors can be seen as a distinctive marker for the elastic stage and the yielding stage. Carbon fibers show great potential for working as strain sensors before the dramatic change in gauge factor and as damage detection after the dramatic change in gauge factor. Measures should be taken when the gauge factor of carbon fibers changes dramatically in practical applications to avoid further damage to the structures.

ACCEPTED MANUSCRIPT Table 5. Gauge factors of 20% fill density with 3K fibers under the uniaxial tension test and three-point bending test. Three-point bending test

Range of ε1

Gauge factor

R2

Range of ε2

Gauge factor

R2

[0,1.22]

0.68661

0.99679

[0,1.05]

0.21556

0.97793

[1.69,1.86]

6.64753

0.96505

[1.27,3.39]

0.09245

0.97469

[0,1.36]

0.49924

0.98338

[0,0.91]

0.73932

0.96644

[1.71,1.89]

4.1753

0.97373

[1.03,3.06]

0.08608

0.95419

[0,1.54]

0.54517

0.99132

[0,0.89]

PT

Uniaxial tension test Specimens

0.52824

0.97028

[1.75,1.92]

4.56727

0.97132

[1.11,3.29]

0.06039

0.94455

20%-3K1

20%-3K2

SC

RI

20%-3K3

AC

CE

PT E

D

MA

NU

Carbon fibers embedded in 3D printed structures work as effective structural reinforcement. For example, the tensile strength of a 20% fill density structure integrated with 3K carbon fibers was 38.68% higher than the specimen without carbon fibers, while the flexural strength improved more than 4.66%. One point to be taken into consideration is that the adhesion between the fibers and the thermoplastic resin (PLA) is insufficient, since carbon fibers are separated from the specimens during the tensile tests (Fig. 8a) and flexural tests (Fig. 8b). Figure 8c shows the interface between the fibers and the PLA materials. The adhesion between the fibers and the PLA materials is weak since there could be no or little chemical bonding at the interface. The deformation of carbon fibers and PLA materials is almost simultaneous before the interface is failure, while the deformation of PLA materials is predominant due to the relative slip between carbon fibers and PLA materials interface after the interface is failure. This might help to understand the bilinear relationship that represented in Fig. 6. Further enhancement in the 3D printed structures’ mechanical properties can be achieved by treatment of the fiber surfaces to improve the adhesion between the fibers and the thermoplastic resin.

SC

RI

PT

ACCEPTED MANUSCRIPT

NU

Fig. 8. Image of the fracture surface of the specimens (20% fill density): (a) tensile fracture surface, (b) flexural fracture surface, and (c) interface of carbon fibers and PLA materials.

MA

Lightweight and print duration reduction is expected to be achieved when combining 3D printing technology with carbon fiber materials. As described in Fig. 9,

PT E

D

which calculated the strength–weight ratio of the specimens from the data in Table 3 and Table 4, the specimens with carbon fibers and low fill density show higher

CE

strength per unit of weight than most of those without carbon fibers but with high fill density. That is to say, lightweight can be achieved by decreasing the fill density but

AC

keeping the structural strength in carbon fiber-embedded 3D printed structures. In addition, there was also a 11.41% print time reduction achieved in the uniaxial tensile specimens (calculated by comparing 20% fill density with 3K fibers to 60% fill density without fibers). The print duration can be reduced even more by optimizing the parameters and the lightweight capability can be promoted by improving the adhesion between fibers and the thermoplastic resin.

SC

RI

PT

ACCEPTED MANUSCRIPT

NU

Fig. 9. Strength–weight ratio for specimens from the uniaxial tension and three-point bending tests.

MA

5. Practical Application—Artificial hand

AC

CE

PT E

D

3D printing technology has been applied in the field of prostheses recently and makes possible the building of products for the disabled, such as limb prostheses [31, 32], ankle-foot orthoses [33], robotic running feet and legs [34] and femur bones [35]. Since 3D printed structures integrated with carbon fibers show a great improvement in strength and a capability for self-sensing of structural health, a carbon fiber-based artificial hand was manufactured as a demonstration to evaluate the benefits of using carbon fibers as the structural reinforcement and the structural health sensors in a more practical application. Figure 10 shows an overview of the conception carbon fiber-embedded 3D printed artificial hand. Carbon fibers were placed in five fingers separately during the fabrication process and the ends of the carbon fibers were connected to electrodes composed of a silver paste and enameled wires. Taken the middle finger as an example, the carbon fiber was 40mm long and was placed at 2mm distance from the surface, as shown in Fig. 10b. The resistance of carbon fibers in the fingers will vary with heavy loads applying on the hand and the fractional change in resistance can help to monitor the condition of fingers.

