Paper ID #23551

Understanding Surface Quality: Beyond Average Roughness (Ra)

Dr. Chittaranjan Sahay P.E., University of Hartford
Dr. Sahay has been an active researcher and educator in Mechanical and Manufacturing Engineering for
the past four decades in the areas of Design, Solid Mechanics, Manufacturing Processes, and Metrology.
He is a member of ASME, SME,and CASE.
Dr. Suhash Ghosh, University of Hartford
Dr. Ghosh has been actively working in the areas of advanced laser manufacturing processes modeling
and simulations for the past 12 years. His particular areas of interests are thermal, structural and materials
modeling/simulation using Finite Element Analysis tools. His areas of interests also include Mechanical
Design and Metrology.

American
c Society for Engineering Education, 2018
 Understanding Surface Quality: Beyond Average Roughness (Ra)

Abstract

Design of machine parts routinely focus on the dimensional and form tolerances. In applications
where surface quality is critical and requires a characterizing indicator, surface roughness
parameters, Ra (roughness average) is predominantly used. Traditionally, surface texture has
been used more as an index of the variation in the process due to tool wear, machine tool
vibration, damaged machine elements, etc., than as a measure of the performance of the
component. There are many reasons that contribute to this tendency: average roughness remains
so easy to calculate, it is well understood, and vast amount of published literature explains it, and
historical part data is based upon it. It has been seen that Ra, typically, proves too general to
describe surface’s true functional nature. Additionally, the push for complex geometry, coupled
with the emerging technological advances in establishing new limits in manufacturing tolerances
and better understanding of the tribological phenomena, implies the need for surface
characterization to correlate surface quality with desirable function of the surface. In turn, the
surface quality over the entire area, not just the 2D Ra parameter, dictates the performance and
reliability of the part.

Both ISO and ASME current standards on surface texture have a range of 3D surface quality
parameters. This is further aided by the availability of modern equipment to accurately measure
them. Despite these advances, design and quality professionals continue to specify surface finish
based solely on the value of Ra. The same outlook trails in graduate and undergraduate education
and their textbooks. This article explores how these multitudes of 2D and 3D surface quality
parameters are to be understood in the design and development of high performance surfaces,
and the strong need for them to be incorporated into graduate undergraduate engineering
curriculum, and be taught as an improved toolkit to the aspiring engineers, process engineers and
quality control professionals. Included case studies can be used to captivate the attention of the
students (target audience would include industry professionals as well) and route their
inquisitiveness into why they need to think beyond Ra in this era of advanced manufacturing.

Introduction

Choice of tool; feed and speed of the tool; machine geometry; and environmental conditions in
machining processes result in the irregularity of machined surface. This irregularity consists of
high and low spots machined into a surface by the tool bit. These peaks and valleys can be
measured and used to define the condition and sometimes the performance of the surface [1]. In
today’s world, there are more than 100 ways to measure a surface and analyze the results, but the
most common measurement of the mark made by the tool, or the surface texture, is the roughness
measurement.

Numerous articles [2-11] have tried to address the dependence of the condition of surface and its
performance on the machining process parameters. However, they have thus far, focused on only
one parameter, namely, Ra. The average roughness, Ra is also known as AA (arithmetic average)
or CLA (center line average). Traditionally, this has been the only roughness parameter to appear
on the drawings, and parts have been only been inspected for this parameter. Critical concepts in
wear, friction and lubrication, fatigue, etc. have been discussed and analyzed based on Ra alone.
While Ra remains useful as a general guideline of surface texture, it typically proves too general
to describe the surface’s functional nature. A surface with sharp spikes, deep pits, or general
isotropy may all yield the same average roughness value, Ra makes no distinction between peaks
and valleys, nor does it provide information about spatial structure [12].

In order to understand what other parameters exist to effectively and efficiently characterize
surface conditions, we need to understand them in their mathematical context as well as in the
context of their design/application requirements. In this paper, we will discuss the limitations of
2D parameters and, more importantly, how 3D parameters can be employed to provide greater
insight into surface finish and performance. Additionally, few case studies have been presented
to corroborate the same.

