University of Huddersfield Repository

Zeng, Wenhan, Jiang, Xiang and Blunt, Liam

Surface Characterisation Based Tool Wear Monitoring in Peripheral milling

Original Citation

Zeng, Wenhan, Jiang, Xiang and Blunt, Liam (2009) Surface Characterisation Based Tool Wear
Monitoring in Peripheral milling. International Journal of Advanced Manufacturing Technology, 40
(3). pp. 226-233. ISSN 0268-3768

This version is available at http://eprints.hud.ac.uk/id/eprint/31745/

The University Repository is a digital collection of the research output of the

University, available on Open Access. Copyright and Moral Rights for the items
on this site are retained by the individual author and/or other copyright owners.
Users may access full items free of charge; copies of full text items generally
can be reproduced, displayed or performed and given to third parties in any
format or medium for personal research or study, educational or not-for-profit
purposes without prior permission or charge, provided:

• The authors, title and full bibliographic details is credited in any copy;
• A hyperlink and/or URL is included for the original metadata page; and
• The content is not changed in any way.

For more information, including our policy and submission procedure, please
contact the Repository Team at: E.mailbox@hud.ac.uk.

http://eprints.hud.ac.uk/
Surface Characterisation Based Tool Wear Monitoring in Peripheral milling

W. Zeng*, X. Jiang, L. Blunt

Centre for Precision Technologies, School of Computing and Engineering, University of

Huddersfield, Queensgate, Huddersfield, HD1 3DH, UK

Abstract: The progress of surface metrology in the last decade has led to improved 3D

characterisation of surfaces which offers the possibility of monitoring manufacturing

operations to give highly detailed information regarding the machine tool condition. This

paper presents a case study where areal surface characterisation is used to monitor tool wear

in peripheral milling. Due to the fact that tool wear has a direct effect on the machined

workpiece surface, the machined surface topography contains much information concerning

the machining conditions including the tool wear state. Through analysing the often subtle

changes in the surface topography the tool wear state can be highlighted. This paper utilises

areal surface characterization, areal auto-correlation function (AACF) and pattern analysis to

illustrate the effect of tool wear on the workpiece surface. The result shows that: (1) tool

wear, previously difficult to detect will influence almost all of the areal surface parameters; (2)

the pattern features of AACF spectrum can reflect the subtle surface texture variation with

increasing tool wear. The authors consider that, combined analysis of the surface roughness

and its AACF spectrum are a good choice for monitoring the tool wear state especially with

the latest developments in on-machine surface metrology.

Keywords: surface metrology; tool wear; areal surface texture parameters; areal

auto-correlation function

*Corresponding author Tel.: +44-1484-473613; Fax: +44-1484-472161;

E-mail address: z.wenhan@hud.ac.uk

1 Introduction

With increasing demands for higher productivity and quality, there have been increased

interests in monitoring all aspects of the machining process. As tool wear directly affects the

precision and surface finish of the product, it is important to monitor the state of cutting tool if

precision is to be maintained over time [1].

Metal Cutting is achieved through shear mechanisms and the relative motion between the

workpiece and the cutting tool. As a result of the cutting motion the surface of workpiece

will be influenced by cutting conditions (cutting parameters, cutting force, cutting tool state),

and the surface topography of the workpiece will include much information pertaining to the

cutting process [2-6]. Thus, through monitoring the machined surface topography of the

workpiece and extracting the relevant information the cutting process and tool wear state

should be able to be monitored and quantified. This paper proposes a surface metrology

based tool wear monitoring methodology through analysing the change of the surface

topography features during tool wear. The work forms part of a large study where surface

metrology is being used to measure all aspects of the machining in combination with an

on-line metrology tool.

2 3D Surface Analysis Techniques

2.1 3D Surface Parameters

The traditional surface evaluation methods use surface profile evaluation based on a section

profile of the surface. Cutting process however are in Euclidean space, the information

concerning tool wear must be three dimensional in nature, but default surface profile

evaluation can only recover limited information (2D) from the cutting process and will be

limited in its ability to monitor development of tool wear for example.

