Available online at www.sciencedirect.

com

ScienceDirect
Procedia Materials Science 6 (2014) 967 – 974

3rd International Conference on Materials Processing and Characterisation (ICMPC 2014)

Thrust, Torque Analyses and Optimization in Microdrilling of

GFRP Using Full Factorial Design Integrated CNSGA-II Algorithm
Shiba Narayan Sahua* Debasis Nayakb, Hemanta kumar Ranac
a,b
Gandhi Institute for Education and Technology,Baniatangi, Po: Bajapur,
Khurda, Bhubaneswar 752060, India
c
Centurion University of Technology and Management, At - Ramchandrapur
P.O. - Jatni, Dist: Khurda,Bhubaneswar 752050, India

Abstract

Glass fibre reinforced plastics (GFRP) are having good weight to strength ratio, so they have wide range of application for
many purposes. Micro drilling of GFRP produces de-lamination in it. So it is always needed to optimize the process parameters
by minimizing thrust force and torque. The process parameters and response parameters values are used here for analysing and
optimizing the micro drilling of GFRP. The introduction of peck micro drilling in lieu of direct micro drilling was found to be
responsible for lower value of torque and thrust force. The optimized result of thrust force and torque was found to be 40000
and 5 of rpm and feed respectively for both direct and peck type of micro drilling.
© 2014 The Authors. Published by Elsevier Ltd.
© 2014 Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Selection and peer review under responsibility of the Gokaraju Rangaraju Institute of Engineering and Technology (GRIET)

Keywords:Micro-drilling;GFRP; Direct drilling; Peck drilling; ANOVA; CNSGA-II

1. Introduction

Due to its high strength to weight ratio, high stiffness, low thermal conductivity and corrosion resistance Glass
fibre reinforced plastic (GFRP) have wide application in automobile, aerospace and electronics industries
[Rahamathullahand et al.(2011),Rahamathullahand et al. (2012)]. GFRP require micro machining, when used for
specific applications like making holes in printed circuit board (PCB). Machining of holes with diameters less than
0.5mm is called as micro-drilling. Micro-holes can be produced by both traditional and non-traditional process.

* Corresponding author. Tel.: +91-6755-243603.

E-mail address:shibacvrce381@gmail.com

2211-8128 © 2014 Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Selection and peer review under responsibility of the Gokaraju Rangaraju Institute of Engineering and Technology (GRIET)
doi:10.1016/j.mspro.2014.07.167
968 Shiba Narayan Sahu et al. / Procedia Materials Science 6 (2014) 967 – 974

Traditional process is machining by micro-drills, while non-traditional processes are Ultra-sonic drilling, Laser
beam drilling, EDM, Water-Jet drilling. Because of materials like cupper, fibre and resin in PCB having different
optical and material properties, it is difficult for Laser beam with particular characteristics to cut cleanly and
efficiently through the board [Rahamathullahand et al (2011, 2012),Chinnock(1995)].

Micro-drilling is the cheapest way to produce these holes. So it is necessary to optimize the drilling
parameters on surface roughness and thrust force. The main difficulty that is incurred with drilling GFRP is Fibre
pullout, Fuzziness, de-lamination of glass fibres. This happens due to increase in thrust force and tool flank wear
(Palanikumar et al., 2012). This is because of the anisotropy of the composite material. This can be reduced by using
lubricants over the work piece, changing the tool geometry, applying interrupted cutting [Kudla(2006)] and
vibratory drilling [Zhang et al. (2001)]. Fuzzy and rough cuts are produced when feed rate is too low, so that the
accumulated heat around the tool edge destroys the matrix stability. Thrust force is mainly affected by feed rate for
GFRP composite materials, which is obtained from the results of the experiments carried by Davim et al. [Davim et
al.(2004)]. Similarly increase of cutting speed reduces the thrust force [Khashaba(2004)]. The damages of drilling
polymeric composites can be categorized into four types namely, (1) de-lamination at drill entry, (2) geometric
defects, (3) temperature related damages and (4) de-lamination at drill exit. The first case is not always present. The
second case is associated with the orientation of fibres and cutting edges, due to which elliptical holes are created
because of dissimilar nature of fibre and matrix. The minor axis of the ellipse is in the direction of fibre orientation.
The temperature related damages happens due to friction between the drill and the wall of hole. The fourth case is
the cause of development of normal stress, which opens the matrix and fibre interface [Lachaud et al.(2011)].The
empirical relationship among the influence of feed and drill size on thrust force and torque on drilling have been
coined, which can be extended for large number of composite systems and various material parameters [ Mohan et
al. (2005)].The de-lamination of composite material by Taguchi with various drill bits have also been analysed and
the result have been found that during the drilling process drill diameter and feed rate have largest contribution
[Tsao andHocheng (2004)].

