Experimental investigation for machinability aspects of graphene oxide/carbon fiber reinforced polymer nanocomposites and predictive modeling using hybrid approach

Jogendra Kumar,Rajesh Kumar Verma

Department of Mechanical Engineering,Madan Mohan Malaviya University of Technology,Gorakhpur,273010,India

Keywords: Graphene oxide Polymer Carbon fiber RSM ANOVA GR-PCA

ABSTRACT This article explores the drilling behavior of polymer nanocomposites reinforced by Graphene oxide/Carbon fiber using a hybrid method of Grey theory and Principal component analysis(GR-PCA).An online digital dynamometer was employed for the evaluation of Thrust Force and Torque.The image processing technique computes the delamination.Response surface methodology (RSM) considers the parameters,namely,drilling speed(S),feed rate(F),Graphene Oxide wt.%(G)in designing the experimentation array.Principal component analysis (PCA) was used to tackle the response priority weight during the combination of multiple functions.Analysis of variance (ANOVA) scrutinized the influence of parameters and intended the regression models to predict the response.GR-PCA evaluated the optimal parametric setting and remarked that feed rate acts as the most predominant factor.The higher feed rate and wt.%of G is responsible for surface damages like fiber pull-out,fiber fracture and cracks.A significant improvement in drilling responses has been obtained and also validates through confirmatory test and microstructure investigations.

1.Introduction

Nowadays,Polymer nanocomposites efficiently replacing the traditional engineering materials due to its reduced weight and improved mechanical features.From the last two decades,Carbon/polymeric composites play a significant role in aerospace components,automobile,aircraft,bio-device,sensors,electric boards,etc.[1].It is possible due to exceptional properties,such as corrosion resistivity,ultralightweight,high stiffness and bulk production aptness.These materials can be made at various levels and patterns to achieve different mechanical characteristics for a variety of products.In this modern era,the manufacturing sector is proliferating due to advanced machinery,equipment,and,most important is the varying demand of the customer.Due to the different choices of the users,industries have to adopted accordingly.In the last few years,new trends are adopted by the polymer manufacturing sector by doping or reinforcing the nanomaterials into the matrix phase.It enriches the mechanical characteristics of the polymer composites in a significant way.The nanofillers materials can be with multiple dimensions (D),i.e.,1D,2D,and 3D.The carbonbased nanofillers like Graphene oxide (GO),Carbon nanotube(CNT),Carbon nano onions(CNO),Carbon nanorods(CNR),reduced GO,etc.are extensively used in carbon-based polymer composites.It possesses superior properties such as exceptional shape &size,aspects ratio,surface area and other physical and mechanical features[2].Hence it is recognized as the most responsive nanofillers with polymer matrix[3,4].Graphene is a single layer carbon atom arranged into hexagonal cells of wax known as a honeycomb of the cave.The measurable factors are superlative,such as strength,stiffness,elasticity,thermal and electrical conductivity[5-7].It has greater carrying mobility with 10000 cm2/VS ambient temperature,specific surface area (2630 m2/g),optical visibility (97.7%),high young module (1 TPa) and an exceptional conductivity(3000-5000 W/mK) [8-11].Due to these extremely excellent properties,it can extend the maximum up to 1/4th of its original[12].

Some studies show that the rise in the mechanical and electrical features is not only due to Graphene nanofillers,but it also depends upon the matrix material.It has been remarked that most of the studies preferred the epoxy material as a matrix phase for the development of nanocomposites[13-15].The intensity of dispersal of Graphene oxide into the epoxy matrix,point distribution and surface adhesiveness of GO/epoxy also compete for a substantial role in the composite strength [7,16].The carbon fiber/polymer composites replaced the traditional engineering metallic alloys due to improved features.Sometimes these improved properties play a sophisticated role during the machining of polymeric materials.The non-homogeneity and anisotropic conducts make them different from existing materials.Hence,the machinability phases of carbon fiber/polymer materials are remarkably different from a metallic material and their alloys.Drilling is regarded as the extremely demanding,indispensable machining procedure in manufacturing trades for the assembly of the major and minor structures.However,fiber strength in the machined surface is weakened when the polymers are drilled.The drilling-induced damages/faults,such as cracks,debonding,fiber fracture,fuzzing,and delamination,are created in plastics.The choice of suitable drilling constraints can control these inevitable damages [17,18].These defects and damages inspections are feasible only by highresolution magnification devices and practices.Such types of drilling damages significantly decline the overall efficacy of polymer material and also disrupt the service life and load capability[19].In the aviation industry,the most accountable factor for refusal of elements during production of the final composition is the delamination and its related damages,about 60%of total refusal is carried by delamination only [18,20].Hence,trades are too much apprehensive about delamination,and it turns out to be an issue of interest and investigation.The efficiency of any component predominantly depends upon the surface and machined hole quality[21].The crucial factor for minimizing the delamination and surface damages during the manufacturing of fiber/polymer composites is viable through the cutting force control.It is possible through the monitoring of feed rate,spindle speed and nanofiller wt.% combination [18].During manufacture,it is important to examine the kinetics of drilling operations affecting the cutting tool action[22].

