Sage Journals: Discover world-class research

Abstract

In-situ consolidation of thermoplastic composites using Automated Fibre Placement (AFP) technology is an emerging manufacturing technique, offering tailored composite properties through customised processing parameters. Multiple competing parameters during AFP manufacturing influence the quality and mechanical performance of the laminates. These lay-up parameters are interrelated, and often require comprehensive experimental characterisation which is costly and time-intensive. This study aims to optimise the fracture toughness of in-situ consolidated thermoplastic composite (AS4/APC-2) and investigate the mechanisms that contribute to it. Taguchi’s method is employed to efficiently analyse the effect of various process parameters at multiple levels. Based on the obtained results, a considerable effect of process parameters on Mode I and II fracture toughness is observed. The statistical analysis reveals that the Hot Gas Torch (HGT) temperature required for AFP processing significantly affects the Mode I fracture toughness, contributing to 33.8%. Whereas, the consolidation force, another key processing parameter in AFP notably affects Mode II fracture toughness, with the contribution of 81.8%. The analysis of variance (ANOVA) reveals interdependent processing parameter relations for both fracture modes. A validation test showed good agreement between the predicted fracture toughness and the experimental test.

Keywords

Automated fibre placement polymer-matrix composites fracture toughness crystallinity statistical analysis taguchi method

Introduction

Fibre reinforced polymer composites are increasingly used in applications where high specific strength and stiffness are desired. However, laminated composites exhibit poor through-thickness mechanical properties and are susceptible to delamination damage caused by impact loadings.^1–4 Delamination initiation and growth are governed by the interlaminar fracture toughness. Thermoplastic matrices are an excellent alternative to epoxy matrices due to their excellent Mode I and II interlaminar fracture toughness, and impact resistance.⁵ The initial perception of thermoplastic composites as expensive and challenging to process is changing. As such, thermoplastic composites are being considered for aircraft primary structures over traditional epoxy-based composites.^6,7 One of the key technical enablers is automated fibre placement (AFP) by in-situ consolidation of thermoplastic composites. In-situ consolidation describes the process of rapidly melting and bonding the thermoplastic matrix as the towpreg is being laid by the robot.⁸ The highly automated tow laying process promises to reduce manufacturing costs⁹ and manufacture composite structures with highly tailored behaviour.^10–12

The selection of AFP process parameters affects the final manufactured quality of the in-situ consolidated thermoplastic composite laminates. Hence, significant research interests have gone towards optimising parameters such as deposition rate, consolidation force, tool temperature, and process temperature in attempt to achieve autoclave level quality in in-situ consolidated laminates. The typical means of assessing manufacturing quality are porosity content and interlaminar strength development as the two parameters are closely related. Higher porosity content leads to lower interlaminar strength. Khan et al.¹³ found porosity content to reduce with decreasing deposition rate but high heating did not guarantee maximum consolidation. Repass treatment was found to improve the laminate’s surface finish¹⁴ and reduce porosity in the upper layers.¹⁵ Chen et al.¹⁶ reported reduction in porosity content with decreasing deposition rate and certain tool temperature had an important role in eliminating voids. By improving consolidation and promoting conditions for polymer chain diffusion, high tool temperature improves the degree of bonding and, consequently, interlaminar strength.^17,18 With sufficient polymer chain diffusion across the towpreg interface, strength development occurs. A combination of high hot gas torch temperature and consolidation force yielded the highest interlaminar strength whilst keeping the deposition rate constant.¹⁹ However, further increase beyond the optimum point risks fibre damage and thinning of samples. Comer et al.²⁰ reported that in-situ consolidated specimens using laser-assisted AFP had lower interlaminar strength than autoclaved consolidated specimens. This difference can be due to the short dwell time available for bonding to fully develop during the manufacturing process of laser-AFP. With a modified hot gas torch nozzle, Qureshi et al.²¹ reported the deposition rate had a significant effect on interlaminar shear strength whilst the consolidation force showed no notable effect. A high deposition rate decreases the thermal load on the towpregs and reduces the processing window to melt the polymer matrix. Schiel et al.²² investigated laser-assisted AFP process parameters and reported that an elevated tool temperature had a beneficial effect on the interlaminar strength development of in-situ consolidated thermoplastic composites. Similarly, when using laser-assisted AFP, Stokes-Griffin and Compston²³ found that the interlaminar strength remained relatively independent of the nip-point temperature (for a sufficiently high temperature) at deposition rate of 100 mm/s. This nip-point temperature dependency became more pronounced at higher deposition rate of 400 mm/s. Heathman et al.²⁴ performed optimisation studies on laser-AFP manufactured low-melt polyaryl-ether-ketone (LM-PAEK) and reported manufacturing defects similar to PEEK thermoplastic composites. This suggests that the manufacturing defects are related to the manufacturing process, rather than material specific issues.

Interlaminar fracture toughness is another interfacial property to consider in the in-situ consolidation quality. The microstructure of semi-crystalline thermoplastic polymer matrices, such as polyether-ether-ketone (PEEK), are affected by the processing conditions. This can lead to variation in fracture toughness that can affect out-of-plane impact resistance. Ray et al.²⁵ reported that the fracture toughness of laser assisted tape placement consolidated PEEK composites was higher than samples consolidated by autoclave. Oromiehie et al.²⁶ also concluded similar findings for hot gas torch AFP consolidated PEEK composites. Shafaq et al.²⁷ established that the deposition rate is critical to the fracture toughness properties of in-situ consolidated PEEK composites. Higher deposition rates lead to lower degree of crystallinity and higher fracture toughness. However, this is only valid when the tows have been properly consolidated. The brief literature review highlights the complexity of the in-situ consolidation processing parameters which influence the laminate quality. While many of the studies performed parametric optimisation with each processing parameters in isolation, the inter-dependence of the parameters cannot be ignored.