Mathematical Understanding of Surface

Quantifying surface irregularities means assessing them by categorizing them by height, depth,
and interval. They are then analyzed by a predetermined method and calculated per industrial
quantities standards. The form and size of surface irregularities and the way the finished product
will be used determine if the surface roughness acts in a favorable or an unfavorable way.
Painted surfaces should be easy for paint to stick to, while drive surfaces should rotate easily and
resist wear. It is important to manage surface roughness so that it is suitable for the component in
terms of quality and performance. There are close to 100 different parameters for describing
surface finishes or conditions. Selecting the most suitable can be difficult, and usually it takes at
least two to reasonably describe a surface. A typical engineering surface consists of a range of
spatial frequencies. The high frequency or short wavelength components are referred to as
roughness, the medium frequencies as waviness and low frequency components as form. Figure
1 illustrates this [13].

Figure 1. Form Error, Waviness and Roughness [13].

Historically, it has been accepted that different aspects of the manufacturing process generate the
different wavelength regimes and these affect the function of the part differently [13,14]. As
shown in figure 2, by separating surface profile into various bands, we can map the frequency
spectrum of each band to the manufacturing process that generated it. Thus, filtering of surface
profiles serves as a useful tool for process control and diagnostics.

Figure 2. Separation of surface into frequency bands [14].

While engineers commonly trace manufacturing process variations based on surface profile data,
mapping functional performance of a component based on surface profile information has been a
challenge. The different wavelength regimes play a key role in critical parts. Thus, separation of
signal into various bandwidths has to be viewed from a functional standpoint as well. Digital
filtering is a common practice to separate a signal into various regimes. Analog RC filters have
been replaced by the now common Gaussian filters. Form errors (like flatness, straightness,
roundness and cylindricity) include the crudest (highest wavelength) irregularities on the surface.
They are commonly measured using layout gages and Coordinate Measuring Machines (CMM)
based on Geometric Dimensioning and Tolerancing (GD&T) standards. These irregularities can
be quantified based on deviations measured from a nominal surface. Roughness includes the
finest (shortest wavelength) irregularities of a surface. Roughness generally results from a
particular production process or material condition. Waviness includes the more widely spaced
(longer wavelength) deviations of a surface from its nominal shape. Waviness errors are
intermediate in wavelength between roughness and form error. Lay refers to the predominant
direction of the surface texture.

Surface Roughness Measurement Tools (Contact-Type)

In most of the contact-based measurement tools, traces are done 90° to “lay” with a conical
diamond stylus. The radius of the stylus tip is usually in the range of 2-10 µm. In a skidded
measurement tool (as shown in figure 3), the skid and the diamond stylus are independent, and
are in contact with the surface. The skid and diamond follow the surface during measurement.
The surface deviations are measured by the change in the diamond position relative to the plane
of the skid. Skidded instruments measure only roughness. Waviness is filtered out by the skid
following the surface. Most portable instruments are skidded. In skidless instruments (as shown
in figure 4), the diamond stylus alone follows the surface during measurement. Deviations are
measured by the change in the diamond position relative to the drive datum guide. Skidless
instruments measure roughness, waviness and profile.
 Figure 3. Skidded Measurement Tool.

Figure 4. Skidless Measurement Tool.

Filters in Surface Roughness

To separate wavelengths, a filter is applied to the profile data. Digital filters commonly used are
RC type (simulated old analog electrical “resistor capacitor”) and Gaussian type. The user selects
the “cutoff” setting used by the filter to separate profile into roughness and waviness. Filtered
data is centered around a mean line. This is shown in figure 5.
 Figure 5. Filter Transmission and Cutoff.

The cutoff value is the longest nominal wavelength to be included in roughness. Wavelengths
longer than the roughness cutoff are included in waviness. Cutoff functions in a manner similar
to a sieve (shown in figure 6). The cutoff selected must be short enough to exclude long
wavelengths (waviness). Figure 7 demonstrates the effects of cutoff filter selection. For the same
profile data, selection of λc = 0.08 mm results in Ra = 0.560 µm and waviness Wa = 0.827 µm.,
while selection of λc = 0.8 mm results in Ra = 1.149 µm and waviness Wa = 0.229 µm.. Hence
it demonstrates that higher wavelengths were used in the second situation that raised Ra while
reducing Wa.

Figure 6. The role of roughness “Cutoff” (λc) filter.

 Figure 7. Effect of roughness “Cutoff” (λc) filter setting.

Mathematical Modeling of Profiling Methods

Although there are about 100 parameters in existence, for the brevity of this paper, only a few
have been explained here. For mathematical definition of the rest, reader can look up reference
[14]. Key quantities that distinguish one profile from another are their height deviations from
nominal profile and the distances between comparable deviations. Various mathematical
combinations of surface profile heights and spacings have been devised to compare certain
features of profiles numerically.

