Journal of

Manufacturing and
Materials Processing

Article
Using the Segmented Iterative Learning Control
Method to Generate Volumetric Error-Compensated
Part Programs for Three-Axis CNC Milling
Machine Tools
Ying-Chen Lu 1 and Syh-Shiuh Yeh 2, * ID

1 Institute of Mechatronic Engineering, National Taipei University of Technology, Taipei 10608, Taiwan;
[email protected]
2 Department of Mechanical Engineering, National Taipei University of Technology, Taipei 10608, Taiwan
* Correspondence: [email protected]; Tel.: +886-2-2771-2171




Received: 17 July 2018; Accepted: 10 August 2018; Published: 13 August 2018


Abstract: This study proposes using the iterative learning control method to adjust the volumetric
error-compensated tool path, where the working volume motion accuracy of three-axis computerized
numerical control (CNC) milling machine tools is increased by segmented modification of the part
program. As the three-axis CNC milling machine tools generally have volumetric error of working
volume, this study refers to the measured and established table of volumetric errors and uses
the method of the modifying part program for volumetric error compensation of machine tools.
This study proposes using part-program single-block positioning segmented for volumetric error
compensation, as the generated compensated part program with multiple compensated blocks can
effectively compensate the volumetric error of working volume in the tool moving process. In terms
of the compensated tool path computing method, this study uses the iterative learning control
(ILC) method and refers to compensated tool path and volumetric errors along the compensated
tool path for iterative computation. Finally, a part program with multiple blocks is modified
by the converged optimal compensated tool path, in order that the modified part program has
higher-precision volumetric error compensation effect. The simulation result shows that the rate of
improvement of error of the volumetric error compensation method proposed in this study is 70%.
The result of cutting tests shows that the average rate of improvement of the straightness error of the
test workpiece is 60%, while the average rate of improvement of height error is 80%. Therefore, the
results of simulation and cutting tests can prove the feasibility of using the ILC method for segmented
modification of the volumetric error-compensated part programs proposed in this study.

Keywords: iterative learning control; volumetric error; part program; CNC machine tools

1. Introduction
Manufacturing mechanical parts with higher accuracy has been the machining target of various
computerized numerical control (CNC) machine tools through the ages, as the conventional pitch
error measurement and compensation methods cannot meet the requirement. Using volumetric
error measurement and implementing the geometric error compensation of machine tool working
volume according to this measurement result have become an important method to improve
machine tool machining quality [1–5]. There are six static geometric errors resulting from the axial
movement of three-axis CNC milling machine tools, including displacement errors and angular errors.
The displacement errors are linear displacement errors, vertical straightness errors, and horizontal
straightness errors. The angular errors are pitch angular errors, yaw angular errors, and roll angular

J. Manuf. Mater. Process. 2018, 2, 53; doi:10.3390/jmmp2030053 www.mdpi.com/journal/jmmp

J. Manuf. Mater. Process. 2018, 2, 53 2 of 15

errors. In addition, the three-axis CNC milling machine tools have three squareness errors; therefore,
for the three-axis CNC milling machine tools, there are 21 static geometric errors, leading to the
volumetric error of the working volume of the machine tool [6–10].
The volumetric error has significant effect on the motion accuracy of a machine tool, and severely
influences the geometric dimension accuracy of the workpiece, thus, measurement and compensation
are required [11–16]. Xiang et al. [17] developed a volumetric error compensation technique for
real-time and simultaneous compensation of the volumetric errors of multiple machine tools by
integrating an ethernet distributed numerical control system with a generalized volumetric error
kinematic model, which was also developed in this study. Due to the increased importance of
integrating metrology systems to machine tools to online compensate volumetric errors of machine
tools for improving machining efficiency and accuracy, Wang and Maropoulos [2] developed a real-time
volumetric error compensation algorithm to improve the dynamic path accuracy of a machine tool
using a laser tracker-assisted 3-axis positioning system. Holub et al. [18] identified volumetric errors
of a small three-axis vertical machining center using a laser interferometer, and ballbar tests were used
to validate the volumetric error compensation results. Later, Holub et al. [19] performed on-the-fly
tracking experiments on the small three-axis machining center to measure the volumetric errors of
the small three-axis machining center using a self-tracking interferometer, and then compared the
measurement time and results with conventional trigger- or time-based measurements to show the
superior performance of on-the-fly measurement. Circular tests based on ISO 230-4 were also employed
for verification. Owing to the influence of thermal distortion on the manufacturing accuracy of large
machine tools, Gomez–Acedo et al. [20] developed a thermal distortion compensation system using
parametric state-space representation as the model architecture, and the Kalman filter as the estimator
for the thermal drift of the machine tool center point for different spindle speeds and temperatures of
the motor gearbox and room air, respectively. Later, Gomez–Acedo [21] developed a new methodology
for the measurement of thermal distortion in large machine tools using a laser tracking interferometer to
achieve low measurement uncertainty. The measurement method for the angular thermal distortion of
the horizontal plane was also developed by referring to the temperatures of the machine headstock and
column. López De Lacalle et al. [22] developed a methodology for the geometric accuracy estimation
of five-axis machine tools using the Denavit–Hartenberg (DH) convention. The assembly errors were
considered as the additional geometric parameters in DH transformation matrices in order to present
the imperfect tool position in absolute reference system; therefore, the developed methodology can be
applied to precisely estimate the dimensional errors of a manufactured workpiece. Based on the DH
convention, Díaz–Tena et al. [23] developed a methodology for the geometric accuracy assessment of a
multitasking machine consisting of a swivel tool-spindle head, two spindles, and a turret. Assembly
errors were also considered as additional geometric parameters in transformation matrices in order to
analyze the propagation and influence of assembly errors on the tool position and orientation in detail.
While there are high-end machine tool controllers with volumetric error compensation
functionalities, it is expensive, which limits the generality of volumetric error compensation.
To make the conventional machine tool controller have volumetric error compensation functionalities,
the software compensation methods of part-program generation and modification have become
an important trend to improve the influence of the volumetric error of machine tools [24–27].
Eskandari et al. [28] utilized a neuro-fuzzy algorithm to model the geometric errors of a machine tool
using a laser interferometer, and then, developed a compensation method to modify the Numerical
Control (NC) part program with G-codes, which consider the developed error model and the
volumetric errors calculated through kinematic analysis of the machine tool.
Among numerous volumetric error software compensation methods, using CAD/CAM to
generate part programs with volumetric error compensation generally has excessive complexity
and difficult operation. The design of the modifying part program calculates the volumetric error and
makes compensation according to the part program characterized tool moving path. While the method
is simple, its volumetric error compensation effect is limited. As the volumetric errors induced by the
J. Manuf. Mater. Process. 2018, 2, 53 3 of 15