Roadblocks remain in achieving full mechanical performance of thermoplastic composites by AFP in-situ consolidation process before industry widespread adoption. The process is highly sensitive to the process parameters selected. There are likely several competing factors relating the process parameters to the laminate’s quality and mechanical performance. With slower deposition rate, there is more time for the molten polymer to flow and molecular diffusion across the interface. Keeping a constant heating power, a slower rate yields a higher intensity of heating, ultimately resulting in the attainment of a higher nip-point temperature. However, increasing the hot gas torch temperature with slow deposition rate increases the risk of thermal degradation to the polymer matrix. Often, these effects are non-linear, and will be in combination of two or more factors. Many of the presented works investigated the influence of individual parameters on in-situ consolidation without precisely quantifying their interdependence. In order to quantify the interdependence of the parameters, a significant experimental programme must be undertaken which can be prohibitively expensive. The Taguchi’s design of experiments (DOE) offers a practical means to reduce number of experiments in a robust manner.²⁸ The Taguchi’s orthogonal array systemically selects certain combinations of the process parameters for characterisation. The Taguchi’s DOE had been used to investigate the interdependence of fused deposition modelling process parameters, such as print speed, infill density, and layer thickness, to optimise the strength of the built parts.^29,30 Heathman et al.²⁴ conducted a DOE analysis of deposition rate, nip-point temperature and consolidation force on LM-PAEK composites to identify an optimum set of parameters. Khan³¹ performed DOE repair conditions of crack healing of Elium thermoplastic composites. Belhaj et.al.³² conducted a study on the effect of AFP process parameters to examine tackiness and optimise the lay-up process. Through their research, they successfully identified a solution to the problem with a minimal number of trials. Therefore, the effect of AFP process parameters should be studied synergistically.

The objective of this research is to optimise the process parameters influencing fracture toughness and to analyse the challenges that dictate the performance characteristics of the laminate. The paper investigates the inter-dependence of processing parameters on Mode I and II interlaminar fracture toughness to maximise the efficiency and performance of AS4/APC-2 laminates. The processing parameters for manufacturing selected for the DOE study are deposition rate, hot gas torch (HGT) temperature, consolidation force and tool temperature and analysed using Taguchi’s orthogonal array. The statistical analysis of the obtained results identifies the interdependence and the optimised process parameters that affect interlaminar fracture toughness of laminates.

Experimental method

Material and manufacturing

In this study, carbon fibre PEEK composites, AS4/APC-2 towpregs of 6.5 mm width and a fibre volume fraction of 65%, was used. The glass transition temperature (T_g) from the supplied datasheet was 143°C. A quasi-isotropic panel of 350 mm × 250 mm was consolidated using an AFP machine. The AFP machine consists of a seven-axis robot platform, stainless steel consolidation roller and a HGT. The HGT can expel hot nitrogen gases up to 1000°C and a flow rate up to 100 standard litres per minute. The steel consolidation roller, with a diameter of 12 mm, is used to ensure uniform distribution of consolidation force. The heated tool plate was designed and built in-house, schematically shown in Figure 1. The tool plate was constructed from 6 mm thick aluminium with heating elements placed on the underside. A thermocouple was embedded on the topside of the tool plate to monitor the temperature where a PID controller was used to control the tool temperature. The heated tool plate was secured onto the spindle and thermal insulation was placed all round to minimise heat losses during layup and consolidation process.

Figure 1.

(a) Schematic of heated tool plate arrangement (b) Heated tool plate surface with initial layers attached.

The stacking sequence of the panels were [0/45/90/-45/0]_4S which gave a nominal manufactured thickness of 6 mm. After the 20^th layer was laid, the manufacturing process was paused to place a steel shim of 25 μm²⁶ at the pre-crack location. A release agent was applied prior to the application of the steel shims and secured with Kapton tape. The steel shim acted as a starter crack for the Mode I and II fracture toughness tests. After the steel shim was secured, the consolidation of the subsequent layers continued.

Selection of process parameters

From the literature review, the deposition rate, HGT temperature, tool temperature, and consolidation force was identified to have substantial influence on the consolidation quality and interlaminar shear strength. Previous comprehensive investigations by the research group were performed on individual processing parameters to obtain a set of optimum parameters. The studies concluded that improvements in properties can be achieved by increasing or decreasing a parameter at one time to a certain limit. Increasing the values beyond this limit typically resulted in a reduction in properties. Table 1 lists the parameters which will be studied in this work and at three levels. The levels were selected based on prior investigations.^19,27 Level 2 typically represent an “optimum” value for a given set of parameters. It was found that an optimal bonding strength is achieved at a deposition rate of 76 mm/s.¹⁹ A very high deposition rate hinders consolidation within the laminate. Therefore, the deposition rate was maintained between 76 and 124 mm/s. Higher rates of deposition require higher HGT temperatures to ensure proper consolidation of the prepreg layers. Reduced amount of voids could be obtained by heating the tool to a temperature of 60°C, however improved toughness are obtained by increasing the tool temperature up to 105°C.²¹ Hence, 100°C was chosen as upper limit of tool temperatures. A minimum force of 180 N is required for consolidation.¹⁹ By increasing the lay-up speed, a higher consolidation force is needed to achieve better cross-linking of resin. However, applying a force beyond 300 N at a deposition rate of 76 mm/s results in polymer degradation. Therefore, the levels 1 to 3 were selected considering all of the above factors. These levels and factors will give a full factorial of 81 combinations to investigate the inter-dependence of the selected processing parameters.

Table 1.

Manufacturing parameters and their levels.

Factors	Level 1	Level 2	Level 3
Deposition rate (mm/s)	76	100	124
HGT temperature (˚C)	850	900	950
Tool temperature (˚C)	20	60	100
Consolidation force (N)	180	230	300

Design of experiments

Design of experiments entails developing a series of tests where the factors are methodically set. This method employs a distinctive design of orthogonal arrays to efficiently analyse the complete parameter set with a minimal number of experiments. The outcome of the experiments helps in the identification of optimum settings. The process of the Taguchi’s design of experiment approach is described in Figure 2. Using the set conditions, an experimental program based on the full factorial design would be too costly and arduous. Thus, the fractional factorial design using Taguchi’s orthogonal arrays was performed in MINITAB software to overcome the issue. In this study, four control factors were used, with three levels each. The DOE analysis reduced the number of combinations from 81 to just 9. The experimental design followed an L9 array, based on the Taguchi method. The reduction in the number of samples from 81 to 9 is possible through the use of orthogonal arrays. These arrays are specifically designed to efficiently explore the entire parameter space with a minimal number of experiments, while still providing crucial information about the impact of each factor on the outcome. In orthogonal arrays, the levels of each factor are “orthogonal” to the levels of other factors. This means that each level of any factor is paired with every level of the other factors an equal number of times. This ensures that the effects of each factor can be evaluated independently. The selected L9 orthogonal array of the fractional factorial DOE is summarised in Table 2, which will be carried out to investigate the inter-dependence of the parameters on the Mode I and II fracture toughness of the manufactured laminates. The orthogonal array means the processing parameters are weighted equally in the DOE. Therefore, each factor can be assessed independently.

Figure 2.

Taguchi’s method of parametric design approach.

Table 2.

Experimental testing trial based on design of experiments.

