Fiber laser processing of GFRP composites and multi-objective optimization of the process using response surface methodology

Abstract

In the present investigation, a continuous wave fiber laser with maximum power of 400 W was used to cut a glass fiber reinforced plastic sheet of 4.56 mm thickness using Nitrogen as assisting gas. The influence processing parameters such as laser irradiance, gas pressure, and cutting speed on the cut surface quality were investigated by using response surface methodology. The different responses of laser cut surface such as upper kerf width, taper percentage along the cut depth, and heat-affected zone on the top surface were measured to analyze the influence of input process parameters on the responses. A statistical analysis on the obtained results was conducted and found that the optimum values of different input process parameters were laser irradiance: 8.28 × 10⁵ watt/cm², cutting speed: 600 mm/min and assisting gas pressure: 7.84 bar. The corresponding values of responses were upper kerf width: 177.4 µm, taper 0.73%, and heat-affected zone on top surface: 109.23 µm. The confirmation experiments were conducted with the obtained optimum parameter setting and observed that the predicted values and experimental values for upper kerf width, taper percentage and top surface heat-affected zone were within the error limits of 2.52%, 1.84%, and 0.45%, respectively. Furthermore, damages like loose fibers, interlayer fractures, evaporation of matrix material and fiber breakages were observed.

Keywords

Multi-response optimization response surface methodology glass fiber reinforced plastic laser beam cutting fiber laser kerf width heat-affected zone taper percentage

Introduction

Fiber reinforced plastics (FRPs) has become one of the most popular advanced material in industries for their outstanding properties such as lightweight, high stiffness, high strength-to-weight ratio, wear and temperature resistance. Their wide varieties of applications include automobile, aerospace, sports, transportation, infrastructure, and biomedical industries.^1–4 Glass fiber reinforced plastic (GFRP) is an advanced epoxy matrix-based composite material, which is widely used in aerospace industries for making aircrafts, satellites, rockets, and missiles due to its insulating properties and dimensional stability at high-temperature working environments.^5–8 Machining of the FRPs has been a challenging task, which is very different from machining of metals and alloys due to their anisotropic orientation of fibers, heterogeneous material properties, low thermal conductivity as well as heat sensitivity. Due to its extensive use, the machining of GFRP has become very important; in this connection, different processes are required to be explored to impart high cutting speed, good dimensional accuracy, high surface quality at the cutting zone, and less alteration of mechanical properties of the base material.⁹ Conventional methods like turning, milling, grinding, and drilling are popular for cutting/machining of GFRP. But these methods have major drawbacks, as the abrasive nature of glass fibers generate high tool wear which subsequently leads to high cutting force and low surface quality. The use of cubic boron nitride (CBN), cemented carbide, and diamond (PCD) tools have solved few of the above problems to some extent. The problems like fiber pull-out and delamination are some factors in conventional methods which hamper the cutting process.^10–14 In this perspective, the laser beam cutting may be an attractive option as it does not use a physical tool.

Laser beam cutting is a thermal process where material removal takes place by melting and vaporization when a highly collimated laser beam is focused on a localized area. The high laser beam energy helps in melting of the material and the molten material is removed from the cutting zone with a high-velocity co-axial gas jet leaving a clean edge. This process is gaining popularity due to the contactless cutting with no tool wear, controlled machining environment, varied energy density, flexibility in selection of process parameters, ease of automation, minimum heat-affected zone (HAZ), fast processing, and precise cuts.¹⁵ The most common lasers that are used in the industry are neodymium-doped yttrium aluminum garnet (Nd: YAG) laser and carbon dioxide (CO₂) laser. But at present, the disk and fiber lasers have been increasingly used to cut composites materials due to their high flexibility and low maintenance cost.¹⁶ In case of laser cutting of FRPs, the cut surface quality can be measured by analyzing the kerf width, the taper percentage of the machined surface, HAZ, microstructure analysis, striation lines, dross formation, etc. Caprino et al.¹⁷ used CO₂ laser with multimodal Gaussian power distribution to cut GFRP, i.e. the power is non-axially symmetrically distributed. They further put forth an analytical model that can predict the kerf depth as a relationship with laser beam travel in accordance with the workpiece material. They proposed that the kerf width depends on the spatial distribution of power and fiber distribution in the matrix.