tool moving process cannot be neglected, some studies segmented and compensated the tool moving
path [29–31]. Zhu et al. [32] employed statistical analysis methods to extract the component errors of a
machine tool, and then, a B-Spline mathematical model was further developed to represent component
error functions using the least-squares fitting method. Here, the NC part program with initial G-codes
was recursively subdivided and segmented according to the developed component error functions
with the calculated deviation errors. In addition, in order to improve the volumetric error induced by
the compensated tool path, some studies used the iteration method to compute more accurate tool
moving paths for machine tools [29,30]. Based on the measurement and modeling of volumetric errors,
Wang and Lin [27] developed a software-based error compensation method with an error interpolation
scheme to improve machining accuracy of a machine tool by recursively compensating the cutting
path and machining trajectory. Tool path deviation, as induced by the kinematic and geometric errors
of a CNC machine tool, usually corrupts the dimensional accuracy of the machined mechanical parts;
therefore, based on the developed error estimation model, Vahebi Nojedeh et al. [26] developed NC
program editor software to modify and generate the errors the compensated NC program by recursively
computing the compensated positions along the modified tool path. Lee et al. [33] developed a kind
of capacitance-sensor to measure the five geometric errors of a miniaturized machine tool, and then,
a volumetric error model was developed to synthesize and recursively compensate the geometric
errors of the machine tool to efficiently improve its positioning accuracy.
While the iterative computation of the volumetric error-compensated tool moving path can
effectively improve the influence of the volumetric errors in the tool moving process of a machine tool,
the unit recursive gain may fail to effectively converge to appropriate compensated tool path positions
due to the machine tool working volume error distribution. Therefore, this study refers to the measured
and established table of volumetric errors, as based on the part-program modification method, to
segment the tool moving path, and uses the concept of ILC (iterative learning control) adjustment
to design a segmented compensation method for the volumetric error of the part program, meaning
the part program can be modified to have volumetric error compensation effect with higher tool
moving accuracy. This study uses the Laser Doppler Displacement Meter (LDDM) laser displacement
measurement system and volumetric error analysis software to establish a table of the volumetric
errors of the machine tool [34,35], and then, performs path segmentation according to the part program
characterized tool moving path. Finally, the segmented tool moving path of the part program is
iteratively adjusted according to the concept of ILC. The ILC registers the actuation input, system
output, and output error information of the control system execution process through an iterative
computation process as reference for adjusting the control system actuation input of the next iteration,
in order that the output error can converge to a limited range within a limited number of iterative
learning. According to the segmented tool moving path of the part program, this study uses the
interpolation table of volumetric errors to compute the volumetric errors induced by the tool moving
path, and modifies the segmented tool moving path according to the volumetric errors. At this point,
to reduce the volumetric errors induced by the execution of the segmented tool moving path after
modification, the segmented tool moving path is adjusted by ILC iteratively till it converges to the
optimal segmented tool moving path.
The structure of this paper is described, as follows. Section 2 describes the experimental and
measuring equipment used in this study. Section 3 describes the volumetric error compensation
and calculation method proposed in this study, including computation for error compensation and
application of the iterative learning control method. Section 4 tests volumetric error compensation and
discusses the results, thus, the feasibility of the volumetric error compensation method, as proposed in
this study, can be validated. Section 5 summarizes this paper.