Recent advanced techniques have taken place in the ability to measure and characterise
areal surface topography [7-12]. A set of areal surface roughness parameters have been

defined and as shown in Fig.1. The areal surface texture parameters not only describe the

statistical amplitude properties, but also reflect texture distribution properties. As discussed

earlier, tool wear will influence both the surface roughness and the texture distribution. In

this paper, the use of areal surface texture parameters to characterise the machined surfaces of

workpieces in order to evaluate the tool wear is outlined. Clearly not all of the areal

parameters will reflect the tool wear state significantly, and the aim here is to define a subset

of parameters for the purpose of tool wear monitoring.

2.2 Areal Autocorrelation Function (AACF)

The autocorrelation function is a very useful tool for processing random signals. It describes

the general dependence of the topographical values at one position on the topographical

values at another position. For areal surface evaluation, it can not only describe the spatial

relation dependences of the surface topography, but also describe the direction and periodicity

of the surface texture [7-8].

From the authors previous research, the AACF has been used to describe different

machining methods as each has very different surface textures patterns and consequently

very different patterns for their AACF spectra. Furthermore, the AACF spectra can reflect

the texture periodicity and directionality more clearly than visualisation of a surface

topography map. The authors have previously used the AACF to monitor the development

of chatter in peripheral milling. It was observed that with different degrees of chatter the

surface texture on the workpiece showed which were more apparent than simple visualisation.

AACF analysis of machined surface has the capability to reflect the information of the

machining state more clearly than observing the raw data. Consequently in this paper the

pattern features of the AACF spectrum of machined surface are used to monitor the tool wear
state.

3 Simulated Tool Wear Experiment

Under normal conditions, tool wear develops over a period of time. In the present study

however tool wear has been “induced” by using a diamond lap to manually wear the tool edge

and hence control the amount of tool wear. The procedure in detail is as follows: measure

new tool→machine a set of workpieces→measure the workpiece→lap the tool to induce

wear→measure the tool wear→machine a new set of workpieces→measure the

workpieces→lap the tool to induce wear→measure the tool wear→repeat……. Through

this procedure a full set of machined workpieces were obtained under defined amounts of tool

wear.

To eliminate the influence of random factors and to enhance comparability, eight

groups of cutting conditions as shown in table 1 were selected (all the parameters selected

according to the recommendation of the tool’s manufacturer). In the table, nomenclature of

the machined workpieces is as follows: “TWnC”, “TW” represents “tool wear”, the number

“n” represents the tool wear amount of each group of tests, n is 1,2,3,4, and 1 represents the

new tool, the first wear, the second wear and the third wear level; the last character represents

the different combination of the cutting parameters in each group of cutting tests, from A to H.

The material of the workpiece is EN8 with hardness 243Hv. The peripheral milling cutter is

an SSM2200 (radius 20mm, two teeth and 300 helix). For each cut the same position on the

cutter was used, a Talysurf PGI was used to measure the tool wear amount, and for each tooth

four positions at 10mm intervals from the tip were selected for measurement. All machining

was carried out on a Cincinnati Arrow2-500 machine centre.

To measure the machined surface a Talysurf PGI was used, the evaluation area was

6*10mm2, with a sample spacing of 10um, polynomial fitting was applied to remove form
errors firstly and then a Gaussian regression filter was applied to obtain the surface texture,

the cutoff wavelength of the filter c was 2.5mm.

4 Analysis of the simulated tool wear

4.1 Measurement of Tool Wear

Figure 2 shows the section profile of the tool in the new, first wear, second wear and final

wear state, the amounts of the tool wear are approximately 0um, 90um, 110um and 130um.

Tool wear is defined as the departure from idealized tip geometry and is measured using the

form Talysurf PGI.

4.2 Variation of surface texture

Fig.3 compares the 3D surface topography under different tool wear conditions. It is

observed that the surface topography has good texture in feed direction, the surface is very

flat and has low levels of waviness when tool is new; with increasing tool wear the regular

texture loses “strength” with scuffs and furrows appearing, with further tool wear, the surface

appears very irregular having alternate rough and smooth zones.