The various eminent scholar has explored the machining conduct of Carbon fiber/polymers and proposed different tool and models to achieve a favorable machining environment.In this series,Miyake et al.[23]explored the drilling of uni-directional glass fiber reinforced polymer (UD-GFRP) and estimated the generated stress through micro-Raman spectroscopy.Heidarya et al.[24]planned a Taguchi concept based drilling of woven E-glass/MWCNT polymers to examine the Thrust,delamination,and residual flexural strength using Grey theory.The study revealed that feed rate is the prominent factor for the control of Thrust force and delamination.Quan et al.[25]evaluated the impact of drill speed and feed rate,as well as the geometry during MWCNT polymer drilling.The conclusions of the study divulge that at the exit of the hole,more damages were found.Karatas et al.[26] carried out an investigational study for the prediction of the optimum parametric condition during the abrasive water jet process of carbon/polymer composites.The planned works targets to lowers the average surface roughness and delamination factor.In this,experimental design finalized with three different process parameters and levels under the TaguchiL27orthogonal array.Variance analysis(ANOVA)conducted to verify the process performance and to test the model adequacy.They found that water pressure and hole diameters have a more significant effect on average roughness of the surface and were affecting delamination by water pressure and stand distance.The Abrasive jet machining is found feasible in the machining of polymers [27].Another machining study of Karatas? et al.[28] performed on carbon polymer materials in which the fiber orientation was selected from three different angles.Taguchi concept based L16orthogonal design employed for drilling procedure and combined approach of Grey model and Taguchi concept was proposed for parametric appraisal.Kumar[29]attempted the experimental work according to central composite design (CCD) based on RSM to examine the machining damages and subsurface defects during the GFRP composite drilling.The results concluded that the feed rate is the most affecting factor for delamination and surface roughness during GFRP drilling.Also,the response diagrams and threedimensional interaction plots used to elaborate on the effect of parameters.Ragunath et al.[30]proposed an efficient optimization method to assess delamination during machining of the fiber/polymer composites.The study validates that feed rate is the most leading factor amongst all the constraints such as applied load,sliding distance,drill bit diameter,point angle,and a chisel edge.

In the optimization of different process constraints,many attempts were made by eminent scholars to evaluate the optimal conditions for the efficient manufacturing environment.Gopalsamy et al.[31] explored the machining phases of hardened steel using a Grey model-based optimization tool.This study was planned to obtain the desired values of the material removal rate(MRR),surface roughness (Ra),tool wear rate (TWR).Variance analysis estimates the effect of the dominant factor during hard machining.Kumar et al.[32] established a robust module to optimize varying cutting parameters and production circumstances.For this,experimentation was organized according toL18orthogonal array used and Grey based Taguchi technique was used to optimize the conflicting responses.PCA was used to assess the response weight values for the aggregation of multiple machining performances.A comparative model was created between the experimental outcomes and the result of the confirmation.GRA method is broadly utilized to assess the optimal combinations of machine constraints for contradictory machining response analyses during traditional manufacturing processes.Besides this,it is also widely used by eminent scholars in advance manufacturing operations such as welding [33],casting [34],milling [35],turning [36],etc.PCA method is a useful statistical tactic primarily used to assess the response priority weight having conflicting characteristics.It is also extensively used in the weldability assessment and joining process[37,38],Wire electric discharge machining (WEDM) [39,40],turning [41],etc.for weight assessment during aggregation of multiples output.Earlier,the combined approach of GRA-PCA was employed for quality enhancement and multi-criteria decisionmaking case studies of manufacturing industries [42,43].The application,as mentioned earlier,shows the feasibility of the GRAPCA method in manufacturing science and technology.