2. Introduction to Experimental System and Architecture

The three-axis CNC milling machine tool used in this study is shown in Figure 1. Its working
volume is 560 × 410 × 450 mm3 , the controller is a FANUC 32i series controller, which is suitable for
J. Manuf. Mater. Process. 2018, 2, 53 4 of 15

J. Manuf. Mater. Process. 2018, 2, x FOR PEER REVIEW 4 of 15

the application
J. Manuf. Mater.of a high-speed
Process. 2018, 2, x FORCNC
PEERmachine
REVIEW tool [36]. The machine tool uses a C-shaped mechanical 4 of 15
the application
structure; the Y-axisoffeed a high-speed CNC machineistool
drive servomechanism [36]. The
installed on themachine tool the
base, and usesballscrew
a C-shaped drives
mechanical
the structure;
application of a the Y-axis
high-speed feed
CNC drive servomechanism
machine tool [36]. The is installed
machine
the saddle. The X-axis feed drive servomechanism for driving the working table is installed on the on
tool the
usesbase,
a and
C-shaped the
ballscrew drives the saddle. The X-axis feed drive servomechanism for driving onthe
mechanical
Y-axis feed drive structure; the Y-axis
servomechanism, feed the
and drive servomechanism
workpiece is installed
to be machined is fixed toworking
the base,
the table
and
working is
thetable.
installed on the Y-axis feed drive servomechanism, and the workpiece to be machined is fixed to the
Theballscrew
Z-axis feed drives theservomechanism,
drive saddle. The X-axis whichfeed drive servomechanism
is perpendicular to for
thedriving
working thetable,
working table is on
is installed
working on
installed table.
theThe
Y-axisZ-axis feed drive servomechanism, which is perpendicular to the working table,
the column to control thefeed drive
vertical servomechanism,
movement of theand the workpiece
spindle. to be tool
The cutting machined
used isforfixed to
machining the is
is installed on the column to control the vertical movement of the spindle. The
working table. The Z-axis feed drive servomechanism, which is perpendicular to the working table, cutting tool used for
installed on theis spindle.
machining installed on the spindle.
is installed on the column to control the vertical movement of the spindle. The cutting tool used for
machining is installed on the spindle.

Figure 1. Experimental CNC machine tool.

Figure 1. Experimental CNC machine tool.
Figure 1. Experimental CNC machine tool.
The LDDM™ laser displacement measurement system used in this study is shown in Figure 2,
TheThe
which LDDM™
measures
LDDMTM laser
the displacement
volumetric
laser errormeasurement
displacement of machine tool
measurement system
working
system used
used inin
volume this
this study
and
study is shown
generates
is shown the in Figure
forward
in Figure 2, 2,
whichand reverse
measures table
the of volumetric
volumetric errors
error ofon each
machine motion
tool axis.
workingThe laser
volume Doppler
which measures the volumetric error of machine tool working volume and generates the forward and displacement
generates the meter,
forward
LDDM™,
andandreverse
reverse istable
table anofinstrument
volumetric
of system
volumetric errors with
errors on1/1,000,000
on each
each motion
motion (1 axis.
ppm) displacement
axis.The
The laser
laser measurement
Doppler
Doppler accuracy.
displacement
displacement meter, meter,
The maximum
LDDM™,
LDDM is, an
TM is an measurement
instrument
instrument systemrange
system withis1/1,000,000
with 15 m, the (1standard
1/1,000,000 (1ppm) resolution is
ppm)displacement
displacement 0.01 ìm, theaccuracy.
measurement
measurement highest
accuracy.
measurement
The The maximum
maximum speed is 3600
measurement
measurement mm/s,
range is and
range isthe
15 m, thetemperature
15 m, range is
the standard
standard resolution 15.5
resolution°C to
is 0.01 is32the
ìm, °C.highest
0.01 In
ìm,this
the study,
highest
measurement all
measurementsspeed
measurement andis compensation
3600 mm/sec, andprocesses
the were ◦ performed
temperature range is ◦
15.5 under
°C to 32 similar
°C. In temperature
this study, all
speed is 3600 mm/s, and the temperature range is 15.5 C to 32 C. In this study, all measurements
(approximately and 20 °C) and air humidity (relative humidity approximately 50%)temperature
for ◦ a fair
and measurements
compensation processes compensation
were performedprocesses
underwere
similarperformed
temperature under similar
(approximately 20 C) and air
comparison.
(approximately The measurement
20 °C) and instrument
air humidity for the cutting
(relative test
humidity is the NVC322
approximately fully automatic
50%) for avision
fair
humidity (relative humidity approximately 50%) for a fair comparison. The measurement instrument
measuring machine produced by 3DFAMILY Technology for workpiece
comparison. The measurement instrument for the cutting test is the NVC322 fully automatic vision straightness error
for the cutting testThe
measurement. is the NVC322
TESA fully automatic
MICRO-HITE Plus M350 vision
heightmeasuring
gauge for machine
produced by produced
Chiyeung, by 3DFAMILY
is used for
measuring machine produced by 3DFAMILY Technology workpiece straightness error
Technology
height for workpiece
error measurement. straightness error measurement. The TESA MICRO-HITE Plus M350 height
measurement. The TESA MICRO-HITE Plus M350 height gauge produced by Chiyeung, is used for
gauge produced by Chiyeung, is used for height error measurement.
height error measurement.