4.3 Areal Surface Texture Parameter Analysis

Figure 4 shows the variation trend of the 3D amplitude parameters Sq, Ssk, Sku, S5z with the

increase in tool wear. Sq, the areal root mean square roughness, increases with tool wear

increase, with the amplitude of the roughness rising with the increasing tool wear. i.e between

tw1 and tw2, the tool wear changes 90um while between tw3 to tw2 and tw4 to tw3 only

changes 20um. Ssk is the measurement of asymmetry of surface deviations about the mean

/reference plane. This parameter can effectively be used to describe the height distribution

of surface topography. From the figure, Ssk is near zero when the cutter is in good condition,

which means that the surface height distribution curve is very like a standard normal

distribution and a symmetrical distribution. When the tool is worn, Ssk becomes
increasingly negative, due to the fact that the height distribution curve is changed to an

asymmetrical distribution with a negative skew, which shows that the height of the surface is

mainly above the mean plane with the surface tending towards having a “flatter top” with

some deep valleys below the mean surface plane. Sku characterises the spread of the height

distribution. These curves indicate that the Sku is near 3 when cutter in good conditions,

which shows that the surface height distribution is very close to a Gaussian distribution.

When tool is worn, the machined surface has greater “peakedness”, with the Sku being are far

bigger than 3. The parameter S5z increases with the tool wear, which indicates that some

random scores and deep furrows occur on the surface and surface quality deteriorates.

From the combined analysis of the amplitude parameters, it can be concluded that the

surface quality deteriorates with the increasing tool wear. Additionally when the tool is

worn, the blunt cutter edge has a rubbing effect on the surface that results in the surface

tending towards a plateau type surface with a relatively flat tops and some deep scored

valleys.

Figure 5 shows the variation of the areal spacing parameters Sds, Str, Sal with tool wear

increase. The spacing parameters refer to the spatial properties of surfaces. Sds is the

number of summits of a unit sampling area. The curve indicates that the number of the

random peaks and valleys on surface increases with a tool wear increase. Str is defined as the

ratio of the fastest to slowest decay to 0.2, of the AACF of the surface area. The Str is used

to identify texture strength i.e. uniformity of the texture aspect. From fig.3 the surface has a

strong directionality in feed direction when cutter is new; when the tool is worn, the texture

distribution strength in feed direction is weakened. With the increasing tool wear the effect

of randomly occurring deeper furrows along the feed direction on the machined surface is

more apparent. A large value of Sal denotes that the surface is dominated by low frequency
(or long wavelength) components, while a small value of the Sal denotes the opposite case.

So from the figures it is clear that, with increasing tool wear distinct waviness appears on the

surface. This could be due to the Built Up Edge effects along the tool edge and consequent

inconsistent cutting of deep furrows on the surface along the feed direction as well as

ploughing zone of lower roughness indicating poor cutting zones.

Compared with the amplitude parameters, the spacing parameters can reflect the tool

wear’s effect on surface texture distribution more distinctly.

Figure 6 shows the variation of hybrid parameters with increasing tool wear. The hybrid

parameters are parameters based on both amplitude and spatial information. Sdq is the root

mean square value of the surface slope with the sampling area, Sdr is the ratio of the

increment of the interfacial area of a surface over the sampling area, and Ssc is defined as the

arithmetic mean of the principal curvatures of the summits within the sampling area. They

define numerically hybrid topography properties such as the slope of the surface, the

curvature of high spots, and the interfacial area. Any changes that occur in either amplitude

or spacing may have an effect on the hybrid property. From these figures, it is clear that

most of the hybrid parameter Sdq, Ssc, Sdr will increase with increasing tool wear. This

reflects the development of random deep scores smearing and possible BUE formation.

4.4 AACF spectrum analysis

Figure 7 shows the variation of the AACF of the machined surfaces. Relative to the original

surface texture, the AACF spectra can show the varying trend of the surface texture with

increasing tool wear. In fig.7a, the tool is new and the AACF has the same periodicity as the

original surface texture along the feed direction. In fig.7b, the strength of the periodicity

distribution along feed direction is weakened, and the regular machined surface lay is

somewhat destroyed with increasing tool wear. In fig.7c, new periodical waviness
components appear in the vertical direction these are the result of inconsistent cutting along

the feed direction because of the probable presence of BUE. In fig.7d, the periodicity

distribution along feed direction has almost disappeared and that in the vertical direction is

also weakened, this shows that at this stage the tool wear is very serious and less real cutting

of the surface is occurring and the rubbing effects of the tool edge radius are becoming much

more significant.

The AACF illustrates clearly the main features of the surface texture and its pattern

reflects the tool wear state. From the extraction and recognition of the pattern features of the

AACF it is feasible to monitor the state of tool wear, and this is the subject of ongoing

research.

