INTRODUCTION
Wood composites find wide applications in construction and furniture industries.
The demand for wood based Medium Density fiber boards have increased greatly
due to population growth while the timber resources are alarmingly depleted.
These composites are subjected to various machining operations while in assembly
of parts. Among the machining operations, drilling is most critical due to characteristics
of composites. The poor hole drilling had been reported to account for 75% of
all part rejections (Irle and Loxton, 1996). Because
drilling is preferred on nearly completed parts, such defects are very costly.
Also the quality of holes has been linked to the strength of the joint being
assembled and life of the part (Aguilera et al.,
2000; Neese et al., 2004). Surface finish
is an important parameter in manufacturing engineering, which can influence
the performance of final parts and production.
In Metal drilling and turning had been studied extensively in the literature
but MDF drilling has not received much attention. However, many works of various
authors (Hiziroglu et al., 2004; Hiziroglu
and Suzuki, 2007; Engin et al., 2000) have
presented about the machining of MDF. They strongly recommended that the machinability
is dependent on the mechanism of cutting tool and work piece material.
From the literature, it has been asserted that machining MDF is strongly dependent
on the machining parameters. Philip and Gordon (2006)
studied the application of PCD tool in machining MDF. According to his study
the friction on the rake is small and the pressure exerted by uncut chip on
the rake face mainly dominates the force on the rake face. Lin
et al. (2006) reports about the machinabilty of MDF. This author
confirms that the board densities were found to have major influence on the
machinability characteristics of the panel. Davim et
al. (2008) presents the study of surface roughness aspect in milling
MDF. In his study the surface roughness in milling decreases with an increase
of spindle speed and increase with feed rate. Prakash and
Palanikumar (2010) formulated a mathematical model for the prediction of
surface roughness in drilling MDF panel with Tin coated carbide drill. The result
shows that the feed rate is the major factor which influences surface roughness
followed by drill diameter associated with spindle speed. Recently, Prakash
et al. (2011) evaluated modeling for surface roughness parameters
(R_{a}, Rz) in drilling of MDF panel using BoxBehnken Experimental
Design (BBD) with Brad and spur drills. In his study, the surface roughness,
both R_{a} and R_{z} increases with increasing feed but decreases
with increasing spindle speed.
SURFACE ROUGHNESS
Surface roughness is a measure of the technological quality of a product and
a factor that greatly influences the manufacturing cost. There are several methods
to describe surface roughness, such as roughness average (R_{a}), rootmeansquare
(RMS) Roughness (R_{q}) and maximum peaktovalley roughness (R_{max})
etc. The surface roughness used in this study is the arithmetic mean average
surface roughness (R_{a}), which is mostly used in the industry. It
can be expressed by the following mathematical relationship (Prakash
et al., 2011):
where, Ra is arithmetic average deviation from the mean line and Y is ordinate
of the profile curve and L is the sampling length.
EXPERIMENTAL STUDIES
Materials choice: In this study, MDF panel has been used for experimentation
with a tensile strength of 0.8 N mm^{2}, density of 700 kg mm^{3}
and modulus of rupture of 28 N mm^{2} was selected. The panels are
supplied by ASIS, India, which is manufactured by them. These panels are commercial
available and used for furniture industry. The important properties of the board
as per ISO 12406 are given in Table 1.
Drilling procedure and design of experiments: For conducting, the experiments
an VMC 100 machining centre with following specifications: Table size: 1270x230
mm; spindle speed 605000 rpm; maximum feed rate: 4000 mm min^{1} was
employed. The drill bit used in the investigation is ‘Step
drill’ carbide type, having drill
diameter of 4, 8 and 12 mm was used. The experimental setup and the drill bit
used are presented in Fig. 1. In the experimental plan, the
most dominant process parameters such as Feed rate (f), spindle speed (N) and
Drill diameter (d) were varied at three levels. The three process parameters
and their factor levels are summarized in Table 2.
In order to measure the average surface roughness (Ra) of MDF panels, Taylor
Hobson surface roughness measuring device was used.

Fig. 1: 
Experimental setup and drill bit used 
A Scanning Electron Microscope (SEM) was used to investigate the machined surface.
Normally one need to conduct many number of experiments, when three factors,
each varied at three levels are considered, using experimental design. In order
to save on experimental cost and time, Taguchi’s
orthogonal array was applied to obtain the surface roughness of MDF panel drilling
process. A L_{27 }orthogonal array was found to be appropriate and it
was chosen. The layout of the L_{27} orthogonal array and the measured
surface roughness values are shown in Table 3.
Determination of regression analysis model for surface roughness: Regression
analysis method includes the experimental investigations, mathematical methods
and statistical analysis. In the present investigation, a whole analysis was
done using the experimental data in Table 3.
Table 2: 
Machine settings used in the experiments 

A response surface regression analysis was performed to predict the surface
roughness using Minitab 16 software. Specially, with a sample on n observations
of the dependent variable Y (Ra), the regression model can be expressed as:
where, k is the number of factors (3) β_{0} is the free term,
β_{1} is the linear effect, β_{ii} is the squared
effect and β_{ij} is the interaction effect (Montgomery,
1997). The second order polynomial regression equation representing the
surface roughness (Ra) can be expressed as function drilling parameters such
as spindle speed, feed rate and drill diameter. Eq. 4 can
be rewritten to build relationship between the drilling process parameters and
surface roughness as follows:
RESULTS AND DISCUSSION
Results of regression analysis, used to establish inputoutput relationship
in drill MDF panels are shown and discussed below.
Table 3: 
Experimental design using L _{27 }orthogonal array
and experimental results 