A comprehensive literature review shows that pioneer researcher did ample work for the synthesis and development of Graphene oxide nanomaterial.The existing data demonstrate the feasibility and application potential of Graphene Oxide in engineering applications such as bio-medical,sensors,battery,structure applications of automotive and aircraft.The CFRP materials efficiently substitute the heavyweight metals and their alloys,as it possesses lightweight,enhanced strength and durable,costeffective properties.It has been observed that limited data exist on the mixing of Graphene Oxide epoxy/Carbon fiber polymer nanocomposites.Literature reveals that eminent scholars staged ample efforts for machining of CFRP by considering conflicting machining performance such as MRR,Ra,Cutting force,Tool wear,etc.and various optimization tools such as desirability function analysis(DFA),GRA,Technique for Order of Preference by Similarity to Ideal Solution(TOPSIS),Utility,Taguchi.But very limited work is available on machining aspects of Graphene oxide doped epoxy/CFRP composites and simultaneous optimization of machining characteristics of these novel materials.Also,during multiple machining performance optimizations of contradictory responses,it is found that most of the studies assume negligible response correlation and uniform priority weight assignment,which is not possible in real practice and leads to ambiguity,inaccuracy,and error in the solutions.The prior work shows that the Grey concept is widely accepted by the manufacturing science researchers in the optimization of machining parameters.But,still,there is a shortage of robust optimization module to handle the priority weight issues of the conflicting responses.These critical issues have been efficiently tackled in this research by the exploration of the GRA-PCA hybridization approach.The machinability aspects are passing through an opening stage and work is not satisfactorily explored to fully utilized the exceptional properties of these novel nanocomposites.Therefore,it can become a prospective field of research,development and commercialization for practicing engineers,polymer sector and academia.In this study,tests performed to appraise the drilling efficiency of Graphene oxide modified epoxy and Carbon fiber reinforced polymer composite plates using the GR-PCA hybrid approach.The ability of the machined sample is improved by using the Titanium aluminium nitride (TiAlN) SiC coated tool.The varying process parameters,namely,spindle speed,feed rate,and wt.% of G was considered to obtain the desired machining characteristics,i.e.,thrust force (Th),Torque (T),and delamination factor (IN and OUT).Present work targets to discuss the experimental results and minimize the unavoidable defect and damages during drilling of polymer nanocomposites.The outcomes of the work demonstrate the application potential of the proposed GR-PCA hybrid approach in a manufacturing environment.An effort aimed to introduce a robust optimization tool that can overwhelm the drawbacks of the prevailing optimization techniques.

2.Experimentation and methodology

2.1.Development of composite samples

The machining of Graphene Oxide/Carbon Fiber (GO/CF) polymer nanocomposites having a thickness of 10 mm was drilled using TiAlN SiC coated drill tools with a diameter of 5 mm.The specimens were fabricated by using Graphene oxide (Mesh size 200),CFRP prepregs uni-directional (UD),epoxy resin-520 supplied by M/s.Platonic Pvt Ltd Jharkhand,India and hardener-D from M/s.CF composites,Delhi India.The Graphene Oxide strengthened polymer composite made by modified epoxy resin using the hand layup method (Graphene by wt.% of 1,2 and 3%).The sample size was taken as 100×100×10 mm.Table 1,shows the mechanical properties of the proposed GO/CFRP nanocomposites.

Table 1 Mechanical properties.

Graphene oxide nanosheets at a high resolution of 500 μm were characterized by transmission electron microscopy (TEM).Fig.1 displays TEM images of graphene oxide.In general,the nanosheets of graphene oxide producing multilayer agglomerates appear to congregate together.The sizes of the individual nanosheets range from tens to several hundred square nanometers.Since,GO sheets 2-D in a blending state are thermodynamically stable.The findings show that,even after higher rates of oxidation,the graphite stacking order is retained in the lattice and few interlamellar spacings.It indicate that even after higher oxidation rates,the graphite stacking order is preserved in the system.The graphite sheet layered a usual ring,showing the polycrystallineexistence.

Fig.1.TEM analysis of Graphene oxide.

X-ray diffraction (XRD) method is primarily applied to determine the level of the graphitization and to provide knowledge on the degree of nanoparticles[44].The structures of Graphene Oxide were examined with XRD,as presented in Fig.2.The outcome revealed that the graphite diffraction peak was 26.54°,lower peaks at 43.2°,which indicates the graphite plane existence [45,46].The spacing of the interlayer was 0.34 nm,measured from peak level using the Bragg equation [9,47].Due to a particular oxygen group on the graphite surface when the graphite becomes oxidized,which contributed to an increase in interlamellar spacing [7,9,13].

2.2.Experimentation

The drilling was executed on a vertical Computer numerical control (CNC) machine with a maximum 10000 rpm,Model No BMV35 TC20 (Fig.3).The three parameters and three levels have been considered,according to Table 2.The drilling was carried out under the RSM based box Behnken design (BBD),as reported in Table 3.The Drill tool having a point angle of 135°and a helix angle of 30°,as displayed in Fig.4.

Table 2 Process parameters.

Table 3 Box Behnken Design and corresponding observed data and standard deviation.

Fig.3.Machining setup.