Cenna and Mathew¹⁸ have developed a mathematical model to predict the laser processing parameters for different composites to achieve different responses such as inlet and outlet kerf width, kerf angle, and energy loss through transmission in the kerf. They have validated the model and found out an excellent agreement between the experimental results and the predicted data. However, for GFRP material, their predicted results differ from the experimental results. Hirogaki et al.¹⁹ used a CO₂ laser of the maximum power of 100 W to drill blind holes on GFRP with a spot diameter of 0.3 mm. They found considerable surface roughness on the wall of the hole due to the difference in evaporation temperature between resin and fiber. Tuersley et al.²⁰ machined a composite made up of borosilicate glass matrix and SiC fibers with the help of a pulsed Nd:YAG laser of maximum power of 400 W.They found that the material removal rate (MRR) was greatly affected by maximum laser power and pulse energy had a greater influence than the pulse duration. They have also conducted experiments by using various assist gases with different pressures and suggested to avoid oxygen as assist gas because it oxidizes the molten matrix material. But when the nitrogen is used as the assist gas it improved the situation as it stopped the burning process and allowed the vaporization of material. Yung et al.²¹ studied the HAZ of 500 µm holes in printed circuit board made of 1.6-mm-thick GFRP drilled by YAG laser. They proposed that the pulse frequency and laser power are the most important parameters that influence the structure of HAZ formed during drilling. Higher pulse frequency and laser power increase the width as with high temperature the fibers get melted and the resin is vaporized. Some of the melted materials adhere to the hole creating a charcoal-like structure which decreases the quality of the holes. Goeke and Emmelmann²² conducted experiments to analyze the influence of laser cutting parameters like laser power, cutting feed rate, laser mode or wavelength on carbon fiber reinforced part quality. They found that at constant laser power both kerf width and HAZ significantly decreased with the increase in cutting feed rate. Quintero et al.²³ studied the impact of laser cutting parameters on surface integrity of cut surface during the CO₂ laser cutting of phenolic resin with the help of a design-of-experiment (DOE) technique. Besides, investigating the process parameters, they studied the potential hazardous gases that are generated during laser cutting of phenolic resin and found the presence of benzene and formaldehyde in the vapors, which are toxic in nature. The authors have also concluded that like the conventional machining, laser cutting has also some drawbacks due to the heterogeneous reaction of fiber and matrix. It creates charring, uneven evaporation, delamination, and matrix decomposition. Riverio et al.²⁴ examined the performance of both continuous wave (CW) and pulsed mode CO₂ gas laser to cut 3-mm-thick sheets of CFRP composite and investigated the influence of process parameters on the cut surface qualities like taper, kerf width, and HAZ. They observed large HAZ in CW mode in comparison to the pulsed mode and with an increase in laser power, the HAZ increased linearly. Choudhury and Chuan²⁵ used 500 W CW CO₂ laser to cut GFRP with both single-pass and double-pass beam mode to cut the GFRP and access the cut surface quality. They reported that the double-pass laser beam produces much better surface quality than the single-pass beam. They also studied the influence of laser input parameters like nozzle diameter, cutting speed, and material thickness on output parameters like surface roughness and kerf width at the top surface and a bottom surface. Between single-pass and double-pass beam, the former produced better surface finish but the later produced lower kerf width. Moreover, they found the nozzle diameter a significant parameter as better surface finish and lower kerf width is achieved in smaller nozzle diameter. Patel et al.²⁶ have used artificial neural network (ANN) modeling to predict HAZ for a given laser power, assist gas pressure, and cutting speed during fiber laser cutting of GFRP. They found that the cutting speed is the most influencing parameter that determines the quality and the dimension of HAZ followed by laser power and gas pressure. Shyha et al.²⁷ have done a comparative study of CO₂ laser cutting of both CFRP and GFRP and found that the latter can be easily processed through laser cutting than the former due to the lower melting temperature of glass fibers. They also found that the surface roughness in case of laser cutting GFRP is higher due to the re-deposition of burnt matrix material on the cut surface.

From the literature review, it is observed that investigation on GFRP cutting through Nd: YAG, CO₂ lasers have been reported widely. But very limited studies on fiber laser cutting of GFRP have been reported so far. It has advantages like superior beam quality, low cost of operation, compact size, and is easy to use. In a fiber laser, the beam can be focused to achieve a high level of precision and greater intensity with longer processing distances.²⁸ In the present investigation, a CW fiber laser with maximum power 400 W is used for cutting the GFRP and the influence of the process parameters like laser irradiance, cutting speed, and gas pressure on the cut surface quality has been investigated. The surface quality parameters considered here are upper kerf widths, taper percentage, and top face HAZ. Response surface methodology (RSM) is followed for cutting parameter optimization in which experiments were planned as per the central composite design (CCD) and optimal parameters are searched using the multi-response optimization technique. Additionally, the significance of the developed model has been identified using analysis of variance (ANOVA) and the model is validated by doing confirmation tests. The microstructure, surface integrity analysis has been carried out on the laser-cut surfaces.

Materials and methods

The experiments were conducted in a CW fiber laser (400 W, 1070 nm, make: SPI, UK) Workstation (Figure 1). A fine laser cutting head with nozzle diameter 0.4 mm was used which can be moved along the z-axis. A focused beam diameter of 0.2 mm was obtained by maintaining the stand-off distance of 0.4 mm between the workpiece and tip of the nozzle. Nitrogen was purged through the nozzle during the cutting operation. GFRP composite sheet of dimensions 100 mm × 50 mm × 4.56 mm was used as the workpiece and mounted in a fixture to conduct laser cutting experiments. The matrix of the GFRP was epoxy (thermosetting polymer) and the reinforcement was the E-glass fiber whose properties of are given in Table 1.

Table 1.

Properties of E-glass fiber.