Table 4: 
ANOVA test results for surface roughness 

R^{2}: 0.97, DF: Degree of freedom, SS: Sum of squares,
MS: Mean square 

Fig. 2: 
Comparison of regression model results with experimental measurements 
Estimation of surface roughness by regression analysis: From the results
(Table 3), the final regression model for surface roughness
obtained is as follows:
Note that some interaction terms are removed from full model because of their
no significant effect. Regression statistics R^{2} and R^{2}_{Adj
}are obtained equal to 97.49 and 96.17%, respectively. The R^{2}
value indicates that the drilling parameter explain 97.49% of variance in surface
roughness. This value indicates that the presented model fits the data very
well. The Analysis of Variance (ANOVA) for regression analysis is shown in Table
4.
The pvalue shows that the model terms are significant influence on surface
roughness. The results predicated by regression model are compared with experimental
measurements results in Fig. 2.
It can be seen from Fig. 2 that model prediction presents
a good agreement with the experimental data.
Analysis of the S/N ratio: In the Taguchi method, the term ‘signal’
represents the desirable value (mean) for the output characteristics and the
term ‘noise’ represents the undesirable value (standard deviation)
for the output characteristics. Therefore, the S/N ratio is the ratio of the
mean to the standard deviation. There are several S/N ratios available, depending
on the type of characteristics: Lower Is Better (LB), Nominal Is Best (NB),
or Higher Is Better (HB) (Montgomery, 1997).
The lower the better quality characteristics can be formulated as (Engin
et al., 2000):
where, n is the number of measurements in a trail/row, in this case, n = 3
and y_{i} is the ith measured value in a row/run. The S/N ratio values
are calculated by taking into consideration Eq. 5. The surface
roughness values measured from the experiments and their corresponding S/N ratio
values are listed in Table 3.
The response table for each level of process parameters was created in the
integrated manner and the results are given in Table 5.
Based on the S/N analysis ratio and the ranking position in the response (Table
5), the most influencing parameter is feed rate (rank 1), followed by drill
diameter (rank 2) and spindle speed (rank 3). Figure 35
show the effect of drilling parameters on the surface roughness.

Fig. 3: 
Feed rate vs. S/N ratio 

Fig. 4: 
Spindle speed vs. S/N ratio 

Fig. 5: 
Drill dia vs. S/N ratio 
THE CHARACTERISTICS OF DRILLED SURFACE OF MDF PANELS
The drilled MDF panels were observed under Scanning Electron Microscope (SEM)
for analyzing the quality of surfaces. Figure 6 and 7
show, SEM image of the wood dust observed in drilling of MDF composites.

Fig. 6: 
Scanning electron micrograph of a broken fiber 

Fig. 7: 
SEM micrograph of dust form at higher magnification 
Figure 6 shows how the wood fiber fracture takes place while
drilling of wood composite materials which consists of lingocellulosic fibers.
The fracture is takes place in the middle of the weaving used in drilling of
composites. It does not show the combination of fibers and matrix materials,
it may be due to the weak bond between the lingocellulosic fiber and the matrix
materials, since the composite is manufactured through dry processing technology.
The separation of adhesive and lingocellulosic fibers is observed in the Fig.
7, which is taken at high magnification (100X). In drilling of composite
materials, the chips formed are in a loose form with completely discontinued
powder chips. The chips study indicates that the chip production process in
MDF composites is completely different from the drilling of metals. Figure
8 shows, the microstructure of the specimen which has been taken at low
feed rate of 100 mm min^{1} for drill diameter 4 mm.

Fig. 8: 
SEM micrograph at a low feed rate of 100 mm min^{1}
for 4 mm drill diameter 

Fig. 9: 
SEM micrograph at a medium feed rate of 300 mm min^{1}
for 4 mm drill diameter 

Fig. 10: 
SEM micrograph at a high feed rate of 500 mm min^{1}
for 4 mm drill diameter 
The surface of the microstructure indicated that smooth surface is observed
in drilling of MDF composites. In some portion of the work piece, there is a
small fracture and tiny tots are observed. It is the common phenomenon in drilling
of MDF composites.
Figure 9 shows, the microstructure of the work piece observed
in drilling of MDF composites at a medium feed rate of 300 mm min^{1}
for 4 mm drill diameter which indicates that the surface is not smooth. Fussy
surface is observed at 300 mm min^{1} of feed rate. Figure
10 shows, the microstructure of the specimen taken at high feed rate (500
mm min^{1}) which shows the fiber pull out and protruding fibers. Generally
the surface is not smooth.
From these microstructures, it can be predicted that, low feed rate is preferred
for drilling of MDF composite panels. It has been noticed that almost smooth
surface is observed inside the holes. In some places tiny pits and small damages
are observed in the cut section; it is due to the incomplete distribution of
adhesives and removal of some fibers in the composite materials when drilling
is carried out.
CONCLUSION
To determine the relationship between drilling parameters and surface roughness
in MDF panels, regression analysis was carried out based on Taguchi’s
orthogonal array. The drilled surfaces were also examined using Scanning Electron
Microscopy (SEM). Summarizing the main features of the results regression analysis
is seen to be sufficient for estimating surface roughness in drilling the MDF
panels. The predicted process parameters are found to be close correlation with
the actual performance results.
From this, the predictive models can be used for predicting surface roughness
in drilling MDF panels with a higher reliability. The performance can further
be enhanced with large experimental data from full factorial experimentation
and considering the additional performance characteristics. Based on the ANOVA
and Taguchi method, the most dominant parameter on surface roughness was found
to be feed, while the second ranking factor was drill diameter. Spindle speed
is less effective on surface roughness.
Further study could consider more factors different drill properties, point
angle, board thickness in the research to see how these factors would affect
the surface roughness.