RI

PT

ACCEPTED MANUSCRIPT

SC

Fig. 10. Conception of carbon fiber-embedded 3D printed artificial hand: (a) overall schematic, and (b) placement of carbon fibers in the middle finger.

AC

CE

PT E

D

MA

NU

Experiment was carried out to demonstrate the self-monitoring function of carbon fiber-based artificial hand, as shown in Fig. 11. A step tensile load was applied on the middle finger to simulate the actual situation of carrying heavy loads with the hand (Fig. 11a). The step loading was provided by the universal testing machine in 20N increments stepwise to 220N and stopped every 10 seconds. The change in resistance was recorded by the TH2516 DC resistance instrument at the same time. The fractional change in resistance and tensile force versus the loading time were plotted in Fig. 11c. The result shows that the fractional change in resistance of the middle finger can be a real-time monitor of external loading.

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

MA

Fig. 11. Loading experiment for one of the fingers: (a) experiment setup, (b) 3D printed carbon fiber-based middle finger, and (c) fractional change in resistance under axial step loading.

AC

CE

PT E

D

Figure 12 shows the linear fit of the step curve of fractional change in resistance. An obvious bilinear relationship between the fractional change in resistance and the external loading appeared. The slope changes from 0.00162 to 0.00304, approximate twofold mutation. The mutation of the slope can be considered to be the threshold for structural failure. Thus, protection and prevention of overloading can be achieved by presetting the extreme limit of the fractional change in resistance. The example of an artificial hand demonstrates that 3D printing technology combining carbon materials can produce self-sensing applications, particularly for medical support or the prosthetic industry.

PT

ACCEPTED MANUSCRIPT

RI

Fig. 12. Linear fit of fractional change in resistance versus heavy loads (tensile force).

AC

CE

PT E

D

MA

NU

SC

6. Conclusions This study provides a method to incorporate continuous carbon fibers into 3D printed structures. Based on the uniaxial tension and three-point bending tests, carbon fibers significantly improved the mechanical strength of a 3D printed structure by over 70% in tensile strength and 18.7% in bending strength. Meanwhile, up to 26.01% in weight reduction and 11.41% in print time reduction were achieved without decreasing the tensile strength. Moreover, the correlation between fractional change in resistance and the strain was generally repeatable and the reversibility of the resistance change was mostly reversible within the elastic range of the printed structures. There was an approximate tenfold mutation bilinear relationship of the fractional change in electrical resistance during the deformation of the printed structures. Gauge factors ranged from 0.545 up to 6.686 with all correlation coefficients greater than 0.965 for the uniaxial tension test, and from 0.739 down to 0.060 with all correlation coefficients greater than 0.944 for the three-point bending test. The dramatic change in the gauge factors can be seen as a distinctive marker for the elastic stage and yielding stage. Carbon fibers can work as strain sensors before the dramatic change in the gauge factor and as damage detection after the dramatic change in the gauge factor. Therefore, continuous carbon fibers integrated into 3D printed structures can be used as structural reinforcement and sensory agents simultaneously. Indeed, the idea of combining continuous carbon fibers with 3D printing technology should find wide use in applications where high strength and lightweight structures are required, including the automotive and aerospace industries. Furthermore, the technique has enormous potential for use in a variety of smart structures. In order to improve the reinforcement role of carbon fibers, the adhesion between

ACCEPTED MANUSCRIPT the carbon fibers and the thermoplastic resin has to be further enhanced during the 3D printing process. In addition, the carbon fiber placement and self-monitoring strategies should be studied for complicated structures in the future.

AC

CE

PT E

D

MA

NU

SC

RI

PT

Acknowledgments This paper was supported by the Science Fund for Creative Research Groups of the National Natural Science Foundation of China (NO. 51521064), Zhejiang Province Public Projects of China (NO. 2016C31036) and the Fundamental Research Funds for the Central Universities of China (NO. 2015QNA4002).