Taking into account the complex geometry of drill and drilling process, the thrust force and torque prediction
models are based on empirical relations, namely (1) The straight line model, (2) Power model, (3) Logarithmic
model, (4) Quadratic model, (5) Mechanistic model, (6) Finite element model. Power model correlates better with
the experimental results [Shaw and Oxford (1957), Lin and Ting (1995), Wang and Zhang (2008), Jalali et al.
(1991)].

The effect of high speed in the drilling of GFRP have been studied and it was observed that with high
speed machining productivity can be improved by lowering the production cost in manufacturing companies [ Rubio
et al., (2008)].

The mechanical model to predict thrust and torque in vibration drilling of fibre-reinforced composite
materials shows that thrust and torque can be reduced by 20-30 percent compared with conventional drilling. This
happens due to mechanics of vibration cutting analysis and the continuous distributions of thrust and torque along
the lip and the chisel edge of a twist drill[Zhang et al. (2001)].The factors that limit the use of smaller drill size in
drilling PCB and observed that the copper content used in a board should be held to minimum. In order to get this
the thinnest possible copper foil should be used [Hinds et al. (2000)]. The drilling in carbon/epoxy composites
have been investigated by researchers by conducting an experiment and forming a unique three dimensional finite
element model of drilling in composite laminate, the obtained data was optimized and found that the thrust force,
torque and de-lamination damage increases abruptly with an increase in feed rate, but reduces gradually with
increase in cutting speed. Hence they concluded that low feed rates and high cutting speeds are ideal for drilling
carbon/epoxy laminate [Vaibhav et al. (2013)].The Thrust force and Torque in drilling GFRP composites by
multifaceted drill using fuzzy logic have been modelled and analysed .The research has been made by taking Glass
Fibre Reinforced Plastic using 8 facet solid carbide tool. 3 parameters such as spindle speed; feed rate and drill
diameter with each having 3 levels using Taguchi L27 was used. Machine used was ARIX-CNC Machine centre.
Thrust force and torque were taken as judging criteria for optimization of input parameters. Fuzzy rule based
model was developed to indicate thrust force and torque in drilling of GFRP composites. The results suggested that
 Shiba Narayan Sahu et al. / Procedia Materials Science 6 (2014) 967 – 974 969

the model can be effectively used for predicting the response variable by means of which delimitation can be
controlled [Singh et al., (2009)]. Experimental study on peck drilling of GFRP and predicted the drilling induced
damage using ANN and it was observed that increasing pecking length damage factor at the hole exit decreases.
Again increase of cutting speed decreases damage factor both at hole entrance and exit. It is also observed that
increase of feed increases damage factor at hole exit and decrease damage factor at the hole entrance [Kentli,
(2011)].

2. Experimental data (machining parameters collection) and methodology

The machining parameters method of drilling, rpm andfeed were used as input variables while performing the
analysis, modelling and optimisation for the responses thrust force and torque involved in micro drilling of
GFRP.The data of process parameters and response parameters used here for modelling the micro drilling of GFRP
has been referred from RahamathullahandShunmugam[RahamathullahandandShunmugam(2011)].