(a) (b)
(a)
Figure 2. LDDM laser displacement measurement system for this(b) study. (a) LDDM light alignment
with close distance between source mirror and reflection mirror;
Figure 2. LDDM laser displacement measurement system for this study. (b) LDDM (a)light
LDDM alignment with far
light alignment
Figure 2. LDDM
distance laser
between displacement
source mirror and measurement
reflection system for this study. (a) LDDM light alignment
mirror.
with close distance between source mirror and reflection mirror; (b) LDDM light alignment with far
with close distance between source mirror and reflection mirror; (b) LDDM light alignment with far
distance between source mirror and reflection mirror.
distance between source mirror and reflection mirror.
3. Volumetric Error Calculation and Compensation
J. Manuf. Mater. Process. 2018, 2, 53 5 of 15

3. Volumetric Error Calculation and Compensation

As CNC machine tools have many coordinate systems, the part-program coordinate system
used in part program, as well as the machine coordinate system used in machine tool motion, are
described in this study. Generally speaking, for convenient workpiece machining, the operator
usually uses the part-program coordinate system for programming the part program according to
the engineering drawing. The original setting of the coordinate system varies with the tool moving
process or machining process. The machine coordinate system is the referential coordinate system
for mechanical system motion and is a real coordinate system. The origin of the coordinate system
(machine origin) is the reference point of the machine tool, which is usually determined by the
machine tool manufacturer. In the workpiece machining process, the part-program coordinate system
is sometimes different from the machine coordinate system. As the machine tool volumetric error
measurement refers to the machine coordinate system, the measured values of errors are represented by
their positions in the machine coordinate system, thus, this study performs segmented compensation
of volumetric errors by the part-program modification method. The part program characterized tool
moving path must be transformed from the part program coordinate system to the machine coordinate
system, and then, the machine coordinate system refers to the measured and established table of
volumetric errors to compute the appropriate volumetric error compensation positions and values.
The general transformation of coordinate systems is applicable to the tool moving path transformation
between part-program coordinate system and machine coordinate system [37].
When the part program characterized tool moving path is transformed from the part program
coordinate system to the machine coordinate system, the calculation and compensation of volumetric
error in the tool moving process are performed. As shown in Figure 3, if the tool is to move from
→
current position R (machine coordinate system represented R) to target position T (machine coordinate
→
system represented T), due to the influence of the volumetric error of the machine tool, the cutting
→ →
tool moves to actual position P (machine coordinate system represented P ), which leads to PT tool
moving position error. As the table of the volumetric errors of the machine tool has been measured
→
and established, the tool moving position error PT can be calculated by the table of volumetric errors,
which can be used as the reference for the calculation of volumetric error-compensated position Q
→
(machine coordinate system represented Q). The volumetric error-compensated position Q can be
calculated according to the following steps:
→
(S1) The tool moving direction from position R to position T (direction of RT vector) determines the
load of the table of the forward volumetric errors or table of the reverse volumetric errors.
→ →
(S2) The volumetric error vector E (=PT) of the tool moves from position R to position T and is
calculated by the loaded table of the volumetric errors.
→
(S3) The tool moving T position and the calculated volumetric error vector E are added up to generate
the volumetric error-compensated position Q.
→
The volumetric error vector E is calculated by interpolation according to the table of volumetric
→ →
errors and RT vector. For example, as shown in Figure 4, if the RT vector means the X-axis moves from
the origin position to the A position in the machine coordinate system, and it is 0 mm to 15 mm in the
table of volumetric errors on X-axis, the X-axis movement results in the volumetric error on X-axis
direction Dxx(A), expressed as Equation (1):

Dxx(15) − Dxx(0) Dxx(15) − Dxx(A)

= (1)
15 − 0 15 − A

Dxx(15) represents the volumetric error on the X-axis direction when the X-axis moves to 15 mm
in the table of volumetric errors. Dxx(0) represents the volumetric error on the X-axis direction
when X-axis is at 0 mm in the table of volumetric errors. Therefore, when volumetric error vector
J.J. Manuf.
Manuf. Mater.
Mater. Process.
Process. 2018,
2018, 2,
2, xx FOR
FOR PEER
PEER REVIEW
REVIEW 66 of
of 15
15
J. Manuf. Mater. Process. 2018, 2, 53 6 of 15

coordinate system
coordinate system represented
represented Q Q ),), which
which is is equal
equal to to position
position T T (machine
(machine coordinate
coordinate system
system
→  
represented
represented T ))the
T
E is calculated, plus volumetric
volumetric
plus volumetric error vector
vector E
error-compensated
error E :: position Q can be calculated (machine coordinate
→    →
system represented Q), which is equal to position Q
Q =
= TT +(machine
+ E
E coordinate system represented T) plus (2)
(2)
→
volumetric error vector E : → position
→ → Q can be calculated by Equation (2), there
While the
the volumetric
volumetric error-compensated
error-compensated
While Q= T + E Q can be calculated by Equation (2), there
position (2)
are two
are two principal
principal problems
problems in in practical
practical application:
application:
While the volumetric error-compensated position Q can be calculated by Equation (2), there are
● The
● The compensation
compensation and and calculation
calculation of of Equation
Equation (2) (2) only
only consider
consider the the error
error compensation
compensation of of
two principal problems in practical application:
target position
target position T, T, while
while the
the position
position error
error compensation
compensation of of aa tool
tool moving
moving from from R R to
to TT is
is not
not
• considered.
The compensation
considered. In and
other calculation
words, the of Equation
position (2) only
accuracy consider
of the the
tool error
moving
In other words, the position accuracy of the tool moving process cannot be compensation
process of
cannottarget
be
position T,
guaranteed.
guaranteed. while the position error compensation of a tool moving from R to T is not considered.
● The
● In
Theother words, the and
compensation
compensation position
and accuracy
calculation
calculation ofof
of the tool (2)
Equation
Equation moving
(2) process
disregard
disregard thecannot
the effect be
effect guaranteed.
of the
of the volumetric
volumetric error
error
• of of the
Thethe machine
compensation tool when
and the tool
calculation moves
of Equationfrom position
(2) disregard R to
the position
effect
machine tool when the tool moves from position R to position Q . In other words, the of Q
the. In other words,
volumetric error the
of
the machine
actual
actual position
position toolof
ofwhen
the the tool
the tool
tool cannot
cannotmoves
be from
be positionand
guaranteed,
guaranteed, R tothe
and position
the targetQ.
target In otherto
position
position towords, the actual
be reached
be reached has
has
position of the
position error.
position error. tool cannot be guaranteed, and the target position to be reached has position error.