Properties	Value
Tensile Strength	3.5 GPa
Modulus	73.5 GPa
Elongation	4.8%
Density	2.57 g/cm³
Refractive index	1.547
Coefficient of thermal expansion	50–52 × 10⁻⁷m/m^oC

RSM has been extensively used for designing the experiment, development of the regression model, and finding the optimal conditions for desirable responses by reducing the number of experiments.²⁹ In the present investigation, the Central Composite Design (CCD) has been used, where the number of continuous factors or input parameters is three with no categorical factors. Different independent parameters considered in this study are the laser irradiance (A), cutting speed (B), and assist gas pressure (C). A number of trial runs were performed to find out the range of parameters that could produce a through cut of the GFRP sheet. The coded form of the parameters and their respective values are given in Table 2. The other independent parameters such as nozzle diameter, stand-off distance, and material thickness were kept constant. Design Expert software was used for the design of the experiment and analysis of the responses. A total of 20 runs were conducted in a single block by taking three levels of cutting parameters, which are presented in Table 3. The present design block consists of eight cube points (i.e. 2ⁿ, n: no. of factors), six center points in the cube (2 × n), six axial/star points (2 × n), and there is no central point in an axial direction and the distance of axial points from the center of cube i.e. alpha is 1.68 (≅ α = [2³]^1/4 where 3 is the number of variable).

Table 2.

Input factors and their levels.

Input parameters	Units	Factors in coded form	Levels
Input parameters	Units	Factors in coded form	−1	0	1
Laser irradiance	Watt/cm²× 10⁵	A	8.28	9.87	11.46
Cutting speed	mm/min	B	200	400	600
Gas pressure	Bar	C	7.84	10.78	13.73

Table 3.

Experimental parameter settings and measured responses.

Input parameters					Results
Std order	Run order	Laser irradiance (Watt/cm²× 10⁵)	Cutting speed (mm/min)	Gas pressure (bar)	Upper kerf width (µm)	Lower kerf width (µm)	Taper (%)	Top surface HAZ (µm)
1	20	8.28	200	7.84	226.22	170.93	1.21	197.14
2	3	11.46	200	7.84	229.58	167.19	1.37	253.25
3	1	8.28	600	7.84	173.16	142.04	0.68	100.48
4	12	11.46	600	7.84	209.41	145.34	1.40	187.47
5	10	8.28	200	13.73	236.85	133.42	2.27	211.46
6	11	11.46	200	13.73	247.26	150.64	2.12	236.87
7	13	8.28	600	13.73	203.67	172.36	0.69	145.09
8	6	11.46	600	13.73	223.44	184.34	0.86	175.73
9	8	7.19	400	10.78	204.97	146.76	1.28	179.17
10	2	12.54	400	10.78	237.61	168.47	1.52	252.88
11	4	9.87	63.64	10.78	247.61	162.23	1.87	227.88
12	18	9.87	736.36	10.78	195.79	174.17	0.47	114.80
13	7	9.87	400	5.83	200.53	152.20	1.06	196.69
14	16	9.87	400	15.74	218.28	151.29	1.47	219.9
15	14	9.87	400	10.78	208.60	149.43	1.30	167.45
16	9	9.87	400	10.78	200.76	149.81	1.12	186.89
17	17	9.87	400	10.78	210.57	144.35	1.45	189.62
18	19	9.87	400	10.78	210.91	147.33	1.39	167.59
19	5	9.87	400	10.78	210.63	146.60	1.40	184.65
20	15	9.87	400	10.78	202.42	144.63	1.27	186.51

HAZ: heat-affected zone.

For measurement of kerf widths and the HAZ, an optical microscope (Olympus) was used. For each data point, the multiple readings were recorded and the average of all was taken for analysis. The kerf width is the distance between the two faces which are produced due to the melting/burning of material during the laser beam scanning (Figure 2). Here, the upper kerf width i.e. the size of cutting zone at the top surface and lower kerf width i.e. the size of cutting zone at the bottom surface were measured and the taper percentage was calculated with the help of the following formula

Taperpercentage = \frac{{\begin{matrix} Upperkerfwidth \\ - Lowerkerfidth \end{matrix}}}{Thicknessofworkpiece} \times 100

(1)

Figure 1.

Experimental set-up of fiber laser cutting.

Figure 2.

Schematic diagram of kerf width, taper angle, and HAZ. HAZ: heat-affected zone.

Similarly, the width of the top surface HAZ for each experiment was measured multiple times on both sides of the kerf width and the average value was considered for analysis. The HAZ in composite materials is characterized by thermal damages such as matrix decomposition, matrix recession, and delamination, which produces poor surface quality and reduces the life of the component. Figure 2 indicates the different parameters of the laser cut surface.

Results and discussions

In this study, the experiments were conducted as per the design table (Table 3) and the responses, i.e. upper kerf width, lower kerf width, taper percentage, and top surface HAZ, were measured and entered into the design table for analysis. The image of the laser cut specimen is shown in Figure 3. The ANOVA of the results was carried out to find the significant parameters that are influencing the responses. The parameters of statistical analysis like p values, f values, the sum of square errors, and standard error of estimate were calculated for different parameters and the interpretation of results of statistical analysis was carried out in further sections. Then the optimal parameters for narrow kerf width, low taper percentage, and minimum HAZ were found through multi-response optimization technique with the help of Design Expert software.