2.1. Full factorial design

The process parameter method of drilling at two levels, revolution per minute(rpm) of the micro drill bit and
feed each at five levels have been chosen for a full factorial design as shown in Table 1 .All process parameter
and performance parameter were designed by full factorial design methodology with a sum total of fifty number of
runs. The technique, which can be used to decompose the total variability in to its component, is known as the
analysis of variance (ANOVA). After that, ANOVA has been conducted for finding out the significant process
parameters. For level of significance of α=0.05, the cut-off value for the p-value is 0.05, which has been chosen as
one of the criteria for identifying the significant and insignificant parameter. If the calculated p-value of a test
statistic is less than 0.05, then the null hypothesis will be rejected. F-test is another test to determine whether the
interaction and main effects are significant. Larger values of F facilitate the rejection of the null hypothesis.
Table 1. Parameter and their levels

Factor Type Levels Values/Types

Method of Direct/
Fixed 2
Drilling Peck cycle
Speed Fixed 5 20000 25000 30000 35000 40000
Feed Fixed 5 5.0 7.5 10.0 12.5 15.0

2.2. Controlled elitist non dominated sorting Genetic algorithm-II

A controlled elitist non dominated sorting genetic algorithm-II (CNSGA-II) is one of the best evolutionary
algorithms for optimization. Here individuals of best solutions are taken as parent population of non-dominated
fronts. Then child population is created from the parent population keeping in view that the created population of
solutions are sorted as non-dominated fronts. Then the parent and child populations are combined together to form
a set of solutions, from which child population of next generation is created. All the solutions of the child
population have solutions from each front with a desired distribution of individuals, so that solutions of each front
have equal importance. This process continues until the stopping criterions are met. For implementing CNSGA
algorithm a MATLAB code was used with the functions and parameters setting as tabulated in Table 2.

Table 2.Process parameter and functional setting of Controlled NSGA algorithm.

Types of operation and parameter Functions or parameters value used

Population
a. size 60
b. creation function Feasible population
Selection Tournament
970 Shiba Narayan Sahu et al. / Procedia Materials Science 6 (2014) 967 – 974

Reproduction
a. crossover fraction 0.8
Cross over
a. crossover function Intermediate
b. crossover ratio 1.0
Mutation Adaptive feasible
Distance measure function Distance crowding
Stopping criteria
a. generation 800
b. stall generation 100
c. functional tolerance 1x10-6

3. Results and discussion

3.1 Influence on thrust force

In the process of micro drilling of GFRP, the influence of various machining parameters like method of
drilling, rpm and feed have a significant effect on thrust force, as shown in the main effect plot for thrust force in
Fig 1(a).The feed is directly proportional to thrust force. Thrust force is decreasing with the increment of rpm.
Method of drilling is seems to have less significance towards thrust force in comparisons to rpm and speed. But in
direct drilling larger amount of thrust force is being developed as compared to peck drilling. The exact percentage
contribution of sources can be referred from percentage contribution column of table2.

Fig. 1. (a)Main effect for Thrust force, (b) Residual plots for thrust force
 Shiba Narayan Sahu et al. / Procedia Materials Science 6 (2014) 967 – 974 971

From the ANOVA (table 3) it can be concluded that feed has highest contribution towards the development
of thrust force. In this table insignificant factors for 95% of confidence interval i.e. factors having p-value more than
0.05 is shown in ‘*’ in P column.Residual plots are generally used to examine the goodness of model fit. Figure 1(b)
shows the normal plot of residuals, a histogram of residuals, and a plot of residuals versus fits, and a plot of
residuals versus order. From these plots it can be concluded that the model fit here is more accurate one. It can be
seen from the histogram graph that frequency of occurrences of the obtained results of thrust force best suits to the
normal distribution curve and is high for the standardized mean andslight lack of skewnesscan be observed. The
mean of the thrust force is very less with standard deviation of 1.01 for an order of observation 50. The fitted values
are very precise and accurate to the standardized mean.