Figure 3.
Figure 3. The
The influence
The influence of volumetric
influence of volumetric errors
errors in
in the
the tool
tool moving
moving process.
process.

To
To overcomethe
overcome
To overcome theabove
the aboveproblems,
above problems,the
problems, the
the original
original
original tool
tool
tool moving
moving
moving path
path
path is segmented.
is segmented.
is segmented. Figure
Figure
Figure 33 shows
3 showsshows
that
that
that target
target
target position
position
position T becomes
T becomes
T becomes the intermediate
the intermediate
the intermediate target position
target position
target position on the segmented
on the segmented
on the segmented tool moving
tool path,
tool moving moving path,
path,
meaning
meaning
meaning
the positiontheaccuracy
the positionofaccuracy
position accuracy of the
of
the original the original
tool original tool moving
tool
moving process moving process
can beprocess
improved canbybe
can be improved
theimproved
volumetric byerror
by the
the
volumetric
volumetric
compensation error
error compensation
of compensation
the segmentedof of themoving
the
tool segmented
segmented tool
path.tool moving
moving
To make path.
thepath.
actualTotool
To make
make the actual
the
positionactual tool position
and tool position
intermediate
and
and intermediate
targetintermediate target
target
position on the position
positiontool
segmented on the
onmoving segmented
the segmented tool
path havetool moving
moving
small position path
path have
have
error, small
smallfrom
differing position
position error,
error,
the present
differing
differing from the
from the present
iterative computation present
method, iterative
iterative
this studycomputation
computation
uses the ILC method,
method, this
conceptthis study uses
study
to iterativelyuses the ILC
the
compute ILC concept
the concept
volumetric to
to
iteratively
iteratively compute
compute the
the
error-compensated position. volumetric
volumetric error-compensated
error-compensated position.
position.

Figure
Figure 4.
Figure 4. Interpolation
4. Interpolation computing
Interpolation computing method
computing method (Dxx).
method (Dxx).
(Dxx).

ILC was
ILC was proposed
proposed by by Uchiyama
Uchiyama [38],[38], which
which adjusts
adjusts oror modifies
modifies the
the actuation
actuation input
input of
of aa
ILC was proposed by Uchiyama [38], which adjusts or modifies the actuation input of a controlled
controlled plant by the iterative approaching control method, in order that the actual
controlled plant by the iterative approaching control method, in order that the actual output of the output of the
plant by the iterative approaching control method, in order that the actual output of the controlled
controlled plant
controlled plant can
can converge
converge to to the
the referenced
referenced output
output within
within limited
limited numbers
numbers of of iterative,
iterative, thus,
thus, it
it
plant can converge to the referenced output within limited numbers of iterative, thus, it can be
can be implemented in the manipulators and machine tools for executing repeated
can be implemented in the manipulators and machine tools for executing repeated machine motions machine motions
implemented in the manipulators and machine tools for executing repeated machine motions [39,40].
[39,40]. The
[39,40]. The development
development of of the
the ILC
ILC theory
theory receives
receives the
the interest
interest and
and attention
attention ofof control
control field
field
The development of the ILC theory receives the interest and attention of control field scholars; as
scholars; as
scholars; as its
its control
control structure
structure isis simple,
simple, itit is
is applicable
applicable toto both
both linear
linear systems
systems andand nonlinear
nonlinear
its control structure is simple, it is applicable to both linear systems and nonlinear systems and can
systems and can provide effective and practical solutions for repetitive
systems and can provide effective and practical solutions for repetitive tracking control and tracking control and
provide effective and practical solutions for repetitive tracking control and repetitive disturbance
repetitive disturbance rejection problems [41–43]. ILC uses the current actuation input of aa
repetitive disturbance rejection problems [41–43]. ILC uses the current actuation input of
rejection problems [41–43]. ILC uses the current actuation input of a controlled plant and the execution
controlled plant
controlled plant and
and the
the execution
execution output
output error
error as
as the
the calculation
calculation reference
reference for
for next
next actuation
actuation input
input
output error as the calculation reference for next actuation input of the controlled plant, thus, there
of the controlled plant, thus, there is appropriate control system execution output
of the controlled plant, thus, there is appropriate control system execution output error correction error correction
J. Manuf. Mater. Process. 2018, 2, 53 7 of 15