Figure 3.

Laser cut marks on the GFRP. GFRP: glass fiber reinforced plastic.

Analysis of kerf width

ANOVA is one of the most used statistical technique to determine the influence of machining parameters on the surface integrity (kerf width, taper, and HAZ). The ANOVA for the upper kerf width at a confidence level of 95% is presented in Table 4. The p value of the model is less than 0.05 and the f value is 30.90, which indicates that the model for upper kerf width is significant. The model terms, in this case, laser irradiance (A), cutting speed (B), gas pressure (C), squared effects of laser irradiance (A²), cutting speed(B²), the combined influence of laser irradiance and cutting speed are having p values less than 0.05 and hence these are significant parameters. The values of predicted R² and adjusted R² are 0.82 and 0.93, respectively, and the difference is less than 0.2, which shows the predicted R² and the adjusted R² are in reasonable agreement. The adequate precision ratio is 21.17, which indicates an adequate signal. This parameter gives the signal-to-noise ratio and a ratio greater than 4 is acceptable as it compares the range of predicted values at the design points to the average variance of the prediction. The adequacy of the model was also checked by utilizing a graphical strategy by plotting the normal plot of residuals as shown in Figure 4(a). The figure indicates that all the residuals fall in close to the straight line, i.e. errors are distributed normally hence the proposed models for upper kerf width are adequate. The regression equation developed by software for prediction of output is presented in a coded form in equation (2)

\begin{matrix} UpperKerfWidth = + 207.30 + 9.13 A - 15.92 B \\ + 7.52 C + 5.28 AB + 5.07 A^{2} \\ + 5.22 B^{2} + 0.8692 C^{2} \end{matrix}

(2)

Table 4.

Analysis of variance (ANOVA) for upper kerf width.

Source	Sum of squares	df	Mean square	f Value	p Value
Model	6332.25	9	703.58	30.90	<0.0001	significant
A-Laser irradiance	1138.18	1	1138.18	49.98	<0.0001	significant
B-Cutting speed	3460.43	1	3460.43	151.97	<0.0001	significant
C-Gas pressure	772.46	1	772.46	33.92	0.0002	significant
AB	223.13	1	223.13	9.80	0.0107	significant
AC	11.09	1	11.09	0.4871	0.5011
BC	32.93	1	32.93	1.45	0.2569
A²	370.61	1	370.61	16.28	0.0024	significant
B²	392.23	1	392.23	17.23	0.0020	significant
C²	10.89	1	10.89	0.4781	0.5050
Residual	227.71	10	22.77
Lack of fit	124.73	5	24.95	1.21	0.4193	not significant
Pure error	102.97	5	20.59
Cor total	6559.95	19
Standard deviation			4.77	R²		0.9653
Mean			214.92	Adjusted R²		0.9340
Coefficient of variation (%)			2.22	Predicted R²		0.8164
Sum of squares residual error (predicted) (PRESS)			1204.62	Adeq. Precision		21.1679

Figure 4.

(a) Normal probabilities of residuals for upper kerf width, (b) surface plot of upper kerf width with cutting speed and laser irradiance (at gas pressure :10.78 bar), (c) surface plot of upper kerf width with gas pressure and laser irradiance (at cutting speed 400 mm/min), and (d) 3D surface plot of upper kerf width with cutting speed and gas pressure (at laser irradiance 9.87 W/cm²× 10⁵).

In any sheet-cutting process, the kerf width is an important parameter as it shows the MRR of the process. The lesser the material removed, i.e. less kerf width, more the surface quality of the material. In laser cutting the material is removed by the high energy generated by a focused laser beam that evaporates the material along the scanning path of the laser. During laser cutting procedure, a higher kerf width at the upper surface is obtained in comparison to the bottom surface because the absorption of laser energy by the material reduces along the depth or thickness of the workpiece.³⁰

From the ANOVA Table 4, it is found that the individual parameters, i.e. laser irradiance, gas pressure, and cutting speed, are significantly affecting the upper kerf width. Figure 4(b) to (d) represent the 3D surface plots that represent the relationship between different laser cutting parameters and the upper kerf width. The 3D surface plots reveal that the laser irradiance and gas pressure have a proportional relationship with the upper kerf width as higher laser irradiance and gas pressure increases the kerf width. The cutting speed indirectly affects the upper kerf width as it decreases with the increase in cutting speed. As the intensity of the laser beam increases, more energy is absorbed and hence more amount of workpiece material is evaporated causing an increase of kerf width. But, it is observed that a narrow kerf width can be achieved with the increase in cutting speed. As we go on increasing the cutting velocity, the contact time between the laser beam and workpiece would decrease, which tends to the reduction in kerf width.^18,25 But high cutting speed may result in less penetration of laser beam onto the workpiece and thus the cutting speed must be chosen appropriately. In laser cutting, the assist gas (in this case Nitrogen) helps in the ejection of molten material from the cutting zone and prevents high thermal damage.³¹ Low gas pressure causes no ejection and subsequent solidification of the molten matrix, which hampers the cutting of fibers. Consequently, this may cause cutting process unsuccessful throughout the entire thickness of the material. So, with high gas pressure, the molten material gets ejected out easily, producing a higher kerf width.³²