PANEL NO.	Deposition rate (MM/S)	HGT temperature (˚C)	TOOL temperature (˚C)	Consolidation force (N)
P1	76	850	20	180
P2	76	900	60	230
P3	76	950	100	300
P4	100	850	60	300
P5	100	900	100	180
P6	100	950	20	230
P7	124	850	100	230
P8	124	900	20	300
P9	124	950	60	180

After the Mode I and II experimental trials were conducted, a statistical analysis was conducted using MINITAB to determine the relative effects of each processing parameters. In order to do so, the signal to noise ( $S / N$ ) ratio firstly needs to be calculated for each of the experiment conducted. The term “signal” refers to the desired aspect of the output, which is close to its intended target value. On the other hand, the term “noise” refers to the undesired aspect and is a measure of the variability of the output. The Taguchi method utilises the signal-to-noise ratio ( $S / N$ ) to indicate the level of dispersion around the target value. A higher $S / N$ value indicates that the signal is significantly stronger than the random effects of the noise factors.^33,34 In terms of quality, there are three possible categories for the characteristics: (1) smaller is better, (2) nominal is better, and (3) larger is better. For this study, the objective is to maximise the Mode I and Mode II interlaminar fracture toughness, therefore the “larger is better” quality characteristic has been selected. The $S / N$ ratio is given as:

{(S / N)}_{i} = 10 \log \frac{{\bar{y_{i}}}^{2}}{s_{i}^{2}}

(1)

where the subscript

i

is the experimental number,

{\bar{y_{i}}}^{2}

is the mean value of the Mode I and II fracture toughness for a given experiment, and

s_{i}

is the experimental variance for a given experiment. The mean and variance are calculated as:

{\bar{y}}_{i} = \frac{1}{N_{i}} \sum_{u = 1}^{N_{i}} y_{i, u}

(2)

s_{i}^{2} = \frac{1}{N_{i} - 1} \sum_{u = 1}^{N_{i}} (y_{i, u} - \bar{y_{i}})

(3)

and

u

is the trial number. Since the work aims to maximise the response values of Mode I and II fracture toughness, the following

S / N

definition was calculated:

{(S / N)}_{i} = - 10 \log [\frac{1}{N_{i}} \sum_{u = 1}^{N_{i}} \frac{1}{y_{u}^{2}}]

(4)

The optimal value was then determined using the average values for $S / N$ ratio for all response at individual level. The higher obtained value indicates better properties.

Mode I fracture toughness

The Mode I interlaminar fracture toughness of AFP manufactured specimens were tested in accordance with ASTM D5528 standard.³⁵ From the manufactured panels, double cantilever beam (DCB) specimens with the nominal dimension of 195 mm × 22 mm with 80 mm of crack length were machined using a water-cooled diamond saw. Piano hinges of 20 mm length were adhesively bonded to the pre-cracked end of the specimens, as shown in Figure 3(a). This, then, provided an initial crack length, $a_{0}$ , of 60 mm that was measured from the loading point. The side of the DCB specimen was coated with a thin layer of brittle white paint to aid visual detection of the crack tip. A minimum of three specimens was prepared for each manufacturing condition. The Mode I experiment was performed on Instron universal testing with 5 kN load cell. A digital camera was used to aid the crack location and propagation. The specimens were loaded at the loading rate of 1 mm/min. The Mode I $(G_{I c})$ fracture energy was calculated using the modified beam theory and is expressed in equation (5):

G_{I C} = \frac{3 P δ}{2 b (a + Δ)}

(5)

where

a

is the delamination length,

P

is the applied load,

b

is the width of the specimen, and

δ

is the crosshead displacement. The term

Δ

is called the effective delamination extension which was experimentally determined using the least-squares plot of the cubed root of compliance (

C^{1 / 3}

) as a function of the length of delamination.

Figure 3.

Specimen configuration of (a) Mode I (DCB) (b) Mode II (ENF).

Mode II fracture toughness

The Mode II interlaminar fracture toughness test was carried out in accordance with the standard ASTM D7905.³⁶ The end-notched flexure (ENF) specimens were machined from the panels and had the nominal dimension of 195 mm × 22 mm with a 60 mm pre-crack, as shown in Figure 3(b). A thin coating of white paint was applied on the ENF specimen edge. A three-point bending static loading was applied on an Instron machine with a 10 kN load cell. The three-point flexure jig had a span of 100 mm. The support rollers have a diameter of 5 mm while the loading roller diameter was 9 mm. The ENF specimen was loaded at a rate of 0.5 mm/min.

The Mode II ( $G_{I I c}$ ) interlaminar fracture toughness was calculated using the compliance calibration method and is expressed in equation (6):

G_{I I C} = \frac{3 m P_{\max}^{2} a_{0}^{2}}{2 b}

(6)

where

a_{0}

is the initial delamination length,

P_{\max}

is the peak load prior to the propagation of the delamination,

b

is the specimen width, and

m

is the compliance calibration coefficient. The compliance calibration coefficient,

m

, was determined experimentally using a linear least squares regression analysis of the compliance,

C

, against the delamination length cubed,

a^{3}

. This is given as:

C = A + m a^{3}

(7)

In order to perform this analysis, the ENF specimens were marked with 20, 30, and 40 mm from the delamination tip. The ENF specimen was firstly loaded to a sub-failure load of 130 N at the delamination length at 20 mm (a_o-10 mm) and 40 mm (a_o+10 mm). The final loading was performed at 30 mm (a_o) until delamination propagation initiated. The compliance calibration coefficient was then determined at this known initial crack length. The process was repeated for the pre-cracked position (i.e., after crack propagation).

Degree of crystallinity

Degree of crystallinity was carried out using Differential Scanning Calorimetry (DSC) on DSC 204 F1 Phoenix. Each specimen was cut and weighed 13.6 g on average. The tests were performed with the heating and cooling rate of 10°C/min under nitrogen atmosphere up to 400°C. NETZSCH Proteus® software was used for data evaluation process.