Fig.4.TiAlN (SiC coated) drill bit.

2.3.Response measurement

The online thrust force was assessed using the CNC dynamometer (MEDILAB MLB-PML 300),as shown in Fig.5.The computerized data collection system gathers and monitors the experimental data.A connecting cable type attaches the dynamometer to a 3-channel load amplifier device,which is connected by a pin cable to the computer.The tests have been performed carefully to prevent any chance of experimental failures.

Fig.2.XRD analysis of graphene oxide/carbon fiber nanocomposites.

Due to the difficulty of gauging the arbitrary area affected by the hole,a few studies conducted over the study of the delamination factor[30,48,49].Eminent scholars have used different techniques for the identification and evaluation of hole damage during drillingof composites.The delamination value was measured on a highresolution (0.7×) vision setup (Model no.SDM -TRZ-3D),as shown in Fig.6.This figure illustrated the damage caused by the delamination of the nanocomposite.The delamination factor at Inlet and Outlet is computed by using Eq.(1),i.e.,the ratio of the maximum diameter of the damaged area to the nominal hole diameter[50,51].

Fig.5.Thrust force measurement setup.

Fig.6.Vision measurement setup.

Here,Fd-delamination factorDmax-damaged maximum diameter of the machined hole,Dnominal-Nominal diameter of the machined hole.

2.4.Grey embedded principal component analysis (GR-PCA)approach

Grey embedded principal component analysis (GR-PCA) theory mainly used in multi-criteria optimization case studies consist of uncertainty,discretion,vagueness,etc.in their outcomes.Hence,it acts as an effective way to analyze many variables with some information.The principal component theory measures the real weight assignment to evaluate the aggregation among multiperformances [43,52,53].The GR-PCA technique evaluates the optimum setting of process parameters for a significant improvement in drilling characteristics[54].The parametric conditions show the minimum thrust force,Torque,Delamination IN and Delamination OUT values.In this way,using the GRA method,the multi-objective issue has been altered into a single-objective function known as Grey relation grade (GRG).The following are the basic steps of the Grey-PCA theory.

Step 1.Normalization data.

Data pre-processing methods are mainly dependent on the required characteristics of the information sequence ’’lower-thebetter-LB′′criteria was performed by using the below equations[53,55].

Step 2.Deviation sequence.

However,the deviation value is calculated using the equation:

WhereXi（k）represents the original series,minxi（k）is the deviation.

Step 3.Grey relation coefficient.

The GRA process evaluated the significance between the two structures and sequences.GRC（ξ）has been computed using the expression:

Where Δidenote the current value of the sequence and deviation sequence（?）betweenXi（k）andXj（k）,generally（?）=0.5 is used[42,56].

Step 4.Weight calculation.

Response priority weight is calculated after the normalization procedure by the exploration of the PCA method.Computation of Eigen value and Eigen vector

Where,γk=eigen value andk=1,2,3,4 … ….N.

Calculation of PC values

Where,Ym1，Ym2，……… are first,second principal component etc.The principal components adapted for variance in the decreasing order,and therefore the first major elementYm1，is the most variance of the data.

Step 5.Grey relation grade.

After the computation of weight,Grey-PCA value is determined using the equation:

3.Results and discussion

The machinability analysis of the GO/CF polymer nanocomposites is investigated for drilling characteristics,namely the delamination factor,Thrust force,and Torque.The parametric appraisal was done by intending different mathematical models using the GR-PCA technique.ANOVA demonstrates the role of the prominent factor affecting machining performances.

3.1.Analysis of variance test

ANOVA test has been implemented at a 95%confidence level to check the influence factor on machining performance [57,58].Fisher test (F-value) has been carried out to check the maximum value machining parameters indicating most influencing significant affecting factors on the process performance.On the other hand,theP-value indicating if the process parameter is less than 0.05 than the factor considered as significant.The variance analysis for Thrust (Th) and Torque (T) are detailed in Tables 4 and 5,respectively.The most significant factor for thrust force is spindle speed and feed rate(Table 4).In Table 5,Torque is primarily affected by feed rate (84.54%) and trail by spindle speed (8.85%),wt.% of Graphene oxide (3.58%) significant.From Tables 6 and 7,thedelamination factor at the entrance is primarily affected by feed rate(51.08%),spindle speed(26.89%)followed by wt.%of Graphene oxide is(8.39%)and at exit primarily affected by feed rate(59.08%),spindle speed (20.65%) followed by wt.% of Graphene oxide is(11.64%),respectively.The model adequacy is found as satisfactory for further investigation.The result obtained from the analysis of variance concludes in this study can be forwarded for further analysis to predict the optimal setting.It can enhance the hole quality in the form of reduced Thrust,Torque and damages(delamination) during drilling of GO/CF polymer nanocomposites.Hence,the ANOVA model has been established in this study.