Of the methods outlined in this paper it appears that the AACF technique shows the

clearest indication of tool wear and the authors would recommend this technique. Numerical

analysis of the AACF however requires further methods. The method currently under

investigation by the authors is Pattern analysis of the AACF data [13]. This technique

allows the development of trends in the AACF data to be extracted and further analysed to

investigate subtle changes in surface topography to be quantitatively characterised and

expressed numerically.

5 Conclusions

In the present investigation the affect on surface texture of tool wear have been investigated

via changing tool edge radius under accelerated wear conditions. The increase of the tool

edge radius and the tool surface roughness has a two fold effect on the surface roughness.

One is that it will lead to possible BUE’s formation, growth and break down and the

irregularity of the tool edge, thus developing furrows on the surface and increasing the general

roughness. On the other hand, increasing the tool edge radius produces a rubbing effect on
the machined surface and the surface will possesses an irregular variation of the rough and

smooth zones. Thus, BUE’s effect will increase the roughness in the initial phase of tool

wear, with the rubbing effect becoming apparent as a later wear phenomena

Clearly, the surface topography reflects tool wear state. Tool wear will influence almost

all the 3D surface parameters. Sq, Sds, Sdq, Sdr have the same varying trend with the tool

wear state; Ssk and Sku reflect the surface amplitude distribution, with a new tool the surface

amplitude distribution has a standard normal Gaussian distribution and with a worn tool it will

departs from the Gaussian distribution; Str can reflect the variation of the surface texture

direction. Thus, a subset of surface texture roughness parameters including Sq, Sds, Str, Sdq,

Sdr is recommend as the indication to monitor the tool wear state.

Overall, the 3D surface parameter’s variation can quantify the tool wear effects on the

surface, the pattern features of AACF spectrum can reflect the surface texture’s variation with

tool wear increase. Thus, combined analysis of the surface roughness and its AACF

spectrum can be a choice to monitor the tool wear state and can be used to monitor other

machine states and could therefore provide a means to optimise the machining process.

The authors recognize that online 3D surface measurement of machined surfaces has yet

to be realized and a fast optical data capture technique appears to be the only option in this

case. However the above analysis is essential when such a data capture device is finally

developed and this is currently the subject of research by the present authors [14]. In addition

the authors also recognise the artificial nature of the generated tool wear. They are however

confident that the observed effects would be replaced for a “real’ tool during wear. In fact

the reported analysis techniques have now been applied to real tools and this work will be

reported in the near future.

6 Acknowledgements

The authors would like to thank EPSRC for the financial support given to this research work

under its programme GR/R13401/01.

References

[1] D.G. Ford, Novel Metrology-based Control Algorithm for Precision Manufacture

(CAPM), 2000, EPSRC Project Proposal; GR/R13401/01

[2] M.C. Shaw, Metal Cutting Principles, Clarendon Press, Oxford, 1984

[3] J.A. Schey, Introduction to manufacturing processes (Second ed.) McGraw-Hill Book

Company, London, 1987

[4] Modern metal cutting, publised by A B Sandvik Coromant, Sweden, ISBN

91-972299-0-3, 1994

[5] Hassui, A.E. Diniz, Correlating surface roughness and vibration on plunge cylindrical

grinding of steel, Int. J. Mach. Tools and Manufact. 43 (2003) 855-862

[6] D. J. Whitehouse, Surfaces a Link Between Manufacture and Function, Proc. Inst.

Mech. Engrs 192 (1978) 179-188

[7] K. J. Stout, P. J. Sullivan, W. P. Dong, E. Mainsah, N. Luo, T. Mathia, H. Zahouani,

The Development of Methods for the Characterisation of Roughness in three

Dimensions. published by Commission of the European Communities, Brussels, ISBN

0 7044 1313 2, 1993

[8] L. Blunt, X. Jiang, Advanced Techniques for Assessment Surface Topography, 1st edn,

London Penton Press, 2003

[9] E.J. Davis, P.J. Sullivan, K.J. Stout, The application of 3-D topography to engine bore

surfaces, Surface Topography 1 (1998) 229-251

[10] W.P. Dong, and K.J. Stout, An integrated approach to the Characterisation of surface

wear I: Qualitative charactrisation, Wear 181-183(1995) 700-716

[11] W.P. Dong, P.J. Sullivan, K.J. Stout, Comprehensive study of parameters for

characterising three-dimensional surface topography III: Parameters for characterising

amplitude and some functional properties, Wear 178 (1994) 29-43

[12] W.P. Dong, P.J. Sullivan, K.J. Stout, Comprehensive study of parameter for

characterizing three dimensional surface topography IV: Parameters for characterising

spatial and hybrid properties, Wear 178 (1994) 45-60

[13] P.J. Scott, An algorithm to extract critical points from lattice height data, Int. J. Mach.