Results

Experimental data analysis

The experimental measurements of $G_{I C}^{p}$ based on crack propagation values and $G_{I I c}$ fracture toughness based on pre-cracked values are illustrated in Figure 4. The highest Mode I fracture toughness was obtained in P3. The processing conditions for P3 were a deposition rate of 76 mm/s, HGT temperature of 950°C, tool temperature of 100°C, and consolidation force of 300 N. The highest Mode II fracture toughness was observed in P8. The processing conditions for P8 were deposition rate of 124 mm/s, HGT temperature of 900°C, tool temperature of 20°C, and consolidation force of 300 N. From the initial observations, it can be seen that improvements in Mode I fracture toughness favoured a slower deposition rate and a higher tool temperature while improvements in Mode II favoured a higher deposition rate and a cold tool temperature. This illustrates that an improvement in fracture toughness can be achieved with a higher consolidation force and process temperature. In P4, with a deposition rate of 100 mm/s, fracture toughness close to that of P3 was obtained. The lowest Mode I and II fracture toughness was observed in P7. P7’s processing conditions correspond to highest deposition rate. The low HGT temperature and low consolidation at this rate contributed to the lower fracture toughness. Moreover, a higher tool temperature is considered inappropriate for material properties. However, P8 and P9, manufactured at a deposition rate of 124 mm/s, with the difference in HGT temperature of 900°C and 950°C, tool temperature of 20°C and 60°C, and consolidation force of 300 N and 180 N, showed higher Mode I and II fracture toughness. The numerical values for $G_{I c}^{p}$ and $G_{I I c}$ are compiled in Table 3.

Figure 4.

Mode I, Mode II interlaminar fracture toughness and crystallinity content.

Table 3.

Numerical values of $G_{I c}$ and $G_{I I c}$ results using Taguchi’s DOE analysis.

Panel	Deposition rate (mm/s)	HGT temperature (˚C)	Tool temperature (˚C)	Consolidation force (N)	$G_{I c}^{p}$ (kJ/m²)	$G_{I I c}$ (kJ/m²)
P1	76	850	20	180	0.55	0.58
P2	76	900	60	230	0.92	1.99
P3	76	950	100	300	1.43	2.73
P4	100	850	60	300	0.96	3.07
P5	100	900	100	180	0.80	0.83
P6	100	950	20	230	0.78	1.55
P7	124	850	100	230	0.35	1.73
P8	124	900	20	300	0.54	3.72
P9	124	950	60	180	0.90	1.81

The amount of crystallinity is also summed up in Figure 4. The highest crystallinity content was found in P7, which is attributed to a higher tool temperature with a high consolidation force. During the lay-up, the cooling rates can be controlled by thermal conduction through the tool plate. Higher tool temperatures reduce thermal conduction, raising ambient temperatures, and slow down the cooling rates. As a result, this leads to higher degree of polymer crystallisation. In P1, a higher crystallinity can also be observed despite low HGT temperature. However, the slow deposition rate possibly resulted in the higher crystallinity. Similarly, the crystallinity content of 30% was obtained in P3 due to slow deposition rate with high HGT temperature. When combined with the slow deposition rates, high temperatures and large consolidation force, P3 exhibited one of the high fracture toughness. On the other hand, the fast deposition rate resulted in low crystallinity in P9. Comparative values were observed for P5, P6 and P8 crystallinity content, suggesting the crystallinity reliance on deposition rate and HGT temperature. These panels exhibited low crystallinity content with an increase in deposition rate. However, due to low consolidation force, it resulted in poor interlaminar properties.

Statistical analysis

Process parameters affecting responses

The effect of each processing parameters (in isolation) on Mode I fracture toughness is shown in Figure 5. The vertical axis is the average response of Mode I fracture toughness for each combination of control factor levels in a static Taguchi analysis. It helps in comparing the effect of various parameters with their relative strength. Increasing the deposition rate resulted in a reduction in Mode I fracture toughness. However, the difference was negligible in between 76 mm/s (level 1) and 100 mm/s (level 2). Increasing the HGT temperature shows an increase in the Mode I fracture toughness. Conversely, Mode I toughness was highest with a tool temperature of 60°C (level 2) and consolidation force at 300 N (level 3).

Figure 5.

Mean effects plot for Mode I ( $G_{I C}$ ).

The mean response for Mode II fracture toughness, in Figure 6, shows an opposite trend for the deposition rate parameter compared to Mode I. Increasing the deposition rate to 124 mm/s (level 3) yielded highest Mode II fracture toughness. There is negligible increase and then decrease in HGT temperature and tool temperature at level 2. Consolidation force has the most significant influence on Mode II fracture toughness among all influencing factors, since it has the widest range of values. An improvement in fracture toughness is observed with the increase of the consolidation force. It is observed that, at lower deposition rate, Mode I fracture toughness initially increased with an increase of HGT temperature, tool temperature and consolidation force. Mode II fracture toughness increases at level 3 of deposition rate, and consolidation force with HGT and tool temperature at level 2. By isolating each processing parameter, the response effect of the parameters on Mode I and Mode II properties can be observed. In the following section, the significance of each parameter can be analysed.

Figure 6.

Main effects plot for Mode II ( $G_{I I C}$ ).

Main effect plot for S/N ratio

The $S / N$ ratio of the processing parameters effect on Mode I, and Mode II fracture toughness are plotted in Figure 7(a) and (b), respectively. The higher $S / N$ ratio signifies that the level within the selected processing parameter has a higher effect on Mode I and II fracture toughness. For example, Mode I response favours slow deposition rate (level 1), high HGT temperature and consolidation force (level 3), and tool temperature at 60°C (level 2). While in case of Mode II, the favoured processing parameters are fast deposition rate (level 3), high HGT temperature (level 3) and consolidation force (level 3), and a tool temperature at 60°C (level 2). The influence of each factor on the responses was analysed and acquired through the response table of mean of $S / N$ ratio in Table 4. The larger $S / N$ ratio range implies the greater influence of a processing parameter on the respective measured properties (i.e. causing the most change in response). The most significant processing parameter for Mode I, according to the given ranking, is the HGT temperature, followed by deposition rate, tool temperature, and consolidation force. For Mode II, the most significant processing parameter is the consolidation force. Due to opening of the crack under Mode I, there needs to be high degree of polymer chain entanglement to resist the crack propagation. This typically requires high temperature to enable sufficient mobility of the chains when the polymer matrix is in a molten state. High consolidation forces are effective in improving the Mode II results due to reduction in interlaminar voids.

Figure 7.

Main effects plot for $S / N$ ratio (a) Mode I (b) Mode II.

Table 4.

Response table of $S / N$ ratio.