Table 4 ANOVA for Thrust force.

Table 5 ANOVA for torque.

Table 7 ANOVA for the Delamination factor OUT.

3.2.Nonlinear mathematical model

The quadratic mathematical models are used to the developed correlation between independent and dependent process parameters.Eqs.(9-12)shows the regression analysis generated by using RSM technique:

Here,Th-thrust force(N),T-torque(Nm),Fd In-delamination at inlet,Fd Out-delamination at outlet,S-spindle speed(rpm),F-feed rate(mm/min) andG-wt.% of Graphene oxide.

3.3.Influence of process parameters

The thrust force(Th)and Torque(T)performs an essential role in producing a high-quality machined product and also ensuring the machinability of the polymer composites.During drilling of GO/CF epoxy composites,ThandTwere evaluated online through dynamometer-CNC attachment.From ANOVA analysis ofThandT,it has been found that the combination of feed rate and spindle speed plays a key role during the drilling process.Hence,In this section,the Thrust and Torque generation plot is presented.Fig.7(a-c) illustrates the development ofThandTduring the drilling ofG-1%,G-2% andG-3% doped nanocomposites,respectively.The drilling procedure consists of three phases to create a through-hole.In the first stage,the workpiece touches and starts to penetrate,and a constantThvalue is achieved,as shown in Fig.7(a-c).During the second phase,when the full cutting edges come in contact with the workpiece highest peak point achieved a higher value of thrust force.

Fig.7.Thrust and Torque signals at (a) S-2400, F-160 G-1% (b) S-2400 F-80 G-2% (c) S-1600 F-240 G-3%.

At last,exits out of or splits across the workpiece takes place in the third phase of the drill and then again constant thrust force generated.The third phase shows the less values of the Th.It is mainly due to the less contact of workpiece and tool at this phase;also,the machining is to end at this step[48].From the experiment result in it shows that the thrust force induced by drilling was a minimum of 53.94 N at minimum feed rate (F-80 mm/min) and a maximum of 100.81 N at maximum feed rate (F-240 mm/min)similar for Torque minimum of 0.06 Nm to 0.17 Nm.

In most of the studies,drilling-induced delamination was mainly caused due to the effect of thrust force and Torque [59].In general,the machine tool failure takes place through composites when higher thrust forces and Torque is engendered during drilling[60].It can be regulated through proper assortment of process constraints such as spindle speed and feed rate [61].Fig.7(a-c)displays that the Thrust and Torque force effects significantly depend on the feed rate,also concluded in Table 3.The higher value of feed rate forces the drilling tool to push inside the layer rather than to cut it.Hence,a higher rate of feed rate raises the thrust force and Torque.Most of the related investigations demonstrate that feed rate is the most prominent factor influencing the machining ability of the polymer composites [62,63].The impact on the cutting tool edges is very high against the abrasive fibers;in turn,the resistance force rises,i.e.,the rubbing of drill tool interfaces with the inner surface of the hole increases.It is mainly due to amplified thrust force [64,65].

Fig.8(a and b) shows the interaction plot between thrust force and process parameters.From Fig.8(a and b),it has been noted that the feed rate increases,even if there is a difference in the thrust force rates until all of the cutting edges reach the workpiece,the graph section where the highest thrust force (state of the thrust force graph)occurs is similarly inclined.The main reason for this is an increased feed rate,which leads to more vibration of the machining component,hence higher thrust force[66].From Fig.9(a and b),the torque value declines with a rise in spindle speed,and higher feed rate values also influence it.The combination of higher feed rate and the lower spindle speed is provided,then Torque again increases[67].In addition,even if the measured thrust forces and Torque differ a little,the processing time depending on the drilling condition was established at the feed rate used in this study at lower (80 mm/min) and higher (240 mm/min).

The manufacturing industries of CFRP nanocomposites are seriously concerned about the drilling-induced delamination.The damage round of the hole was calculated using the Microscope through the image processing technique.The damage was found at the entrance sides during the drilling of polymer materials.Drilling of CFRP laminated nanocomposite materials leads to obvious mechanisms of damage,i.e.,peeling off.Thrust force (Th) is the dominant factor that causes damage during drilling [68].Delamination(IN)takes place when the drill cutting edge touches over the laminate.Due to this,in turn,the nibbled lamina peels off with drill flutes,leading to peeling off of the upper lamina.The uncut thickness of the laminates on the bottom edge decreases so that the uncut layers are elastically bent due to the compressive strength of the drill bit [69].