Analysis of taper percentage

The results of ANOVA analysis for taper percentage are listed in Table 5. The p value of the model developed for the taper percentage is less than 0.05 and f value is 52.41, which indicate significant model. The model terms laser irradiance (A), cutting speed (B), gas pressure (C), and interaction influences between laser irradiance and scanning speed (A × B), scanning speed and gas pressure (B × C), and laser irradiance and gas pressure (A × C) are significant as the p values for the terms is less than 0.05. The predicted R² and adjusted R² give values 0.92 and 0.94, respectively, and the difference is less than 0.2, which shows the predicted R² and the adjusted R² are in reasonable agreement. The model is giving an adequate signal as the adequate precision ratio is 26.288, which measures signal-to-noise ratio and an acceptable ratio is greater than 4. In Figure 5(a), the normal plot of residuals is given, which indicates that all the residuals fall in close to the straight line, i.e. errors are distributed normally hence the proposed model of taper percentage is adequate. The regression relation for prediction of taper percentage is presented in a coded form in the equation (3)

\begin{matrix} Taperpercentage = + 1.31 + 0.0954 A - 0.4165 B \\ + 0.1429 C + 0.1109 AB \\ - 0.1071 AC - 0.2937 BC \end{matrix}

(3)

Table 5.

analysis of variance (ANOVA) for taper percentage.

Source	Sum of squares	df	Mean square	f Value	p Value
Model	3.65	6	0.6087	52.41	<0.0001	Significant
A-Laser irradiance	0.1242	1	0.1242	10.70	0.0061	Significant
B-Cutting speed	2.37	1	2.37	203.96	<0.0001	Significant
C-Gas pressure	0.2787	1	0.2787	24.00	0.0003	Significant
AB	0.0984	1	0.0984	8.47	0.0122	Significant
AC	0.0917	1	0.0917	7.90	0.0147	Significant
BC	0.6902	1	0.6902	59.42	<0.0001	Significant
Residual	0.1510	13	0.0116
Lack of fit	0.0767	8	0.0096	0.64	0.72	Not significant
Pure error	0.0743	5	0.0149
Cor total	3.80	19
Standard deviation			0.12	R²		0.96
Mean			1.31	Adjusted R²		0.94
Coefficient of variation (%)			8.23	Predicted R²		0.92
Sum of squares residual error (predicted) (PRESS)			0.30	Adeq. Precision		26.29

Figure 5.

(a) Normal probabilities of residuals for taper percentage, (b) 3D surface plot of taper percentage with cutting speed and laser irradiance (at gas pressure 10.78 bar), (c) 3D surface plot of taper percentage with gas pressure and laser irradiance (at cutting speed 400 mm/min), (d) 3D surface plot of taper percentage with cutting speed and gas pressure (at laser irradiance 9.87 W/cm²× 10⁵).

The variation of kerf width along the thickness of workpiece after machining is characterized by taper percentage, which gives the geometric accuracy of the machined surface. It is desirable to get a minimum taper percentage for a uniform cut throughout the material thickness. Figure 5(b) to (d) show the response surface plot for variation of taper percentage with respect to different input machining parameters. It is observed that the taper percentage varies proportionally with the laser irradiance and gas pressure whereas it varies indirectly with the cutting speed. As discussed earlier, higher laser irradiance produces larger kerf width at the top surface than the bottom surface, thus producing higher taper percentage. On the other hand, higher cutting speed produces lower taper as at higher cutting speed, the time of exposure of laser beam on the workpiece surface is less in comparison to lower cutting speed.²⁷ This produces low kerf width at the top surface and hence low taper percentage. Increased gas pressure removes a larger amount of molten material from the top surface producing larger kerf width whereas bottom surface remains unaffected. So larger kerf width at the top surface and smaller kerf width at bottom surface produce higher taper percentage.

Analysis of HAZ

The ANOVA analysis of HAZ is given in Table 6. The p value for the developed model for HAZ is less than 0.05 and the f value is 29.81, hence the model is significant. The model terms laser irradiance (A), cutting speed (B), squared effects of laser irradiance (A²), and gas pressure (C²), and combined influence of laser irradiance and gas pressure are significant as they are having p values less than 0.05. The adjusted R² and predicted R² are 0.93 and 0.83, respectively, and the difference is less than 0.2, which implies the model is well capable of predicting the width of HAZ. The model is giving an adequate signal as the adequate precision ratio is 20.59. The adequacy of the proposed models of HAZ is confirmed from the graphical normal plot of residuals (Figure 6(a)) as all the residuals fall in close to straight line, i.e. errors are normally distributed. The regression equation developed by software for prediction of output is presented in a coded form in equation (4)

\begin{matrix} TopsurfaceHAZ = + 180.82 + 23.66 A - 35.16 B \\ - 10.88 AC + 10.21 A^{2} + 7.48 C^{2} \end{matrix}

(4)

Table 6.

Analysis of variance (ANOVA) for HAZ.