	Level	Deposition rate	HGT temperature	Tool temperature	Consolidation force
$G_{I c}$	1	−0.948	−4.944	−4.240	−2.698
	2	−1.469	−2.659	−0.685	−4.016
	3	−5.167	0.020	−2.659	−0.870
	Delta	4.219	4.964	3.554	3.146
	Rank	2	1	3	4
$G_{I I c}$	1	6.784	6.274	6.592	5.017
	2	6.243	5.931	9.04	5.843
	3	8.004	8.825	5.399	10.171
	Delta	1.761	2.894	3.641	5.154
	Rank	4	3	2	1

The optimal levels of the processing parameters are obtained by computing the average values of the $S / N$ ratios for each response at each level. The higher value of the $S / N$ ratios correspond to better characteristics. The optimum combination, obtained in this study for $G_{I c}$ are a deposition rate of 76 mm/s, HGT temperature of 950°C, tool temperature of 60°C, and consolidation force of 300 N. For $G_{I I c}$ , the optimum parameters are a deposition rate of 124 mm/s, HGT temperature of 950°C, tool temperature of 60°C, and consolidation force of 300 N. The methodology can also be extended to optimise other physical and mechanical properties, such as porosity content and interlaminar shear strength.

Analysis of variance

For the contribution analysis of each factor on the response, the analysis of variance (ANOVA) was used. In Table 4, this analysis shows the corresponding significance of process parameters to influence the response of Mode I and II values. The significance of a parameter depends up on least p-value. The p-value indicates the probability of a processing parameter influencing an outcome, which can then be used to determine the contributing factor of each processing parameter. A p-value with less than 0.05 shows that the factor is important and has significance. From Table 5, the HGT temperature is found to be influential factor affecting

G_{I C}

which contributed 33.8% and a p-value of 0.05, followed by deposition rate (26.4%). Consolidation force affects

G_{I I C}

most with the contribution of 81.8% and has a p-value of 0.004 followed by deposition rate (7.62%). The remaining parameters with less percentage of contribution and greater than 0.05 are less significant.

Table 5.

ANOVA for S/N ratio of mode I and mode II.

	SOURCE	Degrees of freedom	SUM of squares	Mean of squares	F	p-value	% contribution
$G_{I C}$	Regression	4	0.650	0.162	4.74	0.080
	Deposition rate	1	0.208	0.208	6.06	0.069	26.41
	HGT temperature	1	0.266	0.266	7.76	0.050	33.80
	Tool temperature	1	0.085	0.085	2.47	0.191	10.78
	Consolidation force	1	0.091	0.091	2.66	0.178	11.59
	Error	4	0.137	0.034			17.42
	Total	8	0.787				100.00
$G_{I I C}$	Regression	4	7.596	1.899	10.17	0.023
	Deposition rate	1	0.636	0.636	3.4	0.139	7.62
	HGT temperature	1	0.082	0.082	0.44	0.545	0.98
	Tool temperature	1	0.054	0.054	0.29	0.621	0.64
	Consolidation force	1	6.825	6.825	36.54	0.004	81.80
	Error	4	0.747	0.187			8.95
	Total	8	8.343				100.00

Regression analysis

The linear regression analysis was performed in MINITAB 21 to find the inter-dependent relationship of the processing parameters. The independent factors are Mode I and II fracture toughness. The inter-dependent factors are deposition rate, HGT temperature, tool temperature, and consolidation force. The regression model is used to determine the relationship between the processing parameters. The normal probability plots of the residual for both Mode I and II are shown in the Figure 8.

Figure 8.

Normal probability plot residuals for (a) Mode I and (b) Mode II.

The plot indicates the value of the response that lie close to the normal probability line. It can be seen that the residuals are adjacent to the normal probability line. The fitness of the regression model was checked by the coefficient of determination R² which is in between 0 and 1. R² stands for the coefficient of multiple determination and the variability reduction in the values of fracture toughness is attained by employing the regressor variables as deposition rate, HGT temperature, tool temperature, and consolidation force. The value closest to one is the indication of a good fit between the dependent and independent variable. In this study, R² values of 0.82 and 0.91 were obtained for Mode I and II respectively which meant that the regression model accounts for 82% and 91% of all variances for the respective outcomes. Therefore, this showed the data points are distributed close to the normal probability line which verify that the processing parameters are significant.

Therefore, from the residual plot, the inter-dependence relationship of the selected processing parameters for both Mode I and Mode II are given as:

G_{I C}^{p} = - 2.87 - 0.00775 v_{m a t} + 0.00421 T_{H G T} + 0.00297 T_{t o o l} + 0.00205 F

(8)

G_{I I C}^{p} = - 5.50 + 0.01356 v_{m a t} + 0.00233 T_{H G T} - 0.00236 T_{t o o l} + 0.0177 F

(9)

where

v_{m a t}

is the material deposition rate in mm/s,

T_{H G T}

and

T_{t o o l}

are the temperature of the HGT and tool respectively in ˚C, and

F

is the consolidation force in N. The experimental versus predicted values for Mode I and Mode II obtained from equations (8) and (9) are plotted in the Figure 9 with an average error of 2.58% and 1.14%, respectively. With the weighted functions of

G_{I C}

and

G_{I I C}

, it will be possible to perform a multi-objective study to determine a single optimum set of processing parameters for the desired Mode I and II fracture toughness.

Figure 9.

Linear regression prediction of inter-dependence of processing parameters for both Mode I and II.

Validation

To validate the analytical results, the validation experiments were conducted to confirm the outcomes of Mode I and Mode II predicted. The selected parameters corresponding to the largest

S / N

ratios. Based on the parameters and their levels tested, the average of the results from confirmation tests are compared with the predicted average. However, some manufacturing issues occurred while manufacturing with the Mode I process parameters. Fibrous material stuck onto the roller which impeded the consolidation process. This was possibly due to the combination of high nip-point temperature at lower deposition rate. Hence, the Mode I processing parameters had been modified to second optimum parameters as 100 mm/s, 950°C, 60°C, and 300 N. The processing parameters for the validation tests are summarised in Table 6.

Table 6.

Validation results for the optimum parameters.

Optimised output	Optimum condition for factors				Predicted values (kJ/m²)	Confirmation results (kJ/m²)
Optimised output	Deposition rate (mm/s)	HGT temperature (˚C)	Tool temperature (˚C)	Consolidation force (N)	Predicted values (kJ/m²)	Confirmation results (kJ/m²)
Mode I	100	950	60	300	1.14	1.18
Mode II	124	950	60	300	3.56	3.63

According to the Table 6, the predicted $G_{I C}$ and $G_{I I C}$ were close to the experimental results. The enhanced fracture toughness can be attributed to better consolidation and lower crystallinity in the polymer matrix. The driving mechanisms may possibly be linked to higher tool temperature with lower deposition rate and HGT temperature. The tool temperature facilitated a more even distribution of temperature in the substrate. In case of Mode II, the faster deposition rates could result in lower crystallinity and increased toughness. As Taguchi’s DOE is primarily a statistical analysis, further investigations are needed to uncover the physical mechanisms driving these improvements.

