Available online at www.sciencedirect.

com

ScienceDirect
Procedia CIRP 77 (2018) 235–238
www.elsevier.com/locate/procedia

8th CIRP Conference on High Performance Cutting (HPC 2018)

Cutting process monitoring with an instrumented boring bar measuring

cutting force and vibration.
Dan Östling*,a, Tormod Jensena, Mathias Tjomslanda, Oddvar Standala Terje Mugaasb
a
Sandvik Teeness AS, Ranheimsveien 127, 7053 Trondheim, Norway
b
SINTEF AS, Strindveien 4, 7054 Trondheim, Norway

* Corresponding author. Tel.: +47 90920498. E-mail address: [email protected]

Abstract

We present an industrialized version of a boring bar with embedded sensors and electronics that measure cutting forces and vibration. The novel
tool gives the operator or process developer valuable insight into the cutting process in components like jet engine shafts and landing gear. Critical
events like chatter and excessive insert wear can be detected and avoided and the quality of the cut can be documented and compared with earlier
cuts. Since the deflection of the bar is proportional to the cutting force, the actual machined diameter is measured in real time which reduces the
need for dimensional probing between cuts. In addition, the vibration measurement gives an indication of the quality of the machined surface and
can reveal problematic cutting parameters or component features. We present results from machining in both Maragin 250 and steel that clearly
shows the potential of the technology.

© 2018 The Authors. Published by Elsevier Ltd.

This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Selection and peer-review under responsibility of the International Scientific Committee of the 8th CIRP Conference on High Performance
Cutting (HPC 2018).
Keywords: Intelligent tools; internal turning; chatter; insert wear

1. Background and introduction of the cutting edge. The typically very hard materials used in
aerospace makes the situation worse.
The internal turning process is inherently difficult to monitor The low stiffness of the bar will also make the process
and develop, especially for many of the large components and vulnerable to changes in cutting forces. The radial cutting force
hard materials used in the aerospace industry. Machining deep will deflect the bar which reduces the actual cutting depth and
holes in jet engine shafts or landing gear, behind closed doors the resulting diameter of the machined hole [4]. Flank wear of
in a machine tool, is a “blind” process and it can be hard to the insert will lead to a gradually increasing radial cutting force
detect detrimental events like chatter, insert breakage or chip which results in a conical hole shape.
jams. Therefore, to develop a process that is stable with respect to
Due to its low stiffness, a slender boring bar will be chatter, insert life time, chip control, surface quality and
inherently prone to chatter and internal damping is necessary dimensional tolerances can be both time consuming and require
for a stable process [1]. Even so, as the tools are getting longer, considerable effort.
the stability margins are reduced, and it is necessary to tune the To give insight into the process we have developed a family
process parameters carefully to maintain stability. The stability of slender, damped boring bars equipped with embedded
margins can often be increased by reducing the nose radius of sensors and electronics for measurements of cutting force,
the insert [2], or by increasing the feed [3], but this may conflict deflection and vibration. The sensor signals are communicated
with the surface roughness requirements or the required lifetime wireless to a client PC or Pad giving the operator and process

2212-8271 © 2018 The Authors. Published by Elsevier Ltd.

This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Selection and peer-review under responsibility of the International Scientific Committee of the 8th CIRP Conference on High Performance Cutting
(HPC 2018).
10.1016/j.procir.2018.09.004

This is a resupply of March 2023 as the template used in the publication of the original article contained errors. The content of the article has remained unaffected.
 236 Dan Östling et al. / Procedia CIRP 77 (2018) 235–238