Height (z) Parameters

(1) Roughness Average (Ra)
Ra is the most commonly specified parameter in USA. Confusion exists between Ra and RMS
values. Ra is developed from RMS, but they are not same. As shown in figures 8 and 9,
Roughness average is the arithmetic average of the absolute values of the roughness profile
ordinates. Analytically, Ra is given by the following equation, where Z(x) is the profile height
function used to represent the point-by-point deviations between the measured profile and
reference mean line. For digital instruments, Z(x) is approximated by a set of digitized values
(Zi) recorded using the sampling interval (d0).

1 𝐿
𝑅𝑎 = ( ) ∫0 |𝑍(𝑥)|𝑑𝑥 (1)
𝐿

Figure 8. Filtered roughness profile with mean line, peak to valley is 10

 Figure 9. Steps for the calculation of Ra.

(2) Root Mean Square (rms) Roughness Average (Rq)

Rq is the root mean square average of the profile height deviations taken within the evaluation
length and measured from the mean line. Analytically, it is given by the following equation,

1 𝐿 1/2
𝑅𝑞 = [(𝐿) ∫0 (𝑍(𝑥))2 𝑑𝑥 ] (2)

The digitized approximation is as follows,

𝑅𝑞 = [(𝑍12 + 𝑍22 + 𝑍32 + ⋯ … … 𝑍𝑁2 )/𝑁]1/2 (3)

(3) Maximum Profile Peak Height (Rp)

It is the distance between the highest point of the profile and the mean line within the evaluation
length. Rpi is the distance between the highest point of the profile and the mean line within a
sampling length segment labeled i (figure 10).

(4) Average Maximum Profile Peak Height (Rpm)

It is the average of the successive values of Rpi calculated over the evaluation length.

(5) Maximum Profile Valley Depth (Rv)

It is the distance between the lowest point of the profile and the mean line within the evaluation
length (figure 10).

(6) Maximum Height of the Profile (Rt)

It is the vertical distance between the highest and lowest points of the profile within the
evaluation length (figure 10).
𝑅𝑡 = 𝑅𝑝 + 𝑅𝑣 (4)

(7) Average Maximum Height of the Profile (Rz)

It is the average of the successive values of Rti calculated over the evaluation length (figure 10).
Rz is the second most commonly specified parameter. Confusion exists since it can be calculated
five different ways. So the user must ensure to understand which standard governs. Do not
attempt to estimate Rz using a conversion factor. Ra, is not greatly influenced by spikes in the
profile. However, Rz, is greatly influenced by those spikes.

(8) Maximum Roughness Depth (Rmax)

It is the largest of the successive values of Rti calculated over the evaluation length (figure 11).

Figure 10. Rt, Rp and Rv Parameters.

Figure 11. Rt and Rmax Parameters.

Waviness Height Parameters

(1) Waviness Height, Wt
It is the peak-to-valley height of the modified profile from which roughness and part form have
been removed by filtering, smoothing, or other means. The measurement is to be taken normal to
the nominal profile within the limits of the waviness evaluation length (figure 12).

Figure 12. Waviness Height, Wt

Spacing Parameters
It is a distance that characterizes the lateral spacings between the individual profile asperities.
(1) Mean Spacing of Profile Irregularities (RSm)
It is the mean value of the spacing between profile irregularities within the evaluation length
(figure 13).

𝑅𝑆𝑚 = (1⁄𝑛) ∑𝑛𝑖=1 𝑆𝑚𝑖 (5)

Figure 13. The Mean Spacing of Profile Irregularities, RSm.

Shape Parameters and Functions

(1) Amplitude Density Function, ADF(z) or p(z)
It is the probability density of surface heights. The amplitude density function is normally
calculated as a histogram of the digitized points on the profile over the evaluation length (figure
14).
(2) Profile Bearing Length
It is the sum of the section lengths obtained by cutting the profile peaks by a line parallel to the
mean line within the evaluation length at a specified level p. The level p may be specified in
several ways including the following:
(a) As a depth from the highest peak (with an optional offset)
(b) As a height from the mean line
(c) As a percentage of the Rt value relative to the highest peak (figure 15).