The inter-dependence equations for both Mode I and Mode II have been demonstrated to predict the resulting fracture toughness. However, it is important to note that Mode I and Mode II are interconnected parameters, so a single set of processing parameters must be utilised in the manufacturing of the specimens. This approach can also be extended to other mechanical or physical properties. Consequently, a multi-objective optimisation approach can be employed, utilising the inter-dependence equations, in order to optimise the processing parameters for two or more desired outcomes.

Discussion

The statistical analyses provide critical information on specific processing parameters that influenced Mode I and II fracture toughness of in-situ consolidated panels. The variance analysis revealed that HGT had the greatest effect on Mode I fracture toughness, followed by deposition rate. Increasing the HGT temperature results in a higher nip point temperature of the towpreg, while a slower deposition rate increases the amount of time the material spends under the consolidation roller at a specific nip-point temperature.²² This extended window allows for increased polymer chain diffusion across the interface, which has been enhanced by the intensity of heat flow over time. Consequently, this diffusion reduces amount of porosity across the interface, as shown in Figure 10(a) for P3. The cross-section of P7 in Figure 10(b) shows huge amount of intralaminar voids and inadequate distribution of fibres, found mostly within the tape. The cross-section of $G_{I C}$ -CP in Figure 10(c) depicts a uniform fibre distribution with the presence of resin rich areas and minor amount of porosity content.

Figure 10.

Cross-sectional optical micrographs of (a) P3, (b) P7 and (c) $G_{I C}$ -CP.

Figure 11(a) represents the DCB load-displacement curves of P7, P3, and $G_{I C}$ -CP which are the panel with the lowest, highest, and optimised Mode I fracture toughness, respectively. The laminates exhibited ‘stick-slip' crack growth. Under displacement control, delamination initiates and propagates until a sub-critical length is reached, where it can resist more load. Compared to epoxy composites, the stick-slip behaviour is more stable as evidenced by the smaller load drops.³⁷ The onset of deviation from linearity in P3 (high $G_{I C}$ ) and $G_{I C}$ -CP curve (optimum result) occurred close to the same load while for P7 (low $G_{I C}$ ), it occurred at a much lower load. This non-linear behaviour indicates an increase in Mode I delamination which is due to the formation of the plastic zone ahead of the delamination front as it propagates. P3 also displays small-scale fibre bridging (Figure 11(b)) behind the crack front. This can further impede the Mode I delamination growth. The porosity of P7 in Figure 10 seems to be more than that of P3 which indicates insufficient consolidation and poor mechanical performance of the panel. The amount of crystallinity is also higher in P7 as compared to P3 which suggested the presence of more amorphous polymer chains in P3 than in P7. The presence of resin-rich areas can be seen from optical micrographs of P3 (Figure 11(c)), indicating steady progress of crack through the matrix. However, due to fibre bridging, when the crack tip reached a tougher region, the propagation of this crack was further hindered until an excessive amount of elastic energy was produced.

Figure 11.

(a) Load-displacement for $G_{I C}$ of highest obtained value [P3], lowest value [P7] and optimum results [CP] (b) occurrence of fibre bridging in P3 (c) optical micrographs of P3.

The load-displacement curve for P8, P1, and the confirmed optimum parameters ( $G_{I I C}$ -CP) for the highest, lowest, and optimum $G_{I I C}$ values, respectively, is shown in Figure 12(a). The curve for P8 exhibits initial non-linear behaviour with the onset of steady crack propagation occurring at an early stage. This behaviour is characterised by lack of sudden load drops and is likely driven by plasticity effects at the crack tips.³⁸ Once the plastic zone, consisting of micro-cracks and damage (Figure 12(b)), fully develops, the micro damages coalesce with the main delamination and form a new delamination front. The curve for P1 also shows a high deviation from linearity before reaching the peak load. However, given that the poor $G_{I I C}$ performance, it is the authors’ opinion that this is due to poor consolidation of the panel (Figure 12(c)). In the $G_{I I C}$ -CP panel, the applied load increases linearly until the onset of crack propagation, which is a typical unstable propagation of Mode II cracking.

Figure 12.

(a) Load-displacement curve for $G_{I I C}$ at P8 (highest value), P1 (lowest value), CP (optimum condition) (b) Crack propagation behaviour in P8 during the test (c) Cross-sectional optical micrograph of P1.

The $G_{I I C}$ values of the P8 and CP panels are higher than the reported values of PEEK composites.²⁶ As shown in Figure 4, (P)8 panel has a crystallinity of 25% due to the higher deposition rates during manufacturing. The higher deposition rate results in a smaller window for polymer chains to crystallise leading to lower crystallinity content. However, this lower crystallinity content contributes to a tougher PEEK polymer matrix which translates to a higher $G_{I I C}$ value.

In this study, the effect of tool temperature on Mode I and II fracture toughness is inconclusive. There is no observable trend in the $S / N$ plot in Figure 7. However, it was found that the degree of crystallinity for P7, at highest tool temperature is 33% which is comparable to the hot-press consolidation (32.8%).³⁹ Intuitively, a heated tool plate will help the laminate to retain the heat from the HGT and extend the window above the melting temperature. This should result in a more even and higher crystallinity at high tool temperature.

Conclusion

The interdependent parameters involved in AFP manufacturing contribute significantly to the quality of laminates and their mechanical properties. To characterise these parameters, extensive experimental tests are often necessary due to their complexity. Optimising the process for enhanced fracture toughness of laminates requires an understanding of the cumulative effect of these parameters. This study examined the inter-dependence of material deposition rate, HGT temperature, tool temperature, and consolidation force and their influence on Mode I and Mode II fracture toughness. By employing Taguchi’s design of experiments to conduct an efficient analysis of the critical parameters, the number of test experiments was significantly reduced. The results were also analysed using analysis of variance (ANOVA) to isolate and investigate the contributing factors of individual processing parameters. The main findings of this work are listed as follows:

• HGT temperature and consolidation force are the most significant factors and directly impact Mode I and Mode II fracture toughness, respectively.

• A linear regression analysis, using MINITAB, determines an empirical formulation of the inter-dependence of processing parameters for both Mode I and II. The R² values for Mode I and Mode II were 0.825 and 0.91, respectively, which indicates a good fit.

• The optimal settings for Mode I, based on selecting the largest $S / N$ ratio, are 76 mm/s deposition rate, 950°C HGT temperature, 60°C tool temperature, and 300 N consolidation force. For Mode II, the optimal settings are 124 mm/s deposition rate, 950°C HGT temperature, 60°C tool temperature, and 300 N consolidation force.