Fig.8.Interaction plot for thrust force, v/s (a) Spindle speed (S),(b) Feed rate.

The process parameters affecting the delamination factor (IN)are displayed through Fig.10(a and b).From Fig.10(a and b),indicates the decrease in the delamination factor at the entrance side by adding CFRP nanocomposites with the wt.%ofG.Furthermore,as shown,the delamination at the entrance decreases with increase in spindle speed.At 2 wt.%Gdoped Nano-CFRP at 2400 rpm,the minimal delamination factor (1.012) at the entry was noticed.The addition of GO improves the interfacial strength between reinforcement and epoxy.It provides the lubrication during the cutting edge of the tool and chip interface,which in turn reduces the friction;it requires less cutting force[70,71].Fig.11(a and b)show the role of feed rate during the machining,and it has also been noted that the higher values of feed rate increase the exit side delamination factor in a significant way.The higher feed rate raises the thrust force that can bend the uncut elastically and increase the bending stress on the laminates.It will reduce the ILSS,and the chances of laminar splitting are there[72].The inter-laminar bond failure causes cracks around the hole to develop and spread as the drilling bit further pushes down.The addition of GO in epoxy can increases the bending strength and the ILSS [51,73,74].

3.4.Multi-response optimization using GR-PCA hybrid approach

Fig.9.Interaction plot for Torque, v/s (a) Spindle speed (S),(b) Feed rate.

In this article,the machinability aspects and process parameter optimization in the drilling of GO/CF reinforced composites are performed.The Grey theory aggregates the multiple responses and PCA assigns the response priority weight.Most of the pioneer scholars assume the uniform response weight during the aggregation of responses.Such theory creates the inaccuracy and ambiguity in the outcomes and finally deviates the efficiency of the module.In the previous section 2.4,the basic steps of the proposed approach have been discussed.The GR-PCA approach initially normalizes the observed data using Eq.(2).as presented in Table 8 to decrease the variability among conflicting machining responses.It is also known as a data pre-processing step.Pre-processing information indicates moving of the initial sequence to a similar order,thus normalizing the observed data from zero to one according to desired criteria.Using Eq.(3) to evaluate the deviation sequences,as depicted in Table 8.Using Eqs.(6) and (7) to calculate the Eigen values and its proportion,as mentioned in Table 9.The Eigen vector values and weight of the PCs are displayed in Table 10.The real weight assigns to the Grey relation coefficient (GRC) for estimation of GRC value using Eq.(8) for computation of the objective function of the proposed GR-PCA method (Table 11).

Fig.10.Interaction plot for Delamination at Inlet, v/s (a) Spindle speed (S),(b) Feed rate.

Fig.12 displays the graph among the predicted values for GRPCA.The effect of drilling parameters examined using response diagrams.The result of input parameters over GR-PCA is found as spindle speed at 2400 rpm,feed rate at 80 mm/min and wt.% of Graphene oxide 1%,which shows the minimum values of thrust force and torque delamination factor(INandOUT).The parametric combination shows that 1% of GO is sufficient to improve the desired drilling characteristics and a combination of higher spindle speed and low feed rate is preferably used for low values of Thrust and Torque.The higher value of feed rate raises the thrust force.Due to the higher feed rate,the force of drilling push on the job sample is higher,which creates a more considerable bending that can reach to crack initiation [75-77].The high spindle speed can soften the matrix phase,which increases the temperature between the machining interface(tool cutting edge and workpiece),in turn,necessitates lower cutting force [21,63].It generates less drillinginduced damages like cracks,fiber fracture and debonding of the matrix.A similar trend was observed in previous work [30,78,79].The Graphene oxide doped CFRP epoxy composites consist of Graphene oxide modified epoxy and carbon fiber in polymer nanocomposites.The delamination and fiber breakage may occur during the drilling process.It can be overwhelmed by the assortment of appropriate process constraints and the addition of GO wt.%.From Table 11,it has found that the highest value of GR-PCA is 1,which corresponds to the orthogonal array setting asS-2400 rpm,F-80 mm/min,andG-2% in the designed experimental array.The main effect plot for the S/N ratio of GR-PCA shows in Fig.12.The optimal condition from Fig.12 are found asS-2400 rpm,F-80 mm/min,andG-1%.

Fig.11.Interaction plot for Delamination at Outlet, v/s (a) Spindle speed (S),(b) Feed rate.

Table 8 Normalize data and Deviation sequence.

Table 9 Eigen Values and proportions for PCs.

Table 10 Eigenvectors for PCs and contributions.

Table 11 GRC and corresponding GR-PCA.

Fig.12.Optimal setting by GR-PCA.