Tools and Manufact. 41 (2001) 1889-1898

[14] X. Jiang, D. Lin, L. Blunt, W. Zhang, L. Zhang, Investigation of Some cirtical Aspects

of On-line Surface Measurement by a Wavelength-Division-Multiplexing Technique,

Measurement Science and Technology 17 (2006) 483-487

FIGURE CAPTIONS

Fig.1 3D parameter set

Fig.2 Development of accelerated tool wear across tool edge

Fig.3 3D surface topography in different tool wear amount (a) 0m, (b) 90m, (c) 110m,

(d) 130m

Fig.4 variation of 3D roughness amplitude parameters with tool wear

Fig.5 variation of 3D roughness spacing parameters with tool wear

Fig.6 variation of 3D roughness hybrid parameters with tool wear

Fig.7 variation of the of AACF with tool wear (a) 0m, (b) 90m, (c) 110m, (d) 130m
 Amplitude Spacing Hybrid Other
Parameters Parameters Parameters Parameters

 Root mean  Density of  Arithmetic  Curve related

square deviation summits – Sds mean peak  Sk family
– Sq  Fastest decay curvature - Ssc
 Ten point height auto-correlation  Root mean
of surface – Sz length - Sal square slope –
 Skewness – Ssk  Texture aspect Sdq
 Kurtosis - Sku ratio - Str  Developed
 Texture interfacial area
direction - Std ratio - Sdr

Fig.1 3D parameter set

(a) 0m (b) 90m (c) 110m (d) 130m

Fig.2 Development of accelerated tool wear across tool edge

 (a) 0m (b) 90m

(c) 110m (d) 130m

Fig.3 3D surface topography in different tool wear amount
 Sq Vs Tool Wear Ssk Vs Tool wear

0.9 0.5
0.8 0
0.7 -0.5

Ssk
Sq

0.6 -1
0.5 -1.5
0.4 -2
TW1 TW2 TW3 TW4 TW1 TW2 TW3 TW4

Sku Vs Tool Wear S5z VS Tool Wear

22 25
20
17
15

S5z
Sku

12
10
7 5
2 0
TW1 TW2 TW3 TW4 TW1 TW2 TW3 TW4

Fig.4 variation of 3D roughness amplitude parameters with tool wear

Sds Vs Tool Wear Str Vs Tool Wear

4.00E+02 0.8
0.6
3.00E+02
Sds

Str

0.4
2.00E+02
0.2
1.00E+02 0
TW1 TW2 TW3 TW4 TW1 TW2 TW3 TW4

Sal Vs Tool Wear

0.25
0.2
Sal

0.15
0.1
0.05
TW1 TW2 TW3 TW4

Fig.5 variation of 3D roughness spacing parameters with tool wear

 Sdq Vs Tool Wear Ssc Vs Tool Wear

0.06 6.50E-04
0.05
0.04 5.50E-04
Sdq

Ssc
0.03 4.50E-04
0.02
0.01 3.50E-04
TW1 TW2 TW3 TW4 TW1 TW2 TW3 TW4

Sdr Vs Tool Wear

0.25
0.2
0.15
Sdr

0.1
0.05
0
TW1 TW2 TW3 TW4

Fig.6 variation of 3D roughness hybrid parameters with tool wear

(a) 0m (b) 90m

(c) 110m (d) 130m

Fig.7 variation of the of AACF with tool wear
 Table 1 Cutting parameters

surface name TWnA TWnB TWnC TWnD TWnE TWnF TWnG TWnH

spindle speed(rpm) 796 796 796 796 796 796 796 796

radial depth(m) 200 200 200 200 200 200 600 200

axial depth(mm) 15 15 15 15 15 15 10 10

feedrate(mm/min) 239 215 271 215 239 271 239 239