developer information about what is going on at the cutting Usually it is the resulting deviation in hole diameter and not
edge. It has been a major goal to deliver a robust and user- the deflection of the bar itself that is the target for calibration.
friendly product. In the process development phase, the system When machining, the radial cutting forces will affect not only
can be used to identify and avoid problematic features or cutting the tool but also the clamping unit and the workpiece. The
data, while in the production phase the results are used for deflection of all these parts causes the hole diameter to differ
process monitoring and documentation [5,6]. from the nominal diameter during cutting. These effects can be
Reference [7] gives a good overview of different machining accounted for by using the radial tool positioning servo to move
monitoring scopes and solutions. Of these, it is the prediction the front of the tool in contact with the work piece and then
of tool wear that has received most attention [8]. Many of the continue with a stepwise bending of the tool while measuring
solutions are based on measurements of either cutting force or the corresponding strain values as indicated in figure 2. From
vibration and we therefore expect that our combined solution the resulting strain-position curve we can then find the
will be able to address many of the most important monitoring calibration factor for the relationship between the measured
tasks. strain and the diameter deviation during cutting.
The paper presents the system in general and shows several
examples of how the monitoring can be used to detect and x
correct various problems like tool deflection, insert wear and
breakage, chip jams and chatter buildup. These are problems x
that will lead to dimensional errors, conical holes and surface
defects.

2. System description
Fig. 2. Deflection calibration by stepwise bending of the tool.
Figure 1 shows the damped boring bar with electronics
mounted for test in a lathe. The bar is equipped with 2.2. Acceleration sensors
accelerometers in the front and strain gauges in the back close
to the clamping area. The back of the bar is reinforced by The accelerometers are of the MEMS type and can measure
tungsten carbide rings that also protect the strain gauges. The acceleration in radial and tangential direction from DC up to a
accelerometer signals are high-pass filtered and contain few hundred Hz, limited by hardware filtering. The DC
information about the vibrations of the tool tip, while the low- capability means that the sensors can measure the gravity and
pass filtered strain signals are proportional to the more slowly this is used for factory calibration of the sensitivity. This
varying cutting forces that bend the tool. Both sensor systems feature is also useful for tool leveling, and the user interface
measure in radial and tangential directions separately. The contains a module that helps the operator to find the right
sensor signals are transferred via cable to a battery powered orientation of the tool with respect to the radial and tangential
electronic box that multiplexes the signals and transfers the machine axes.
measurement data via blue-tooth to the user software running
on a Windows10 platform. 3. Measurement examples

2.1. Force and deflection sensors 3.1. Insert wear and chipping in Maragin 250

Both the cutting forces and the resulting tool deflection can To illustrate the capabilities of the tool we did some test
be determined from the measured strain values. The calibration machining of Maragin 250 in the form of a scrapped jet engine
factors can be determined by loading the bar with a known shaft. The surface layer of the material is hard and abrasive
force and by bending the bar a known amount using the (HRC50) so finding the optimal cutting parameters can be
machine tool positioning servos. The force calibration value tricky. Table 1 describes a sequence of cuts with varying
depends only on the tool rigidity and can be factory calibrated cutting data (Fn=feed, Apnom=nominal cutting depth, Ønom=
either by loading the tool with weights or by using a force programmed diameter, Øact=actual diameter measured by a
sensor. digital caliper, Apact=Ønom-Øact, Deflect=Apnom-Apact).

Table 1. Turning in Maragin 250 with insert TR-DC1308-F, nose radius

Rn=0.4 mm, cutting speed Vc=120 m/min, and initial diameter 84.5 mm.
Cut Fn Apnom Ønom Øact Apact Deflect
(mm/r) (mm) (mm) (mm) (mm) (mm)
1 0.2 1 86,5 85,8 0,65 0,35
2 0.2 1 87,8 86,85 0,52 0,48
3 0.16 1 88,85 87,8 0,48 0,52
4 0.16 0.7 89,2 88,75 0,48 0,22
5 0.12 0.7 90,15 89,72 0,48 0,22
6 0.12 0.5 90,72 90,3 0,29 0,21
7 0.08 0.5 91,3 90,89 0,30 0,20
8 0.08 0.3 91,49 91,18 0,14 0,16
Fig. 1. The tool with electronics for wireless sensor data transfer to user.