Table 3 Analysis of Variance for thrust force

Source DF Seq SS Adj SS Adj MS F P Contribution

(%)
Method of 1 0.038337 0.038337 0.038337 8.65 0.010 1.892574
Drilling

Speed 4 0.202641 0.202641 0.050660 11.44 0.000 10.00373

Feed 4 1.484116 1.484116 0.371029 83.76 0.000 73.26602
Method of 4 0.032342 0.032342 0.008085 1.83 0.173* 1.59662
Drilling *
speed

Method of 4 0.055985 0.055985 0.013996 3.16 0.043 2.763799

Drilling *
Feed

Speed * 16 0.141354 0.141354 0.008835 1.99 0.089 6.978191

Feed
Error 16 0.070879 0.070879 0.004430 3.499067
Total 49 2.025654 100
S=0.0665577, R-Sq = 96.50%, R-Sq(adj) = 89.28%* = Insignificant factor

3.1. Influence on torque

Main effect plot for torque is shown in Figure 2(a), which indicates that with increment of feed torque
increases.From the main effect plot it can be concluded that rpm is inversely proportional to the torque. In peck
micro drilling a lower value of torque will be develop as compared to direct micro drilling subjected to same
operating condition. Referring the percentage contribution column of table 3, it can be concluded that feed is
having a maximum contribution of 42.03% towards the development of torque. From the ANOVA (table 4) all the
main effects and two level interacting effects are found to be having significant contributions towards
torque.Figure 2(b) shows the normal plot of residuals, a histogram of residuals, a plot of residuals versus fits, and a
plot of residuals versus order which can be useful to examine the goodness of model fit.
972 Shiba Narayan Sahu et al. / Procedia Materials Science 6 (2014) 967 – 974

Fig. 2. (a)Main effect for torque, (b) Residual plots for torqe

Table 4 Analysis of Variance for thrust force

Source DF Seq SS Adj SS Adj MS F P Contribution

(%)
Method of 1 0.041210 0.041210 0.041210 8.63 0.010 1.549764
Drilling

Speed 4 0.591034 0.591034 0.147758 30.95 0.000 22.22673

Feed 4 1.117730 1.117730 0.279432 58.54 0.000 42.03393
Method of 4 0.273303 0.273303 0.068326 14.31 0.000 10.27797
Drilling *
speed

Method of 4 0.078262 0.078262 0.019566 4.10 0.018 2.943161

Drilling *
Feed

Speed * 16 0.481198 0.481198 0.030075 6.30 0.000 18.09618

Feed
Error 16 0.076377 0.076377 0.004774 2.872272
Total 49 2.659114 100
S=0.0690912, R-Sq = 97.13%, R-Sq(adj) = 91.20%* = Insignificant factor

3.2. Optimization using CNSGA-II

Pareto-optimal decision vectors are tabulated in table 5 and plotted as Pareto-optimal front, which are shown
in Figure 3.

Table 5 Pareto-optimal decision vectors

 Shiba Narayan Sahu et al. / Procedia Materials Science 6 (2014) 967 – 974 973

Method
Predicted Experimental Experimental Experimental
of RPM Feed
torque thrust force thrust force torque
Drilling

Direct
4000 5 1.0612 0.6842 0.9830 0.6564

Peck
4000 5 1.0476 0.7339 0.9270 0.7163

Figure 3 Pareto-optimal front obtained from CNSGA-II

4. Conclusions

From the main effect plot of full factorial methodology, it can be concluded that feed is directly proportional
to both torque and thrust force. Feed is also having maximum percentage of contribution towards torque and thrust
force. With the decrease in speed the development of both thrust force and torque is reducing. The introduction of
peck micro drilling in lieu of direct microdrilling is responsible for lower value of torque and thrust force. From
the optimization result, optimal process parameter setting for obtaining minimal thrust force and torque was found
to be 40000and 5 of rpm and feed respectively for both direct and peck type of micro drilling.