Source	Sum of squares	df	Mean square	f Value	p Value
Model	29011.03	9	3223.45	29.81	<0.0001	Significant
A-Laser irradiance	7644.52	1	7644.52	70.69	<0.0001	Significant
B-Cutting speed	16879.35	1	16879.35	156.08	<0.0001	Significant
C-Gas pressure	357.06	1	357.06	3.30	0.0993
AB	162.95	1	162.95	1.51	0.2477
AC	946.89	1	946.89	8.76	0.0143	Significant
BC	152.47	1	152.47	1.41	0.2625
A²	1502.00	1	1502.00	13.89	0.0039	Significant
B²	450.11	1	450.11	4.16	0.0686
C²	805.66	1	805.66	7.45	0.0212	Significant
Residual	1081.45	10	108.15
Lack of fit	567.02	5	113.40	1.10	0.4588	Not significant
Pure error	514.44	5	102.89
Cor total	30092.49	19
Standard deviation			10.40	R²		0.96
Mean			189.08	Adjusted R²		0.93
Coefficient of variation (%)			5.50	Predicted R²		0.83
Sum of squares residual error (predicted)(PRESS)			5157.03	Adeq. Precision		20.59

HAZ: heat-affected zone.

Figure 6.

(a) Normal probabilities of residuals for HAZ, (b) 3D surface plot of top face HAZ with cutting speed and laser irradiance (at gas pressure 10.78 bar), (c) 3D surface plot of top face HAZ with gas pressure and laser irradiance (at cutting speed 400 mm/min), (d) 3D surface plot of top face HAZ with cutting speed and gas pressure (at laser irradiance 9.87 W/cm²× 10⁵). HAZ: heat-affected zone.

HAZ in fiber laser cutting of FRPs is characterized by matrix shrinkage and decomposition, fiber debonding, and delamination and char formation. The same phenomenon is evident from the Figure 7(a) and (b), where HAZ formation has occurred along the irregular region parallel to the laser cut and delamination and char formation are the prominent features. As the melting temperature of both the fiber and matrix are different, laser exposer causes evaporation of matrix material and expose the glass fibers. Also, glass fibers dissipate heat easily in comparison to the resin matrix due to more thermal conductivity.^33,34 Figure 6(b) to (d) shows the response surface plot for variation of HAZ with respect to different input machining parameters. It is observed that the HAZ increases linearly with the increase in laser irradiance, as more laser intensity would be absorbed by fibers and matrices causing vaporization of matrices and delamination of fibers. On the other hand, there is a decrease in the top face HAZ with an increase in cutting speed. The high cutting speed results in less interaction time of laser beam, causing less material to melt. Also, due to high cutting speed, the heat conduction by glass fibers and the matrix is reduced, causing less HAZ. ANOVA table shows that there is no significant effect of gas pressure on the top face HAZ, but from the graph, it is having a proportional relationship with the gas pressure.

Figure 7.

(a) Illustration of HAZ, kerf width for std order 18 (9.87 × 10⁵ watt/cm² laser irradiance, 400 mm/min cutting speed and 10.78 bar gas pressure) and (b) Illustration of kerf width, HAZ, and fiber pull-out. HAZ: heat-affected zone. HAZ: heat-affected zone.

Surface integrity of the cut surface

FESEM images of laser-cut surfaces of GFRP showing various surface damages are illustrated in Figure 8. Due to the high temperature, melting or charring of resin is seen in images. The charred materials adhere to the surface degrading the surface quality. These adhered melted materials are predominantly obtained in low gas pressure which is not enough to remove the molten material during machining, see Figure 8(d). From Figure 8(b), it is observed that at high gas pressure 13.73 bar, there is less molten material adhered to the surface. Also, due to high laser irradiance, maximum material evaporated, giving a clean-cut surface. But this also gives rise to matrix losses, which result in loose fibers and fiber pull-out. Also, evidence of inter-layer fractures is present in the images, which reduce the service life of the component. This may have been due to fast cooling rate after the laser cutting and also due to different thermal properties of matrix and fiber, which agrees with previous research.²⁷

Figure 8.

SEM images of surfaces produced at different cutting conditions.

Figure 9.

(a) Bar histogram representation for desirability and (b) ramp representation for overall desirability.

Multi-response optimization

The optimal input parameters were found through multi-response optimization by desirability function approach using Design Expert software. It is one of the most common methods used to optimize single- and multi-response problems. In this method, each response is converted into a dimensionless quantity called desirability, which varies from zero to one (zero gives least desirable and 1 gives most desirable). The maximum desirability of the solution containing the input process parameters is considered to be optimal.³⁵ In the present investigation, the Design Expert software has been utilized to decide the combined set of input process parameters, i.e. laser irradiance, cutting speed, and gas pressure, which fulfill the required objectives of every response and process parameters. The optimum value of the mentioned input parameters is obtained by minimizing the kerf width, taper percentage, and HAZ. The constraints for input parameters and the required goals for responses are recorded in Table 7 with optimal solutions. The optimum parameters for minimal kerf width, taper, and HAZ are found to be 8.28 Watt/cm²× 10⁵ laser irradiance, 600 mm/min cutting speed, and 7.84 bar gas pressure with the highest desirability of 0.91. The optimized bar histogram graph representation for desirability and ramp representation for the overall desirability of kerf width, taper, and HAZ have been demonstrated in Figure 9(a) and (b), respectively. The confirmatory experiments have been performed at optimum level parameters to validate the model and values of responses and error are listed in Table 8. It has been found that the obtained values are in close concurrence with the predicted values with error 2.52%, 1.85%, and 0.46% for kerf width, taper, and HAZ, respectively.