This is a resupply of March 2023 as the template used in the publication of the original article contained errors. The content of the article has remained unaffected.
 Dan Östling et al. / Procedia CIRP 77 (2018) 235–238 237

All cuts in table 1 were made with a TR-DC1308-F insert The insert wear continues in the next cut and at the start of
with nose radius Rn=0.4 mm at cutting speed Vc=120 m/min. cut 3, the radial cutting force is larger than the tangential as
When in cut, the radial cutting force will deflect the tool and shown in figure 5. This is not a normal situation for the selected
cause the diameter of the machined hole to be smaller than the cutting parameters and eventually, 17mm into cut 3, we get a
programmed diameter. Since the deflection of the tool is chipping of the cutting edge. This leads to a sharper edge that
measured using the strain gauges we know the position of the cuts slightly deeper which causes a sudden drop in the radial
insert edge and therefore also the resulting machined diameter. force and a corresponding increase in the tangential force.
The measurements corresponding to the 8 cuts in table 1 are
Cutting forces with insert chipping
shown in figure 3. The insert enters the cylindrical work piece 600
at z=0 and exits at z=33 mm. Since the nominal diameter of cut 3
radial
tangential
cut1 is 86.5 mm the plot starts at this value but as the insert 500
engages the work piece, the diameter changes due to the
deflection. This cut also shows an accelerating deflection 400
which is characteristic for increasing insert wear.

Force (N)
300
Cut history Maragin250
92

cut 8 200
91 cut 7

cut 6
100
90
cut 5
Diamater Ø (mm)

89 cut 4: New insert 0

0 5 10 15 20 25 30 35
Distance z (mm)
88 cut 3

Figure 5. Measured tangential and radial cutting forces for cut 3.

87 cut 2

cut 1
After changing insert and reducing the cutting depth and
86
feed we get a more stable situation as shown in figure 6.
85
0 5 10 15 20 25 30 35 Cutting forces small cutting depth
Distance z (mm) 600
radial
cut 6
tangential
Figure 3. The measured diameter for the sequence of cuts in table 1. 500

600
Figure 4 shows the recorded tangential and radial cutting
Cutting forces with insert chipping
400
cutforces
3 for cut 1, and we see thatradial as the wear develops, the radial
Force (N)

tangential
500 force increases, while the tangential force is reduced due to the 300
increased deflection. The oscillation in the tangential signal is
400
caused by a slightly varying cutting depth caused by imperfect 200
centering of the rotating work piece. This offset is almost
Force (N)

300
removed in the first cut. 100
200
Cutting forces with off senter alignment and insert wear
600 0
100 0 5 10 15 20 25 30 35
radial
cut 1 Distance z (mm)
tangential
0 500
0 5 10 15 20 25 30 35
Distance z (mm) Figure 6. Measured tangential and radial cutting forces for cut 6.
400
3.2. Profiling in steel
Force (N)

300
Another interesting application for a measuring tool is to
200 find the effect of the varying cutting conditions in different
positions along a profile. In the programmed profile shown in
100
figure 7, the engagement angle of the insert and the resulting
chip flow will change along the profile. This will cause the
0
radial cutting forces to change as well, and the machined profile
0 5 10 15 20 25 30 35 will therefore differ from the programmed. The cut starts at z=0
Distance z (mm)
and proceeds towards the left in the graph. The nominal cutting
Figure 4. Measured tangential and radial cutting forces for cut 1.
depth is 0.5 mm all along the pre-machined profile. The upper
plot shows both the programed path and the actual path of the
tool tip. The lower plot shows the difference caused by the tool

This is a resupply of March 2023 as the template used in the publication of the original article contained errors. The content of the article has remained unaffected.
 238 Dan Östling et al. / Procedia CIRP 77 (2018) 235–238

deflection. Knowing the difference makes it possible to tune Figure 8 shows a test cut in 34CrNiMo6 (HRC 34) with
the process and implement compensation schemes. insert DCMT04-PF, cutting speed Vc=200 m/min and cutting
depth ap=1 mm. The feed is manually reduced from 0.1 mm/r
Profiling in steel
to 0.02 mm/r in steps of 0.01 mm/r. The cutting forces in figure
Radial position (mm)