is appropriate control system execution output error correction ability, and the error decreases as
the numbers of iterative learning increases. When the numbers of iterative learning approaches
to infinity, the theoretical error approaches to zero. In this study, ILC is used for the iterative
computation of the volumetric error-compensated tool position for the part program characterized
tool moving path. When the calculated tool moving position error converges, the high-accuracy
volumetric error-compensated tool position can be obtained, to modify the tool moving path of the
part program. By referring to the volumetric error-compensated position computing equation, as
shown in Equation (2), the iterative computation of the ILC volumetric error-compensated position is
designed and expressed as Equation (3).
→ → →
Qi+1 = Qi + γ · E i , i = 0, 1, . . . , n (3)
→
where position Qi (machine coordinate system represented Qi ) represents No. (i) calculated volumetric
→
error-compensated position, position Qi+1 (machine coordinate system represented Qi+1 ) represents
→
No. (i + 1) calculated volumetric error-compensated position, E i represents the calculated volumetric
error of a tool moving from the current position to No. (i) volumetric error-compensated position Qi ,
→
γ represents the learning factor, 0 < γ < 1. Q0 equals target position T (machine coordinate system
→ →
represented T), and E 0 represents the calculated volumetric error of the tool moving from the current
→
position to target position T (machine coordinate system represented T). The setting of learning factor
γ affects the convergence of volumetric error-compensated position calculation. When the learning
factor value is large, the volumetric error-compensated position calculation converges rapidly, and
there is fluctuation, thus, it fails to achieve stable calculation results. When the learning factor value is
small, while the volumetric error-compensated position calculation result is stable, the convergence rate
is low. When learning factor γ is larger than 1, the volumetric error-compensated position calculation
may fail to converge. As this study uses the offline computing method of Equation (3), a small learning
factor value is used to guarantee stable volumetric error-compensated position calculation results.
The calculation convergence of Equation (3) is judged based on whether the Euclidean norm of No. (i)
→
calculated volumetric error E i decreases as the number of iterative computation increases. In other
words, this study uses Equation (3) for the iterative computation of volumetric error-compensated
→ →
position Qi , and checks whether the Euclidean norm (k E i k) of calculated volumetric error E i decreases
as the number of iterative computation (i) increases. When Equation (3) calculates convergence
iteratively, the converged volumetric error-compensated position is used as the tool moving path of
the modified part program.

4. CNC Volumetric Error Compensation Experiment Results and Discussion

The machine tool is often provided with compensation parameters for general pitch error
compensation; however, this pitch error compensation only compensates the displacement error
induced by axial motion and neglects the displacement errors induced by other motion axes. Therefore,
this study discusses the displacement error due to axial motion Exx , Eyy , Ezz , and considers the
displacement errors induced by other motion axes Exy , Exz , Eyx , Eyz , Ezx , Ezy . Where Eij represents the
displacement error on j-axis during i-axis movement. This study considers the aforesaid displacement
errors and generates the error-compensated part program. The table of the volumetric errors of the
experimental machine tool is measured by the LDDM laser Doppler displacement meter. The error
distributions of the forward and reverse motions on each motion axis are shown in Figure 5.
J. Manuf. Mater. Process. 2018, 2, 53 8 of 15
J. Manuf. Mater. Process. 2018, 2, x FOR PEER REVIEW 8 of 15

(a) (b)

(c) (d)

(e) (f)

Figure 5. Volumetric errors of the experimental machine tool measured by LDDM. (a) X-axis forward
Figure 5. Volumetric errors of the experimental machine tool measured by LDDM. (a) X-axis forward
error; (b) X-axis
error; (b) X-axisreverse
reverseerror;
error;(c)(c) Y-axis
Y-axis forward
forward error;
error; (d) (d) Y-axis
Y-axis reverse
reverse error;error; (e) Z-axis
(e) Z-axis forward
forward error;
error; (f) Z-axis reverse
(f) Z-axis reverse error. error.

4.1. Simulation Results and Discussion

4.1. Simulation Results and Discussion
This study designs two different tool moving path simulations for observation, segmentation,
This study designs two different tool moving path simulations for observation, segmentation,
and compensation for the circular path, and segmentation and compensation for the spiral path.
and compensation for the circular path, and segmentation and compensation for the spiral path.
Simulation 1 sets the tool movement path as a circle with a diameter of 20 mm, which is divided into
Simulation 1 sets the tool movement path as a circle with a diameter of 20 mm, which is divided into
20 segment and 30 segment moving paths, as shown in Figures 6 and 7, respectively. Simulation 2
20 segment and 30 segment moving paths, as shown in Figures 6 and 7, respectively. Simulation 2
sets the path as a spiral path with a diameter of 20 mm and a pitch of 1 mm, which is divided into 60
sets the path as a spiral path with a diameter of 20 mm and a pitch of 1 mm, which is divided into
segments of moving paths, as shown in Figure 8. Table 1 shows the numerical analysis results of
60 segments of moving paths, as shown in Figure 8. Table 1 shows the numerical analysis results of
simulation. It is obvious that the segmented tool moving path position error root-mean-square value
simulation. It is obvious that the segmented tool moving path position error root-mean-square value is
is improved after ILC compensation, where the rate of improvement is higher than 70%,
improved after ILC compensation, where the rate of improvement is higher than 70%, demonstrating
demonstrating the feasibility and effect of the ILC volumetric error compensation computation
the feasibility and effect of the ILC volumetric error compensation computation method, as designed in
method, as designed in this study, on the volumetric error compensation of a three-axis CNC milling
this study, on the volumetric error compensation of a three-axis CNC milling machine tool. In addition,
machine tool. In addition, according to the analysis of the number of segments on the tool moving
according to the analysis of the number of segments on the tool moving path, the larger the number of
path, the larger the number of segments, the better the volumetric error compensation effect.
segments, the better the volumetric error compensation effect. However, many segments have limited
However, many segments have limited compensation effect on the volumetric error, as it prolongs
compensation effect on the volumetric error, as it prolongs the computation time of compensation
the computation time of compensation method on the contrary. Therefore, the position of a single
method on the contrary. Therefore, the position of a single block of the part program shall not be
block of the part program shall not be excessively segmented in practical application.
excessively segmented in practical application.
 J. Manuf. Mater. Process. 2018, 2,2,53 99of 15
J.J.Manuf.
J.
Manuf.Mater.
Manuf. Mater.Process.
Process.2018,
Mater. Process.
2018,2,
2018, 2,
xFOR
FORPEER
xx FOR
PEERREVIEW
REVIEW
PEER REVIEW ofof15
99 of 1515