Table 7.

Constraints and optimal solutions.

Name	Goal	Lower limit	Upper limit	Lower weight	Upper weight	Importance	Optimal solutions
A: Laser irradiance (Watt/cm²× 10⁵)	is in range	8.28	11.46	1	1	3	8.28
B: Cutting speed (mm/min)	is in range	200	600	1	1	3	600
C: Gas pressure (bar)	is in range	7.84	13.73	1	1	3	7.84
Upper Kerf width (µm)	minimize	173.16	247.61	1	1	3	177.4
Taper (%)	minimize	0.47	2.26	1	1	3	0.73
Top surface HAZ (µm)	minimize	100.48	253.25	1	1	3	109.23

Table 8.

Experimental and predicted values of top face HAZ, taper percentage, and upper kerf width.

Responses	Predicted	Experimental	Percentage of error
Kerf width (µm)	177.4	172.92	2.52
Taper (%)	0.731	0.717	1.85
HAZ (µm)	109.228	108.72	0.46

HAZ: heat-affected zone.

Conclusions

This study demonstrated the fiber laser cutting of GFRP composite sheet of thickness 4.56 mm. The experimental planning was carried out as per the RSM. The multi-objective optimization of input process parameters, namely laser irradiance, cutting speed, and gas pressure, were performed. The output responses like upper kerf width, taper percentage of cut surface, and top surface HAZ were considered in the investigation. The following conclusions are drawn:

The significant parameters that influence the kerf width at top surface are cutting speed, laser irradiance, and gas pressure. The top surface kerf width varies proportionally with gas pressure and laser irradiance, whereas it reduces with the increase in cutting speed. The lowest kerf width achieved is 173.16 µm.

Laser irradiance, assist gas pressure, and cutting speed are the significant parameters for taper percentage. The taper percentage varies proportionally with laser irradiance and gas pressure, but it has an inverse relation with the cutting speed. The lowest taper percentage of the cut surface is 0.47%.

The HAZ on the top surface increases proportionally with the laser irradiance, whereas it is inversely proportional to the cutting speed. The minimum width of the HAZ is 100.48 µm.

Multi-objective optimization by desirability function has been carried out by defining a goal for each response parameter, i.e. to minimum upper kerf width, taper percentage, and top surface HAZ. The optimal solution of input process parameters is found to be 8.28 Watt/cm²× 10⁵ laser irradiance, 600 mm/min cutting speed, and 7.84 bar gas pressure and corresponding responses are upper kerf width 177.4 µm, taper 0.73%, and top surface HAZ 109.23 µm.

The confirmation experiments were conducted at optimum parameters and the error between predicted values and experimental values for upper kerf width, taper percentage, and top surface HAZ are found to be within 2.52%, 1.84%, and 0.45%, respectively. This indicates the good prediction ability of the developed regression model.

The microstructure analysis has been carried out for laser-cut surfaces and different surface damages are identified such as matrix losses, fiber breakages, loose fibers, interlayer fractures, and adhered fibers and matrix materials to the cut surface. However, the minimization surface damages or reconditioning prospects of the damage surfaces need further investigation.

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors acknowledge the financial support of Defence Research and Development Laboratory, Hyderabad for this work under the CARS grant.

References

Botana-Galvín

Blanco

González-Rovira

et al.

Adhesive behaviour of carbon fibre reinforced plastic panels manufactured using woven and unidirectional tape after ultraviolet laser surface treatment. J Compos Mater 2018; 52: 853–865.

Krishnamoorthy

Boopathy

Palanikumar

. Delamination analysis in drilling of CFRP composites using response surface methodology. J Compos Mater 2009; 43: 2885–2902.

Ghafarizadeh

Lebrun

Chatelain

. Experimental investigation of the cutting temperature and surface quality during milling of unidirectional carbon fiber reinforced plastic. J Compos Mater 2016; 50: 1059–1071.

Mostafa

Ismarrubie

Sapuan

et al.

Fibre prestressed polymer-matrix composites: a review. J Compos Mater 2017; 51: 39–66.

Palanikumar

Campos Rubio

Abrao

et al.

Influence of drill point angle in high speed drilling of glass fiber reinforced plastics. J Compos Mater 2008; 42: 2585–2597.

Kilickap

. Analysis and modeling of delamination factor in drilling glass fiber reinforced plastic using response surface methodology. J Compos Mater 2011; 45: 727–736.

Pinho

Carou

Davim

. Comparative study of the performance of diamond-coated drills on the delamination in drilling of carbon fiber reinforced plastics: assessing the influence of the temperature of the drill. J Compos Mater 2016; 50: 179–189.

Yao

Bijlaard

Kolstein

et al.

Tensile strength and fracture behaviors of complex GFRP composites with a central hole. J Compos Mater 2005; 39: 1247–1260.