172 8a) are gradually reduced because of the reduced feed. The
accelerometer signal in figure 8b) clearly shows that the
170 vibration level is also gradually reduced until chatter starts to
168
Program build up at low feed and dies out when the feed is increased
Actual
again.
-60 -50 -40 -30 -20 -10 0 By fourier transforming the accelerometer signal we can
z (mm)
monitor the frequency spectrum and warn the operator if a
0
chatter vibration starts to build up. The measurement is very
Tool Deflection (mm)

-0.05
sensitive and will react long before the chatter can be heard or
-0.1
seen as a surface defect.
-0.15
-0.2 4. Conclusion and further work
-0.25
-60 -50 -40 -30 -20 -10 0
z (mm) We have presented a damped boring bar with embedded
sensors that targets the most important deviations in the internal
Figure 7. Results from a profiling operation in 34CrNiMo6 (HRC 34). The turning process. The focus of the paper is to highlight some of
deflection causes the actual diameter to differ from the programmed. The the possibilities that lies in this technology without going into
difference varies along the profile due to changing cutting conditions.
too much detail. Work is now ongoing in developing and
implementing signal processing algorithms that can provide
3.3. Detecting vibration level and chatter
information at a higher level than the measurement data we
have shown here. The intention is to continually improve the
Regenerative chatter is often the factor that limits the boring
user value by implementing new process knowledge into smart
operation. It is well known that long tools are problematic and
algorithms that provide the operator with high quality
internal damping is necessary but not always sufficient to
information about the events that takes place inside the
obtain a stable operation. Usually it is the finishing operation,
machined component.
where the cutting depth and feed is small that is most critical
with respect to chatter. Increasing the feed or reducing the nose
References
radius of the insert will often help to suppress chatter but this
will also worsen the theoretical surface roughness. Due to this [1] Rivin E, Kang H. Enhancement of dynamic stability of cantilever tooling
tradeoff between stability and surface finish it is often desirable structures. Int. J. of Machine Tools and Manufacture 1992;32:539–561.
to operate close to the stability limit, which means that if the [2] Rahman M, Matin MA. Effect of tool nose radius on the stability of turning
material properties vary, the same cutting parameters can be processes. J. of Materials Processing technology 1991;26:13-21.
stable in one component and unstable in another. Detecting [3] Landers RG, Ulsoy AG. Nonlinear feed effect in machining chatter
analysis. Transactions of ASME J. of Manufacturing Science and
chatter at an early stage is therefore very interesting for the Engineering 2008;130.
operator. [4] Altintas Y. Manufacturing automation, Cambridge University Press; 2000.
[5] Stavropoulos, P., Chantzis, D., Doukas, C., Papacharalampopoulos, A., &
Cutting in SS2541 - varying feed Chryssolouris, G. (2013). Monitoring and control of manufacturing
200 processes: A review. Procedia CIRP, 8, 421-425.
[6] https://www.sandvik.coromant.com/en-gb/products/silent_tools_turning
Cutting force (N)

150
/Pages/silent-tools-plus.aspx
100 [7] Teti R, Jemielnak K, O’Donnell G, Dornfield D. Advanced monitoring of
machining operations, CIRP Annals – Manufacturing technology
50 radial
tangential
2010;59:717-739.
0 [8] Ambhore N, Kamble D, Chinchanikar S, Wayal V. Tool condition
0 10 20 30 40 50 60 70
Time (s) monitoring system: A review, Materials Today: Proceedings 2015;2:3419-
3428.
40
Acceleration (m/s )
2

20

-20 tangential
radial
-40
0 10 20 30 40 50 60 70
Time (s)

Figure 8. Cutting in steel with a feed that is reduced from 0.1mm/r to

0.02mm/r in steps of 0.01mm. a) Cutting force, b) accelerometer signal.

This is a resupply of March 2023 as the template used in the publication of the original article contained errors. The content of the article has remained unaffected.