The GR-PCA means analysis reported in Table 12,and the values in the experimental array lead to the response table at each level and their respective means.The feed rate of the machining parameter shows the highest delta value of 0.381,which affects GRPCA remarkably.The second influencing variable is spindle speed trailed by wt.% of G with a delta value of 0.256,and 0.149,respectively.

In order to find out the nonlinear regression between process parameters and drilling performances,a response surface methodology (RSM) model is established [80,81].The effects of the varying parameters on GR-PCA were determined by a quadratic model of nonlinear regression analysis.The nonlinear regression proposed that the error between the experimental and predicted result is determined using Eq.(13).In Table 13,it is shown that the maximum prediction error for the hybrid GR-PCA module is 6.897%and the average error percentage is 2.985%.The accuracy of the GRPCA hybrid module is observed as satisfactory.Finally,it can be concluded that the proposed GR-PCA hybrid model effectively tackled the multi-attribute decision-making issue in the drilling of polymer nanocomposites.

3.5.Confirmatory test

The confirmatory test was staged to check the application potential of the proposed hybrid module results.Table 14 shows the optimal combination used in the confirmatory test,while experiment number 2 (Table 11) displays the optimal setting in the experimental array.The confirmatory experiment was conducted to confirm the validation of the predicted result using the GR-PCA approach.The predicted GR-PCA can be calculated using the optimum machining setting as:

Table 12 Response table for means.

Table 13 Error calculation for GR-PCA module.

Here,ξ（k）=GR-PCA mean value;ξ0=mean GR-PCA at optimum level andn=number of experiment parameters.It is noticed that the Thrust force is lowered from 53.94 to 46.07 N,Torque decreased from 0.06 to 0.05 Nm and the Delamination factorINandOUTreduced from 1.012 to 1.005 and 1.022 to 1.015 respectively.

After drilling,the microscope image analysis was executed to examine the delamination factor (INandOUT) of the drilled samples.The testing of the lower layers was carried out by flopping the workpiece materials to overcome this shortcoming.The microscope image were set at 0.7× magnification so that damage features on the entrance and the exit side can be visualized better.Fig.13(a and b) displays a microscope image with specific cutting parameters of the entry and exit of the machined holes.Fig.13(a and b)demonstrates that the efficiency of the holes on both sides is enhanced by applying graphene oxide to the CFRP epoxy content.Material damage on the entrance and exit side is the maximum in the case of 3% GO sample.At the entrance and exit side,delamination,however,reduces significantly with one wt.% of G doping.Doping of GO improves the ILSS and flexural strength of the nanocomposites,which necessitates more cutting force to penetrates the tool inside the laminates.Also,the higher feed rate increases the force on the cutting edge of the tool.The pattern of higher cutting force and feed rate significantly increases the delamination damages on the laminates [82].

3.6.Microstructure analysis (FESEM analysis)

The microstructure analysis was completed to check the surface features and defects generated on the drilled samples at different wt.% ofGdoping (1,2 &3%).The image of the Field Emission Spectroscopy(FESEM)shows in Fig.14(a-c)typical defects induced during drilling of Graphene nanocomposites,including fiber breakage,fiber pull-out,deposition and matrix cracking.The smooth surface shows the good quality of the drilled hole.Form Fig.14(a),it has been observed that the damages at this 1 wt.%ofGis very less.Also,the matrix region is properly bonded.It is mainly due to the balanced addition of wt.% of G into the epoxy matrix[83,84].The small quantity of nanofillers can enrich the desired mechanical properties in a minimal quantity content as compared to micro or macro reinforcing agents [59,85].The GO content improves the polymer mechanical properties and it provides the required lubrication to the tool-chip interface during machining,which causes less surface damages.From Fig.14 (b),the marks of damages such as fiber breakage is clearly seen.The machining induced damages are more observed in Fig.14(c),the GO 3 wt.%is added in the epoxy matrix,which recovers the ILSS and flexural strength of the laminates.It requires more cutting force to penetrates into the laminates with a higher feed rate.Due to the combined effect of higher feed rate and cutting force,the thrust force also increases,which increases the size of the damage zone in the form of delamination.The higher feed rate expands the delamination,which is mainly due to the chisel edge of the drill bit.Due to the high feed rate,the composite layers which directly make contact with the chisel edge are extruded most quickly as compare when it is to be cut [30,86].The force of drilling pushes the job material is higher,which creates the more bending that can start to crack initiation,matrix debonding and fiber pull/fracture[24,87,88].Hence a balanced addition of nanomaterials and the optimal parametric setting is extremely required for a cost-effective machining environment.

Fig.14.FESEM microstructure of drilled hole at 2400 rpm,80 mm/min and for GO/CF reinforced epoxy composite (a) 1 wt.%;(b) 2 wt.%;(c) 3 wt.%,respectively.