ACCEPTED MANUSCRIPT References [1]

S. Upcraft, "The rapid prototyping technologies," Assembly Automation, vol. 23, pp. 318-330(13), 2003.

[2]

O. A. Mohamed, S. H. Masood, and J. L. Bhowmik, "Optimization of fused

PT

deposition modeling process parameters:a review of current research and

V. Schöppner and K. P. KTP, "MECHANICAL PROPERTIES OF FUSED DEPOSITION

SC

[3]

RI

future prospects," Advances in Manufacturing, vol. 3, pp. 42-53, 2015.

MODELING PARTS MANUFACTURED WITH ULTEM* 9085," In Proceedings of

NU

69th Annual Technical Conference of the Society of Plastics Engineers

[4]

MA

(ANTEC'11), vol. 2, pp. 1294–1298, 2011.

C. S. Lee, S. G. Kim, H. J. Kim, and S. H. Ahn, "Measurement of anisotropic

D

compressive strength of rapid prototyping parts," Journal of Materials

[5]

PT E

Processing Technology, vol. s 187–188, pp. 627-630, 2007. D. G. Baird, "Effects of processing conditions on short TLCP fiber reinforced

O. Stava, J. Vanek, B. Benes, N. Carr, and R. Měch, "Stress relief: improving

AC

[6]

CE

FDM parts," Rapid Prototyping Journal, vol. 4, pp. 14-25, 2013.

structural strength of 3D printable objects," ACM Transactions on Graphics (TOG), vol. 31, p. 48, 2012. [7]

M. S. Hossain, J. Ramos, D. Espalin, M. Perez, and R. Wicker, "Improving tensile mechanical properties of FDM-manufactured specimens via modifying build parameters," in International Solid Freeform Fabrication Symposium: An Additive Manufacturing Conference. Austin, TX, 2013, pp. 380-392.

ACCEPTED MANUSCRIPT [8]

S. H. Masood and W. Q. Song, "Development of new metal/polymer materials for rapid tooling using Fused deposition modelling," Materials & Design, vol. 25, pp. 587-594, 2004.

[9]

E. Hull, W. Grove, M. Zhang, X. Song, Z. J. Pei, and W. Cong, "Effects of Process

PT

Variables on Extrusion of Carbon Fiber Reinforced ABS Filament for Additive

RI

Manufacturing," in ASME 2015 International Manufacturing Science and

SC

Engineering Conference, 2015.

[10] R. Matsuzaki, M. Ueda, M. Namiki, T.-K. Jeong, H. Asahara, K. Horiguchi, et al.,

NU

"Three-dimensional printing of continuous-fiber composites by in-nozzle

MA

impregnation," Scientific Reports, vol. 6, 2016. [11] J. T. Belter and A. M. Dollar, "Strengthening of 3D printed fused deposition

PT E

e0122915, 2015.

D

manufactured parts using the fill compositing technique," PloS one, vol. 10, p.

[12] S. Christ, M. Schnabel, E. Vorndran, J. Groll, and U. Gbureck, "Fiber

AC

2015.

CE

reinforcement during 3D printing," Materials Letters, vol. 139, pp. 165-168,

[13] W. X., F. X., and L. D., "Strain sensing using carbon fiber," Journal of Materials Research, vol. 14, pp. 790-802, 1999. [14] Q. C., S. Z., and H. H., "Electrical properties of different types of carbon fiber reinforced plastics (CFRPs) and hybrid CFRPs," Carbon, vol. 45, pp. 3027-3035, 2007. [15] D. D. L. Chung, "Self-monitoring structural materials," Materials Science &

ACCEPTED MANUSCRIPT Engineering R Reports, vol. 22, pp. 57-78, 1998. [16] S. Wen and D. D. L. Chung, "Self-sensing of flexural damage and strain in carbon fiber reinforced cement and effect of embedded steel reinforcing bars," Carbon, vol. 44, pp. 1496-1502, 2006.

PT

[17] E. Häntzsche, A. Matthes, A. Nocke, and C. Cherif, "Characteristics of carbon

RI

fiber based strain sensors for structural-health monitoring of textile-reinforced

SC

thermoplastic composites depending on the textile technological integration process," Sensors & Actuators A Physical, vol. 203, pp. 189-203, 2013.

NU

[18] E. Haentzsche, R. Mueller, T. Ruder, A. Nocke, and C. Cherif, "Integrative

MA

Manufacturing of Textile-Based Sensors for Spatially Resolved Structural Health Monitoring Tasks of Large-Scaled Composite Components," in Materials Science

D

Forum, 2015, pp. 571-578.