El-Hofy

Soo

Aspinwall

et al.

Factors affecting workpiece surface integrity in slotting of CFRP. Proc Eng 2011; 19: 9–94.

10.

Lopez de Lacalle

Lamikiz

Campa

et al.

Design and test of a multitooth tool for CFRP milling. J Compos Mater 2009; 43: 3275–3290.

11.

Chang

. Turning of glass–fiber reinforced plastics materials with chamfered main cutting-edge carbide tools. J Mater Process Technol 2006; 180: 117–129.

12.

Azmi

Lin

Bhattacharyya

. Experimental study of machinability of GFRP composites by end milling. Mater Manufact Process 2012; 27: 1045–1050.

13.

Abrão

Faria

Rubio

et al.

Drilling of fiber reinforced plastics: a review. J Mater Process Technol 2007; 186: 1–7.

14.

Rajamurugan

Shanmugam

Palanikumar

. Analysis of delamination in drilling glass fiber reinforced polyester composites. Mater Design 2013; 45: 80–87.

15.

Dubey

Yadava

. Laser beam machining—a review. Int J Machine Tools Manufact 2008; 48: 609–628.

16.

Negarestani R and Li L. Laser machining of fibre-reinforced polymeric composite materials. In: Hocheng H (ed.) Machining technology for composite materials. Philadelphia, PA, USA: Woodhead Publishing, 2012, pp.288–308.

17.

Caprino

Tagliaferri

Covelli

. Cutting glass fibre reinforced composites using CO₂ laser with multimodal-Gaussian distribution. Int J Mach Tools Manufact 1995; 35: 831–840.

18.

Cenna

Mathew

. Analysis and prediction of laser cutting parameters of fibre reinforced plastics (FRP) composite materials. Int J Mach Tools Manufact 2002; 42: 105–113.

19.

Hirogaki

Aoyama

Inoue

et al.

Laser drilling of blind via holes in aramid and glass/epoxy composites for multi-layer printed wiring boards. Compos A Appl Sci Manufact 2001; 32: 963–968.

20.

Tuersley

Hoult

Pashby

. Nd–YAG laser machining of SiC fibre/borosilicate glass composites. Part I. Optimisation of laser pulse parameters. Compos A Appl Sci Manufact 1998; 29: 947–954.

21.

Yung

Mei

Yue

. A study of the heat-affected zone in the UV YAG laser drilling of GFRP materials. J Mater Process Technol 2002; 122: 278–285.

22.

Goeke

Emmelmann

. Influence of laser cutting parameters on CFRP part quality. Phys Proc 2010; 5: 253–258.

23.

Quintero

Riveiro

Lusquiños

et al.

CO₂ laser cutting of phenolic resin boards. J Mater Process Technol 2011; 211: 1710–1718.

24.

Riveiro

Quintero

Lusquiños

et al.

Experimental study on the CO₂ laser cutting of carbon fiber reinforced plastic composite. Compos A Appl Sci Manufact 2012; 43: 1400–1409.

25.

Choudhury

Chuan

. Experimental evaluation of laser cut quality of glass fibre reinforced plastic composite. Optics Lasers Eng 2013; 51: 1125–1132.

26.

Patel

Sheth

Patel

. Experimental analysis and ANN modelling of HAZ in laser cutting of glass fibre reinforced plastic composites. Proc Technol 2016; 23: 406–413.

27.

Shyha

Kuo

Soo

. Workpiece surface integrity and productivity when cutting CFRP and GFRP composites using a CO₂ laser. Int J Mechatr Manufact Sys 2014; 7: 93–107.

28.

Horley R, Norman S and Zervas MN. Progress and development in fibre laser technology. In: Proc. SPIE 6738, Technologies for Optical Countermeasures IV, 67380K, 11 October 2007. doi: 10.1117/12.753171.

29.

Myers RH and Montgomery DC. Response surface methodology: process and product optimization using designed experiments. New York: Wiley, 1995, pp.156–179.

30.

Uslan

. Co₂ laser cutting: kerf width variation during cutting. Proc Inst Mech Eng B J Eng Manufact 2005; 219: 571–577.

31.

Salem

Mansour

Badr

et al.

CW Nd: YAG laser cutting of ultra low carbon steel thin sheets using O2 assist gas. J Mater Process Technol 2008; 196: 64–72.

32.

Fatimah S, Ishak M and Aqida SN. CO₂ laser cutting of glass fiber reinforce polymer composite. In: IOP conference series: materials science and engineering, 2012, Vol. 36, No. 1, p.012033, IOP Publishing.

33.

Pan

Hocheng

. The anisotropic heat-affected zone in the laser grooving of fiber-reinforced composite material. J Mater Process Technol 1996; 62: 54–60.

34.

Pan

Hocheng

. Evaluation of anisotropic thermal conductivity for unidirectional FRP in laser machining. Compos A Appl Sci Manufact 2001; 32: 1657–1667.

35.

Dhanawade

Kumar

. Experimental study of delamination and kerf geometry of carbon epoxy composite machined by abrasive water jet. J Compos Mater 2017; 51: 3373–3390.