Figure 14. Amplitude Density Function – ADF(z) or p(z)

(3) Profile Bearing Length Rati, tp

It is the ratio of the profile bearing length to the evaluation length at a specified level p. The
quantity tp should be expressed in percent.
 𝑏1+𝑏2+⋯+𝑏𝑛
𝑡𝑝 = × 100% (6)
𝐿

Figure 15. The profile Bearing Length

(4) Bearing Area Curve, BAC

It is (also called the Abbott-Firestone curve) related to the cumulative distribution of the ADF. It
shows how the profile bearing length ratio varies with level (figure 16).
(5) Skewness, Rsk
It is a measure of the asymmetry of the profile about the mean line calculated over the evaluation
length (figure 17). In analytical and digitized form,

1 1 𝐿
𝑅𝑠𝑘 = 𝑅 3
𝐿
∫0 [𝑧(𝑥)]3 𝑑𝑥 (7)
𝑞

1 1 3
𝑅𝑠𝑘 = ∑𝑁
𝑗=1 𝑍𝑗 (8)
𝑅𝑞3 𝑁

Figure 16. The Bearing Area Curve and Related Parameters.

 (6) kurtosis, Rku
It is a measure of the peakedness of the profile about the mean line calculated over the evaluation
length (figure 18). In analytical and digitized form,

1 1 𝐿3
𝑅𝑘𝑢 = 𝑅 4
𝐿
∫0 [𝑧(𝑥)]4 𝑑𝑥 (9)
𝑞

1 1
𝑅𝑘𝑢 = ∑𝑁 4
𝑗=1 𝑍𝑗 (10)
𝑅𝑞4 𝑁

Figure 17. Three Surface profiles with Different Skewness.

Figure 18. Three Surface profiles with Different Kurtosis.

Hybrid Parameters

(1) Linear Material ratio Curve Height Parameters

These are related to the profile bearing length in the linear material ratio curve, also known as the
bearing area curve. Parameters Rpk, Rk, Rvk, Mr1, and Mr2 derived from the linear material
ratio curve. As shown in figure 19, Material Ratio expresses the theoretical contact ratio at a
given cutting depth. In this sine wave example, 5 µm depth yields 50% contact ratio.

Figure 19. Example to calculate Material Ratio.

Figure 20 shows more profiles with their Bearing area Curves. Further, figure 21 shows that both
surfaces have the same Ra, however, their Material Ratio at a cutting depth of X is quite different
(the ratio of their Material Ratio is close to 4).

Figure 20. Bearing Area Curves.

 Figure 21. Same Ra surfaces with different Material Ratio

3D Parameters Offer More Details

3D surface parameters readily calculated from 3D topographical measurement data, highlight a

surface’s waviness, microroughness, wear ability and lubricant retention, as well as the angular
orientation of residual machining marks, and much more [12]. In the past decade, significant
efforts have been directed towards developing standard worldwide 3D parameters, the result of
which is a set of standard “S Parameters” in four general categories: amplitude, spatial, hybrid
and functional. Similar to 2D Parameters discussed earlier in this paper, the 3D parameters
commonly used now are,

Amplitude Parameters
Based on overall heights,
(1) Root Mean Square Deviation, Sq- RMS of height distribution
(2) Skewness, Ssk- the degree of asymmetry of a surface height distribution
(3) Kurtosis, Sku – the degree of peakedness of a surface height distribution
(4) Average Height, Sz – average of ten highest and lowest points.

Spatial Parameters
Based on frequencies of features
(1) Sds - Density of Summits,
(2) Str- Texture aspect ratio
(3) Sal – Fastest decay autocorrelation length
(4) Std – Texture direction of surface

Hybrid Parameters
Based on a combination of frequency and height
(1) SDq – Root Mean Square Surface Slope
(2) Ssc – Mean summit Curvature
(3) Sdr – Developed Surface Area Ratio
Functional Parameters
Based on applicability for particular functions
(1) Sbi – Surface Bearing Index
(2) Sci – Core Fluid retention Index
(3) Svi – Valley Fluid Retention Index

Typical applications for various 3D parameters are shown in figure 22.

Figure 22. Shaft Radial Lip Seal [12].

Results and Discussion

The following case studies reflect the very fact why Ra alone is not an indicator of the functional
traits of a surface. Additional parameters are required to specify the design and performance
intent.

Case Study #1
Figure 23 shows a roller contact bearing mounted on a shaft (tight fit). The shaded region (in red)
is the inner seal of the bearing in contact with the shaft. It is imperative that the contact area be
as high as possible for the inner seal to not slip ever over the shaft. Traditionally, however, the
design/drawing specifies only Ra on the inner surface of the seal. As can be seen in Figure 23 (a)
and (b), both conditions delineate to the same Ra value of the surface. However, (b) has clearly
more contact area as compared to (a). This is also obvious based on Material Ratio Curve ( also
known as bearing area curve), that (b) has a much higher bearing area, providing a higher
required contact area for the design to work effectively. Ra is often specified and is valuable for
monitoring process stability, however, it is high time that all drawings for surface finish use
additional parameters that may be needed to monitor for surface function.