• The validation of the analysis reveals that there was good agreement between the predicted fracture toughness for Mode I and Mode II and the validation test results. Therefore, the model developed from this analysis can be effectively utilised to predict the Mode I and Mode II toughness of HGT-assisted-AFP manufactured AS4/APC-2 composite.

• Insufficient consolidation is considered a factor contributing to decrease in the critical energy release rate in panels. Additionally, the presence of fibre bridging can hinder delamination growth in some panels.

The interpretation from the study contributes to a broader understanding of material behaviour under various AFP manufacturing conditions. The results can be used as the preliminary study for further mathematical modelling to control process parameters. The interdependent relations derived from the statistical analysis proved to be effective in predicting the interlaminar fracture toughness of specimens manufactured with optimum parameters. With this predictive capability, thermoplastic composite materials can be designed and fabricated with tailored fracture toughness properties without requiring full-factorial characterisation. However, it was found that the presence of porosity due to low interface bond strength causes poor toughness. Therefore, further investigation of the underlying physical mechanisms is required.

Footnotes

Acknowledgments

This work was supported by Defence Science and Technology Group (DSTG), Australia; the ARC Training Centre for Automated Manufacture of Advanced Composites (IC160100040), under the Australian Research Council’s Industrial Transformation Research Program. The authors also thank S. Mahadevan (UNSW) for technical assistance.

Author contributions

Shafaq: Investigation, Formal analysis, Writing – original draft. M.J. Donough: Conceptualisation, Formal analysis, Writing – Review & Editing. E. Oromiehie: Investigation, Writing – Review & Editing. F. Islam: Formal analysis, Writing – Review & Editing. N.A. St John: Funding acquisition, Writing – Review & Editing. A.W. Philips: Funding acquisition, Writing – Review & Editing. B.G. Prusty: Supervision, Writing – Review & Editing.

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: This work was supported by the ARC Training Centre for Automated Manufacture of Advanced Composites; IC16010040 and Defence Science and Technology Group.

ORCID iDs

Shafaq Shafaq

Ebrahim Oromiehie

B. Gangadhara Prusty

Data availability statement

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.*

References

Nguyen

Donough

, et al. Mechanisms of pop-up delamination in laminated composites pierced by the initial pure waterjet in abrasive waterjet machining. Compos Struct 2022; 297: 115968. DOI: 10.1016/j.compstruct.2022.115968.

Donough

Prusty

Van Donselaar

, et al. In-plane and oblique edge-on impact on thick glass-fibre/epoxy composite laminates. Int J Impact Eng 2023; 171: 104373. DOI: 10.1016/j.ijimpeng.2022.104373.

Saeedifar

Zarouchas

. Damage characterization of laminated composites using acoustic emission: a review. Compos B Eng 2020; 195: 108039. DOI: 10.1016/j.compositesb.2020.108039.

Allix

Ladevéze

Corigliano

. Damage analysis of interlaminar fracture specimens. Compos Struct 1995; 31: 61–74. DOI: 10.1016/0263-8223(95)00002-X.

Vieille

Casado

Bouvet

. Influence of matrix toughness and ductility on the compression-after-impact behavior of woven-ply thermoplastic- and thermosetting-composites: a comparative study. Compos Struct 2014; 110: 207–218. DOI: 10.1016/j.compstruct.2013.12.008.

van Ingen

Waleson

Offringa

, et al. Double curved thermoplastic orthogrid rear fuselage shell. In: SAMPE Europe conference nantes, Nantes, France, 17-19 September 2019, pp. 1–10.

Yao

S-S

Jin

F-L

Rhee

, et al. Recent advances in carbon-fiber-reinforced thermoplastic composites: a review. Compos B Eng 2018; 142: 241–250. DOI: 10.1016/j.compositesb.2017.12.007.

Donough

Shafaq

SJNA

St John

, et al. Process modelling of In-situ consolidated thermoplastic composite by automated fibre placement – a review. Compos Appl Sci Manuf 2022; 163: 107179. DOI: 10.1016/j.compositesa.2022.107179.

Brasington

Sacco

Halbritter

, et al. Automated fiber placement: a review of history, current technologies, and future paths forward. Composites Part C: Open Access 2021; 6: 100182. DOI: 10.1016/j.jcomc.2021.100182.

10.

Zhao

Donough

Prusty

, et al. Pseudo-ductile fracture in grid stiffened structure by automated fibre placement. Compos Struct 2023; 308: 116694. DOI: 10.1016/j.compstruct.2023.116694.

11.

Maung

Prusty

Donough

, et al. Automated manufacture of optimised shape-adaptive composite hydrofoils with curvilinear fibre paths for improved bend-twist performance. Mar Struct 2023; 87: 103327. DOI: 10.1016/j.marstruc.2022.103327.

12.

Gain

Oromiehie

Prusty

. Nanomechanical characterisation of CF-PEEK composites manufactured using automated fibre placement (AFP). Compos Commun 2022; 31: 101109. DOI: 10.1016/j.coco.2022.101109.

13.

Khan

Mitschang

Schledjewski

. Tracing the void content development and identification of its effecting parameters during in situ consolidation of thermoplastic tape material. Polym Polym Compos 2010; 18: 1–15. DOI: 10.1177/096739111001800101.

14.

Shadmehri

Hoa

Fortin-Simpson

, et al. Effect of in situ treatment on the quality of flat thermoplastic composite plates made by automated fiber placement (AFP). Adv Manuf Polym Compos Sci 2018; 4: 41–47. DOI: 10.1080/20550340.2018.1444535.

15.

Khan

Mitschang

Schledjewski

. Identification of some optimal parameters to achieve higher laminate quality through tape placement process. Adv Polym Technol 2010; 29: 98–111. DOI: 10.1002/adv.20177.

16.

Chen

. Understanding processing parameter effects for carbon fibre reinforced thermoplastic composites manufactured by laser-assisted automated fibre placement (AFP). Compos Appl Sci Manuf 2021; 140: 106160. DOI: 10.1016/j.compositesa.2020.106160.

17.

Brzeski

Holschun

Schledjewski

. Effect of tool temperature on laminate properties during in situ consolidation placement process. In: SAMPE international symposium Seattle. Seattle, WA, 17-20 May2010, pp. 17–20.05.

18.

Khan

Mitschang

Schledjewski

. Parametric study on processing parameters and resulting part quality through thermoplastic tape placement process. J Compos Mater 2012; 47: 485–499. DOI: 10.1177/0021998312441810.

19.