(a)
(a) (b)
(b)
(a) (b)
Figure6.6.20
Figure 20segments
segmentsofoftooltoolmoving
movingpath
pathsimulation
simulationresults.
results.(a)
(a)3D
3Ddiagram
diagramofof20 20segments
segmentscircular
circular
Figure
Figure 6.
6. 20
20 segments
segments of of tool
tool moving
moving path
path simulation
simulation results.
results. (a)
(a) 3D
3Ddiagram
diagramof of 20
20 segments
segments circular
circular
toolmoving
movingpath;
path;(b)
(b)Euclidean
Euclideannormnorm ee ofofposition
positionerror
errorofofeach
eachsegment.
segment.(cmd:
(cmd:tool
toolmoving
moving
tool
tool
tool moving
moving path;
path; (b)
(b) Euclidean norm ke
Euclidean norm of position
ek of position error
error ofof each
each segment.
segment. (cmd: tool
tool moving
moving
pathcommand;
path
path command;cmd_realpt:
command; cmd_realpt:simulated
cmd_realpt: simulatedtool
simulated toolmoving
tool movingpath
moving pathunder
path underthe
under theinfluence
the influenceof
influence ofvolumetric
of volumetricerror;
volumetric error;
error;
path command; cmd_realpt: simulated tool moving path under the influence of volumetric error;
final_comp:simulated
final_comp:
final_comp: simulatedtool
simulated toolmoving
tool movingpath
moving pathafter
path afterILC
after ILCcompensation).
ILC compensation).
compensation).
final_comp: simulated tool moving path after ILC compensation).

(a)
(a) (b)
(b)
(a) (b)
Figure7.7.30
Figure 30segments
segmentsofoftool toolmoving
movingpath
pathsimulation
simulationresults.
results.(a)
(a)3D3Ddiagram
diagramofof30 30segments
segmentscircular
circular
Figure
Figure 7.
7. 30
30 segments
segments of of tool
tool moving
moving path
path simulation
simulation results.
results. (a)
(a) 3D
3Ddiagram
diagramof of 30
30 segments
segments circular
circular
toolmoving
tool
tool
tool
movingpath;
moving
moving
path;(b)
path;
path;
(b)Euclidean
(b)
(b) Euclidean
Euclidean
norm ke
Euclideannorm
norm
norm eeek of
ofofposition
of
positionerror
position
position
errorof
error
error ofofeach
of
eachsegment.
each
each
segment.(cmd:
segment.
segment.
(cmd:tool
toolmoving
tool
(cmd: tool
moving
moving
moving
path command;
pathcommand;
path cmd_realpt:
command;cmd_realpt: simulated
cmd_realpt:simulated tool
simulatedtool moving
toolmoving
movingpathpath under
pathunder
underthe the influence
theinfluence
influenceofof volumetric
ofvolumetric error;
volumetricerror;
error;
path command; cmd_realpt: simulated tool moving path under the influence of volumetric error;
final_comp: simulated
final_comp:simulated
final_comp: tool
simulatedtool moving
toolmoving path
movingpath after
pathafter ILC
afterILC compensation).
ILCcompensation).
compensation).
final_comp: simulated tool moving path after ILC compensation).

(a)
(a) (b)
(b)
(a) (b)
Figure8.8.60
Figure
Figure 60segments
60 segmentsof
segments ofspiral
of spiraltool
spiral toolmoving
movingpath
moving pathsimulation
path simulationresults.
simulation results.(a)
results. (a)3D
(a) 3Ddiagram
3D diagramof
diagram of60
of 60segments
60 segments
segments
Figure 8. 60 segments of spiral tool moving path simulation results. (a) 3D diagram of 60 segments
spiraltool
spiral
spiral
spiral
toolmoving
tool
tool
moving
moving
moving
path;
path;
path;
path;
(b)
(b)(b) Euclidean
Euclidean
Euclidean
(b) Euclidean
norm norm
norm
norm eeeposition
kek of ofofposition
of
position
position oferror
eachofof
error error
error of
eachsegment.
segment.
each
each
segment. (cmd:
(cmd: tool
segment. (cmd:
tool
moving
(cmd: tool
tool
path command;
movingpath cmd_realpt:
pathcommand;
command;cmd_realpt:simulated tool
cmd_realpt:simulated moving
simulatedtool path
toolmoving under
movingpath the
pathunder influence
underthe of volumetric
theinfluence error;
influenceofofvolumetric
volumetric
moving
moving pathsimulated
command;toolcmd_realpt: simulated tool moving path under the influence of volumetric
final_comp:
error; final_comp: simulated moving
tool path
movingafter ILC
path compensation).
after ILC compensation).
error; final_comp:
error; final_comp: simulated
simulated tooltool moving
moving path
path after
after ILC
ILC compensation).
compensation).
J. Manuf. Mater. Process. 2018, 2, 53 10 of 15