(a) (b)

Figure 23. Surface profiles with same Ra, but different BAC

Case Study #2
Figure 24 shows a journal bearing whose surface had a Ra specification earlier. Although
inspection resulted in 100% acceptance, some of these bearings seized during operation.
Consequently, the design was revised to a lower 0.2Ra, hoping to fully address the issue. It was
however, realized that most of the bearings failed (seized) after the design revision. A detailed
investigation of profile analysis showed the presence of long wave forms. This poor bearing
contact was not quite obvious or visible by measuring Ra. Waviness height (Wt) was
subsequently specified to give visibility to this condition. No journals failed following this
revision.

Figure 24. Bearing Journal

Case Study #3
Figure 25 shows radial lip seal on a shaft in fluid flow design. The original design/drawing
specified Ra on the shaft-area of the lip seal. However, during inspection it was realized that with
a 0.25 mm cutoff (λc) observed Ra was rejecting parts. Further, lip seal length is shorter than 0.8
mm “default” cutoff. Further studies showed it is important to monitor peak heights Rpm as
peaks are abrasive to the lip seal. More elaborate methods of assessing seal surfaces include new
approaches to lead angle measurement.

Figure 25. Shaft Radial Lip Seal.

Case Study #4
Figure 26 shows three surfaces from a surface comparator strip generated by grinding, Blanchard
grinding and shape turning. Blanchard grinding, also known as rotary surface grinding, is used
to efficiently remove stock from one side of material with a large surface area. It is a far more
economical process than precision grinding. All three surfaces were found to have approximately
the same Ra (680-750 nm). Yet, the functional traits of these very different surfaces are
indistinguishable by Ra. Which surface will wear well? Which will retain fluid? Which will
survive a bearing load, or which is susceptible to stress cracking along machining marks? Ra
provides no information to answer these questions. [12].
 Figure 26. Optical Profiler Images of Surfaces from Grinding, Blanchard Grinding and
Shape Turning Processes [12].

Need in Manufacturing/Engineering curriculum and Education Plan

As explained in this article, despite the availability of measurement equipment and literature on
several surface quality indicators, Ra remains the sole parameter to be taught in engineering
curriculum, and worse, correlated to the functional nature of a surface. Authors believe that it is
high time that the discussed surface parameters, tools, filters and mathematical modeling of
methods be included in all design, manufacturing and capstone project courses (freshmen
through senior year). A better educated workforce would be able to contribute significantly
higher to quality tools and advanced metrology. It is an evolutionary process, i.e., standards keep
constantly changing; however, it must transition smoothly into the curriculum. It is imperative
that course instructors keep up to date with the latest standards, and implement them as short
interactive modules. In an ongoing effort, based upon the contents of this paper, the authors have
made several instructional modules on surface finish. These modules include power point
presentations, ready-to-implement instructor’s kit, in-class and homework problems, and well-
documented hands-on laboratory exercises. These modules and be readily used in existing
mechanical and manufacturing engineering programs, both undergraduate and graduate curricula.
Courses that can directly benefit and have strong potential for implementation are Mechanical
Design, Machine Design, All courses in Manufacturing Processes, Freshmen, Sophomore and
Senior Design courses, Surface Metrology, Precision Engineering, courses in Tribology, friction
and wear, etc. As a trial run, at the University of XXXXXXX, a 3-hour course module was
implemented in the Junior/Senior Manufacturing Processes course. This included a 1-hour
hands-on laboratory exercise. Four of these trial runs have been made since Spring 2016.
University of XXXXXXXX has a state-of-the-art Center for Manufacturing Metrology. It houses
both Contact (stylus-based) and non-contact (white light interferometry-based 3D surface
profiler) surface metrology equipment. The evaluation of the success of this module was based
on the metrics shown in Table 1. Results have been very encouraging thus far. The module has
been appreciated by our students, more by the graduate students who work full time in industry
as they see a direct relevance and impact on their technical skills and knowledge. Additionally,
student surveys are being developed that will be included in courses in Spring 2018. Further, a
survey will be made for the local industry people whose feedback on implementation of surface
finish standards in undergraduate and graduate programs will be sought.