4.Conclusions

This article presents the machinability behavior of relatively advanced polymer nanocomposites developed by Graphene oxide modified epoxy and carbon fiber into the polymer matrix.The consequences of process parameters on machining chracteristics have been investigated by using TiAlN-SiC drill bits.The hybrid approach of GR-PCA did multi-response optimization.The real weight of response is evaluated through the PCA tool.Based on performed machining(drilling)operation and surface morphology outcomes,the following conclusion can be drawn.

? The GR-PCA based statistical model used for the analysis of varying process parameters shows their effect on drilling characteristics.The weight assignment was done using the PCA method.

? The aggregation of multiple responses was done by Grey theory,which is not possible by traditional optimization methods like the Taguchi theory.The outcomes of the proposed GR-PCA model demonstrate the feasibility of optimal parametric setting during drilling of polymer nanocomposites.

? The effect of drilling parameters examined using response diagrams.The result of input parameters over GR-PCA is found as spindle speed at 2400 rpm,feed rate at 80 mm/min and wt.%of Graphene oxide 1%,which shows the minimum values of thrust force and torque delamination factor(INandOUT).

? The parametric combination shows that 1%of GO is sufficient to improve the desired drilling characteristics and a combination of higher spindle speed and low feed rate is preferably used for low values of Thrust and Torque.Also,it has been observed that the damages at this 1 wt.%ofGis very less and the matrix region is adequately bonded.It is mainly due to the balanced addition of wt.% of G into the epoxy matrix.

? The feed rate is the most prominent factor for machining characteristics during drilling of polymer nanocomposites.The collective effect of feed rate and spindle speed can increase the rate of material removal.The higher value of feed rate and wt.%ofGis responsible for surface damages like fiber pull-out,fiber fracture and cracks.

? On increasing the feed rate,the thrust force increases,which in turn raises the delamination damages and it can reduce the bearing strength.Drilling at higher spindle speed raises the temperature of the machining zone (tool-chip),which softens the materials.The cutting edges of tools get heated,which disturbs the matrix stability and worsens the heat-affected area of the drilled hole and takes the lead to a low point bearing strength.

? The maximum prediction error for the hybrid GR-PCA module is 6.897%and the average error percentage is 2.985%.The accuracy of the GR-PCA hybrid module is observed as satisfactory.

This article focuses on the machinability aspects of GO/CF polymer nanocomposites.The outcomes of the proposed model demonstrated the feasibility of this work in a real manufacturing environment for quality and productivity enhancements.The proper selection of parameters can control the drilling-induced damages in a significant way.The inclusion of some other constraints such as Tool geometry,different drill materials,fiber orientations,tool wear,thermal behavior of nanocomposites during machining can be discovered in the future for a better comprehension of nanocomposites machinability.The outcomes of the proposed work can be recommended to polymer industries for continuous quality control and productivity issues.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The authors would like to acknowledge the Uttar Pradesh Council of Science and Technology (UPCST) Lucknow India for expanding all possible help in carrying out this research work directly or indirectly.

Abbreviations

S Spindle speed

F Fee d rate

G% wt.% of Graphene oxide

ANOVA Analysis of variance

CNR Carbon nanorods

AWJ Abrassive water jet machining

GRA Grey relation analysis

TWR Tool wear rate

MRR Material removal rate

Ra Surface roughness

WEDM Wire electrical discharge machining

DFA Desirability function approach

TOPSIS Technique of Order Preference Similarity to the Ideal Solution

IFSS Inter laminar flexural strength

CNC Computer numerical control

CFRP Carbon fiber reinforced polymer

GFRP Glass fiber reinforced polymer

CNT Carbon nanotubes

CNO Carbon nano-onions

GO Graphene oxide

GR-PCA Grey relation embedded principal component analysis

GRC Grey relation coefficient

GO CF-Graphene oxide/carbon fiber

Wt.% of G weight percentage of Graphene oxide

CVD Chemical Vapour deposition

FRP Fiber reinforced polymer

CF Carbon fiber

ILSS Inter-laminar shear strength

TEM Transmission electron microscopy

AFM Atomic force microscopy

SEM Scanning Electron Microscope

EPD Electrophoretic deposition

CCD Central Composite Designs

FTIR Fourier-transform infrared spectroscopy

XPS X-ray photoelectron spectroscopy

RSM Response Surface Method

TiAlN (SiC) Titanium Aluminium Nitride(Silicon carbide)

MWCNT Multi-walled carbon nanotubes

Funding

This research did not receive any specific grant from funding agencies in the public,commercial,or not-for-profit sectors.