Oromiehie

Gain

Prusty

. Processing parameter optimisation for automated fibre placement (AFP) manufactured thermoplastic composites. Compos Struct 2021; 272: 114223. DOI: 10.1016/j.compstruct.2021.114223.

20.

Comer

Ray

Obande

, et al. Mechanical characterisation of carbon fibre–PEEK manufactured by laser-assisted automated-tape-placement and autoclave. Compos Appl Sci Manuf 2015; 69: 10–20. DOI: 10.1016/j.compositesa.2014.10.003.

21.

Qureshi

Swait

Scaife

, et al. In situ consolidation of thermoplastic prepreg tape using automated tape placement technology: potential and possibilities. Compos B Eng 2014; 66: 255–267. DOI: 10.1016/j.compositesb.2014.05.025.

22.

Schiel

Raps

Chadwick

, et al. An investigation of in-situ AFP process parameters using CF/LM-PAEK. Adv Manuf Polym Compos Sci 2020; 6: 191–197.

23.

Stokes-Griffin

Compston

. The effect of processing temperature and placement rate on the short beam strength of carbon fibre–PEEK manufactured using a laser tape placement process. Compos Appl Sci Manuf 2015; 78: 274–283. DOI: 10.1016/j.compositesa.2015.08.008.

24.

Heathman

Koirala

Yap

, et al. In situ consolidation of carbon fiber PAEK via laser-assisted automated fiber placement. Compos B Eng 2023; 249: 110405. DOI: 10.1016/j.compositesb.2022.110405.

25.

Ray

Comer

Lyons

, et al. Fracture toughness of carbon fiber/polyether ether ketone composites manufactured by autoclave and laser-assisted automated tape placement. J Appl Polym Sci. 2014; 132: 41643. DOI: 10.1002/app.41643.

26.

Oromiehie

Gain

Donough

, et al. Fracture toughness assessment of CF-PEEK composites consolidated using hot gas torch assisted automated fibre placement. Compos Struct 2022; 279: 114762. DOI: 10.1016/j.compstruct.2021.114762.

27.

Shafaq

DMJ

Donough

Farnsworth

, et al. Influence of deposition rates on the mode I fracture toughness of in-situ consolidated thermoplastic composites. Compos B Eng 2023; 251: 110474. DOI: 10.1016/j.compositesb.2022.110474.

28.

Antony

. Taguchi or classical design of experiments: a perspective from a practitioner. Sens Rev 2006; 26: 227–230. DOI: 10.1108/02602280610675519.

29.

Zaman

Boesch

Siadat

, et al. Impact of fused deposition modeling (FDM) process parameters on strength of built parts using Taguchi’s design of experiments. Int J Adv Manuf Technol 2019; 101: 1215–1226. DOI: 10.1007/s00170-018-3014-6.

30.

Heidari-Rarani

Ezati

Sadeghi

, et al. Optimization of FDM process parameters for tensile properties of polylactic acid specimens using Taguchi design of experiment method. J Thermoplast Compos Mater 2022; 35: 2435–2452.

31.

Khan

Irfan

Cantwell

, et al. Crack healing in infusible thermoplastic composite laminates. Compos Appl Sci Manuf 2022; 156: 106896. DOI: 10.1016/j.compositesa.2022.106896.

32.

Belhaj

Dodangeh

Hojjati

. Experimental investigation of prepreg tackiness in automated fiber placement. Compos Struct 2021; 262: 113602. DOI: 10.1016/j.compstruct.2021.113602.

33.

Roy

. A primer on the Taguchi method. Southfield, Michigan: Society of Manufacturing Engineers, 2010.

34.

Taguchi

Chowdhury

Taguchi

. Robust engineering. McGraw-Hill, 2000. https://cir.nii.ac.jp/crid/1130282270835758720.

35.

International

. Standard test method for mode 1 interlaminar fracture toughness of undirectional fiber-reinforced polymer matrix composites. West Conshohocken, Pennsylvania: ASTM International, 2007.

36.

ASTM D . Standard test method for determination of the mode II interlaminar fracture toughness of unidirectional fiber-reinforced polymer matrix composites. West Conshohocken, Pennsylvania: ASTM Standard, 2014.

37.

Donough

Gunnion

Orifici

, et al. Scaling parameter for fatigue delamination growth in composites under varying load ratios. Compos Sci Technol 2015; 120: 39–48. DOI: 10.1016/j.compscitech.2015.10.011.

38.

Hunston

Moulton

Johnston

, et al. Matrix resin effects in composite delamination: mode I fracture aspects. West Conshohocken, Pennsylvania: ASTM International, 1987.

39.

Mahat

Alarifi

Alharbi

, et al. Effects of UV light on mechanical properties of carbon fiber reinforced PPS thermoplastic composites. In: Macromolecular Symposia. Hoboken, NJ: Wiley Online Library, 2016, pp. 157–168.

PANEL NO.	Deposition rate (MM/S)	HGT temperature (˚C)	TOOL temperature (˚C)	Consolidation force (N)
P1	76	850	20	180
P2	76	900	60	230
P3	76	950	100	300
P4	100	850	60	300
P5	100	900	100	180
P6	100	950	20	230
P7	124	850	100	230
P8	124	900	20	300
P9	124	950	60	180

PANEL NO.	Deposition rate (MM/S)	HGT temperature (˚C)	TOOL temperature (˚C)	Consolidation force (N)
P1	76	850	20	180
P2	76	900	60	230
P3	76	950	100	300
P4	100	850	60	300
P5	100	900	100	180
P6	100	950	20	230
P7	124	850	100	230
P8	124	900	20	300
P9	124	950	60	180

Parametric process optimisation of automated fibre placement (AFP) based AS4/APC-2 composites for mode I and mode II fracture toughness

Abstract

Keywords

Introduction

Experimental method

Material and manufacturing

Selection of process parameters

Design of experiments

Mode I fracture toughness

Mode II fracture toughness

Degree of crystallinity

Results

Experimental data analysis

Statistical analysis

Process parameters affecting responses

Main effect plot for S/N ratio

Analysis of variance

Regression analysis

Validation

Discussion

Conclusion

Footnotes

Acknowledgments

Author contributions

Declaration of conflicting interests

Funding

ORCID iDs

Data availability statement

References

PANEL NO.	Deposition rate (MM/S)	HGT temperature (˚C)	TOOL temperature (˚C)	Consolidation force (N)
P1	76	850	20	180
P2	76	900	60	230
P3	76	950	100	300
P4	100	850	60	300
P5	100	900	100	180
P6	100	950	20	230
P7	124	850	100	230
P8	124	900	20	300
P9	124	950	60	180