Table 1. Evaluation Plan for the Module on Surface finish

Performance Objectives Evaluation Instrument Performance Criteria Results

Number of courses using Syllabus Minimum two courses 3 courses
Surface Finish Standards in curriculum

Diversity of SDO’s in Syllabus Minimum two SDO’s ASME, ISO

program
Number of students in Enrollment 25 Per Semester More than 50 Senior UG
courses More than 30 Grads
Breadth and Depth of Student/participation/IAC 7 on a scale of 10 9/10
Standards survey Graduate courses and
Implementation short courses
Extent of Standards Examination/Projects 3 in a scale of 4 4
Application

Conclusions

1. Authors, via these case studies, discovered that Ra is not necessarily an effective quality
screen or an adequate measure for development or problem solving.
2. Things can go wrong in surface finish gaging if one doesn’t understand and differentiate
different parameters like, Ra, Rq, Rz, etc.
3. Selecting the wrong gage (skidless vs. skidded) can affect accuracy.
4. Also, relying too much on a “default” cutoff filter or selecting the wrong cutoff filter can
compromise the results of roughness and waviness parameters.
5. Do not attempt to correlate different and unrelated parameters, especially now when we have
more than 100 parameters on surface finish.
6. 3D surface parameters provide more information about a surface than the conventional 2D
parameters. This can efficiently aid product design, development and prototyping.
7. It is high time instructors in design, manufacturing and senior capstone design projects
include discussion of these surface quality parameters in their courses Bite-size modules
(including case studies) dealing with mathematical understanding of surface, cut-off filters,
profile and waviness height parameters, spacing, shape and hybrid parameters must be
included.
References

[1] http://www.mmsonline.com/columns/surface-texture-from-ra-to-rz
[2] Paul, S., Dhar, N.R., and Chattopadhyay, A.B., 2001, “Beneficial effects of cryogenic
treatment over dry and wet machining on tool wear and surface finish in turning AISI 1060
steel,” Journal of Materials Processing Technology, 116, pp. 44-48.
[3] Luca, L., Neagu-Ventzel, S., and Marinescu, L., 2005, “Effects of working parameters on
surface finish in ball-burnishing of hardened steels,” Precision Engineering, 29(2), pp. 253-256.
[4] Schmitz, T.L., Couey, J., Marsh, E., Mauntler, N., and Hughes, D., 2007, “Runout effects in
milling: Surface finish, surface location error, and stability,” International Journal of Machine
Tools and Manufacture, 47(5), pp. 841-851.
[5] Benardos, P.G., and Vosniakos, G.C., 2003, “Predicting surface roughness in machining: a
review,” International Journal of Machine Tools and Manufacture, 43(8), pp. 833-844.
[6] Campbell, R.I., Martorelli, M., and Lee, H.S., 2002, “Surface Roughness Visualization for
rapid prototyping models,” Computer-Aided Design, 34, pp. 717-725.
[7] Bhattacharya, A., Das, S., and Majumder, P, 2009, “Estimating the effect of cutting
parameters on surface finish and power consumption during high speed machining of AISI 1045
steel using Taguchi design and ANOVA,” Production Engineering, 3(1), 31-40.
[8] Pavel, R., Marinescu, I., Deis, M., and Pillar, J., 2005, “Effect of tool wear on surface finish
for a case of continuous and interrupted hard turning,” Journal of Materials Processing
Technology, 170(1-2), pp. 341-349.
[9] Abouelatta, O.B., and Mádl, J., 2001, “Surface roughness prediction based on cutting
parameters and tool vibrations in turning operations,” Journal of Materials Processing
Technology, 118(1-3), pp. 269-277.
[10] Choudhury, S.K., and Bartarya, G., 2003, “Role of temperature and surface finish in
predicting tool wear using neural network and design of experiments,” International Journal of
Machine Tools and Manufacture, 43(7), pp. 747-753.
[11] Kurt, M., Bagci, E., and Kaynak, Y., 2009, “Application of Taguchi methods in the
optimization of cutting parameters for surface finish and hole diameter accuracy in dry drilling
processes,” 40(5-6), pp. 458-469.
[12] https://www.bruker.com/applications/metals/processing.html
[13] ASME B46.1, Surface Texture (Surface Roughness, Waviness, and Lay), 2009.
[14] Whitehouse, D. J., 1978, “Surfaces – a Link between Manufacture and Function”,
Proceedings of the Institution of Mechanical Engineers, Vol. 192, pp. 179-188.