References

Chinnock, C., 1995. Laser-drilled vias could make multichip modules cheap, Microelectronics Mfg, 31, 30-31.
Davim, J.P., Reis, R.Pedro,Conceicao, A., 2004. Experimental Study of Drilling Glass Fibre Reinforced Plastics
(GFRP) Manufactured by Hand Lay-up. Composites Science and Technology, 64, 289–297.
Hinds,B. K. and Treanor,M, 2000. Drilling of Printed Circuit Boards: Factors Limiting the Use of Smaller Drill
Sizes, Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering, 214: 35-45.
Jalali, S. A. and Kolarik, W. J., 1991. Tool Life and Machinability Models for Drilling Steels,Int. J. Mach. Tools
Mf., 31(3), 273–282.
Kentli,Aykut, 2011. Experimental Study on Peck Drilling of GFRP andPrediction of Drilling-induced Damage
using ANN, Scientific Research and Essays Vol. 6(7), pp. 1546-1554.
Khashaba, U.A. ,2004. Delamination in Drilling GFR-Thermoset Composites, Composite Structures, 63, 313–327.
Kudla, L., 2006. Prevention Against Breakage of Miniature Drills,PeriodicaPolytechnica Ser. Mech. Engng, 44,
1413-1421.
974 Shiba Narayan Sahu et al. / Procedia Materials Science 6 (2014) 967 – 974

Lin, S. C. and Ting, C. J., 1995. Tool Wear Monitoring in Drilling Using Force Signals. Wear, 180, 53–60.
Lachaud, F., Piquet, R., Collombet, F., Surcin, L., 2001.Drilling of Composite Structures,52 (3–4), 511–516.
Mohan, N. S., Ramachandra, A., Kulkarni, S. M., 2005.Machining of Fiber-Reinforced Thermoplastics: Influence
of Feed and Drill Size on Thrust Force and Torque during Drilling, Journal of Reinforced Plasticsan
Composites, Vol. 24, pp. 1247-1257.
Palanikumar,K., Latha,B., Senthilkumar,V. S., and Paulo DavimJ.,2012. Analysis on Drilling of Glass Fiber–
Reinforced Polymer (GFRP) Composites Using Grey Relational Analysis, Materials and Manufacturing
Processes, 27: 297–305.
RahamathullahandI,ShunmugamMS, 2011.Thrust and Torque Analyses for Different Strategies Adapted in
Microdrilling of Glass-Fibre-Reinforced Plastic, Proceedings of the Institution of Mechanical Engineers, Part B:
Journal of Engineering manufacturer, pp. 505-519.
Rahamathullahand I,Shunmugam MS, 2013. Analyses of Forces and Hole Quality in Micro-Drilling of Carbon
Fabric Laminate Composites,Journal of Composite Materials, 47: 1129-1140
Rubioa,J.Campos, Abrao,A.M., Faria,P.E., CorreiaA.Esteves, DavimJ.Paulo, 2008. Effects of High Speed in the
Drilling of Glass Fibre Reinforced Plastic: Evaluation of the Delamination Factor, International Journal of
Machine Tools & Manufacture, 48, 715–720.
Singh, R. Vimal, Sam, Latha, B., and Senthilkumar,V.S., 2009. Modelling and analysis of Thrust force and Torque
in drilling GFRP composites by multifaceted drill using fuzzy logic, International Journal of Recent Trends in
Engineering, Vol. 1, No. 5, pp 66-70.
Shaw, M. C. and Oxford, C. J., 1957. On the Drilling of Metals – The Torque and Thrust in Drilling,Trans. ASME,
139–148.
Tsao,C.C., Hocheng,H., 2004. Taguchi Analysis of Delamination Associated with Various Drill Bits inDrilling of
Composite Material, International Journal of Machine Tools & Manufacture,44 ,1085–1090.
Vaibhav, A.Phadnis, Farrukh,Makhdum, Roy, Anish, Vadim,V. Silberschmidt, 2013. Drilling in Carbon/Epoxy
Composites: Experimental Investigations and finite element implementation, Composites: Part A 47 41–51.
Wang, J. and Zhang, Q., 2008. A Study of High-Performance Plane Rake Faced Twist Drills,Part II: Predictive
Force Models. Int. J. Mach. Tools Mf.,48, 1286–1295.
Zhang,L.B., Wang,L.J., Liu,X.Y., Zhao,H.W., Wang,X.,Luo,H.Y., 2001. Mechanical Model for Predicting Thrust
and Torque in Vibration Drilling Fibre-Reinforced Composite Materials, International Journal of Machine Tools
& Manufacture, 41, 641–657.