PT E

[19] M. M. B. Hasan, A. Matthes, P. Schneider, and C. Cherif, "Application of carbon filament (CF) for structural health monitoring of textile reinforced

CE

thermoplastic composites," Materials Technology, vol. 26, pp. 128-134, 2011.

AC

[20] T. M. A Horoschenkoff, C Strössner,K Farmbauer, "Use of carbon-fibre sensors to determine the deflection of composite-beams," 18th International Conference on Composite Materials, International Conference on Composite Materials, Jeju., 2011. [21] R. Salvado, C. Lopes, L. Szojda, P. Araújo, M. Gorski, F. J. Velez, et al., "Carbon fiber epoxy composites for both strengthening and health monitoring of structures," Sensors, vol. 15, pp. 10753-10770, 2015.

ACCEPTED MANUSCRIPT [22] G. Y., B. S., R. O., Q. T., and G. T., "Integrated self-monitoring of carbon based textile reinforced concrete beams under repeated loading in the un-cracked region," Carbon, vol. 98, pp. 238-249, 2016. [23] S. Zhu and D. D. L. Chung, "Theory of piezoresistivity for strain sensing in

PT

carbon fiber reinforced cement under flexure," Journal of Materials Science, vol.

RI

42, pp. 6222-6233, 2007.

SC

[24] F. Y. Yeh, K. C. Chang, and W. C. Liao, "Experimental Investigation of Self-Sensing Carbon Fiber Reinforced Cementitious Composite for Strain Measurement of

NU

an RC Portal Frame," International Journal of Distributed Sensor Networks, vol.

MA

2015, pp. 1-13, 2015.

[25] H. Inada, K. Ishii, and Y. Sugimura, "Carbon fiber sensor for monitoring integrity

D

of concrete piles," Proceedings of SPIE - The International Society for Optical

PT E

Engineering, vol. 5057, 2003.

[26] S. Wen and D. D. L. Chung, "Electrical-resistance-based damage self-sensing in

CE

carbon fiber reinforced cement," Carbon, vol. 45, pp. 710-716, 2007.

AC

[27] S. Wen and D. D. L. Chung, "Spatially resolved self-sensing of strain and damage in carbon fiber cement," Journal of Materials Science, vol. 41, pp. 4823-4831, 2006.

[28] D. D. L. Chung, "Carbon materials for structural self-sensing, electromagnetic shielding and thermal interfacing," Carbon, vol. 50, pp. 3342-3353, 2012. [29] X. Fu, C. Ddl., and W. Lu, "Ozone treatment of carbon fiber for reinforcing cement," Carbon, vol. 36, pp. 1337-1345, 1998.

ACCEPTED MANUSCRIPT [30] "TORAY CARBON FIBERS AMERICA, INC," www.torayusa.com. [31] C. Comotti, D. Regazzoni, C. Rizzi, and A. Vitali, "Multi-material design and 3D printing method of lower limb prosthetic sockets," in Rehab 2015 - Workshop on Icts for Improving Patients Rehabilitation Research Techniques, 2015.

PT

[32] Mei and Zhe, "An Approach for The Development of Low Cost Prosthetic Limbs

RI

with 3D Printing Technology," 2015.

SC

[33] Y. Jin, Y. He, and A. Shih, "Process Planning for the Fuse Deposition Modeling of Ankle-Foot-Othoses," Procedia Cirp, vol. 42, pp. 760-765, 2016.

NU

[34] T. L. Nguyen, S. J. Allen, and S. J. Phee, "Exploiting 3D printing technology to

MA

develop robotic running foot for footwear testing," Virtual & Physical Prototyping, vol. 8, pp. 259-269, 2013.

D

[35] P. K. Maji, "Additive manufacturing in prosthesis development – a case study,"

AC

CE

PT E

Rapid Prototyping Journal, vol. 20, pp. 480-489, 2014.

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

Graphical abstract

ACCEPTED MANUSCRIPT Highlights

 An effective method to resolve the intrinsic inadequate strength of three-dimensional printed structures with continuous carbon fibers was

PT

presented.

RI

 Self-monitoring together with structural reinforcement of carbon-fiber-embedded

SC

three-dimensional printed structures were realized.

 Lightweight and print duration reduction of three-dimensional printed structures

AC

CE

PT E

D

MA

NU

were achieved.