J. Manuf. Mater. Process. 2018, 2, x FOR Table

PEER REVIEW
1. Comparison of simulation results. 10 of 15

Table 1. Comparison
Circular of simulation results.
Circular Spiral
Tool Moving Path
20 Segments 30 Segments 60 Segments
Circular Circular Spiral
Tool Moving
Euclidean norm Path
root-mean-square value of position errors of the 30
segmented tool moving path (µm)
20 Segments Segments 60 Segments
Euclidean norm
Volumetric error root-mean-square value of position errors of the segmented tool moving path (μm)
3.39 3.41 3.31
(before ILC compensation)
Volumetric error
3.39 3.41 3.31
(before
ILC ILC compensation)
compensation
ILC compensation 0.77 0.72 0.71
(after ILC compensation)
0.77 0.72 0.71
(after ILC compensation)
Rate of improvement (%) 77.28 78.89 78.55
Rate of improvement (%) 77.28 78.89 78.55

4.2. Discussion
4.2. Discussion ofof the
the Results
Results ofof Cutting
Cutting Tests
Tests
This study
This studyuses
usesthree-axis
three-axisCNC CNC milling
millingmachine
machine tooltool
slotslot
machining
machiningto detect the straightness
to detect error
the straightness
and height error of a workpiece to develop technical identification. The machining
error and height error of a workpiece to develop technical identification. The machining results are results are shown in
Figure 9inand
shown Table9 2.
Figure andTheTable
cutting tool cutting
2. The for experimentation is a ø10 mm end
tool for experimentation is a mill
ø10 with
mm endtwo flutes, andtwo
mill with the
workpiece material is aluminum alloy A6061. According to the data plot
flutes, and the workpiece material is aluminum alloy A6061. According to the data plot of the of the machining results, the
degree of error
machining convergence
results, the degree is of
known,
error and the relationship
convergence is known, between
and thetherelationship
table of volumetric
betweenerrors and
the table
the actual machining results can be observed. Figure 10 shows the side edge
of volumetric errors and the actual machining results can be observed. Figure 10 shows the side edgestraightness measurement
data distribution
straightness of three machined
measurement slots. Figure
data distribution of11 shows
three the height
machined measurement
slots. Figure 11data showsdistribution
the heightof
the bottom surface
measurement data ofdistribution
three machined of theslots. The experimental
bottom surface of three results show that
machined the The
slots. compensated tool
experimental
moving path has better machining results, the optimum rate of improvement
results show that the compensated tool moving path has better machining results, the optimum rate of the straightness error
is 79.03%,
of the minimum
improvement of the rate of improvement
straightness error is is79.03%,
62.57%,the andminimum
the average raterate
ofofimprovement
improvementisis62.57%,
68.87%.
The optimum rate of improvement of height error is 83.62%, the minimum
and the average rate of improvement is 68.87%. The optimum rate of improvement of height error is rate of improvement is
80.21%, and the average rate of improvement is 81.90%. In addition,
83.62%, the minimum rate of improvement is 80.21%, and the average rate of improvement is the workpiece measurement
result shows
81.90%. that the position
In addition, error ofmeasurement
the workpiece straight cutting increases
result shows gradually.
that the However,
position errorafter volumetric
of straight
error compensation, the distribution of position errors diminishes obviously,
cutting increases gradually. However, after volumetric error compensation, the distribution and fluctuates nearbyof
the zero-error value in the smaller error range. Therefore, the volumetric error
position errors diminishes obviously, and fluctuates nearby the zero-error value in the smaller error compensation method
designed
range. in this study
Therefore, can effectively
the volumetric errorreduce the volumetric
compensation method errors in theintool
designed thismoving
study canof aeffectively
three-axis
CNC milling machine tool.
reduce the volumetric errors in the tool moving of a three-axis CNC milling machine tool.

Figure 9. Machining results.

Figure 9. Machining results.
J. Manuf. Mater. Process. 2018, 2, 53 11 of 15

Table 2. Comparison of results of cutting tests.

Slot Number 1 2 3
Straightness error root-mean-square value comparison (unit: mm)
Without compensation 0.0179 0.0186 0.0160
With ILC compensation 0.0067 0.0039 0.0056
Rate of improvement (%) 62.57 79.03 65.00
Height error root-mean-square value comparison (unit: mm)
Without compensation 0.0177 0.0187 0.0193
With ILC compensation 0.0029 0.0037 0.0035
J. Manuf. Mater. Process. 2018, 2, x FOR PEER REVIEW 11 of 15
Rate of improvement (%) 83.62 80.21 81.87

0.01
0.005
0
-0.005 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Error (mm)

-0.01
No Comp
-0.015
Comp
-0.02
-0.025
-0.03
-0.035
-0.04
Measurement point

(a)
0.01
0.005
0
-0.005 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Error (mm)

-0.01 No Comp
-0.015 Comp

-0.02
-0.025
-0.03
-0.035
Measurement point

(b)
0.015
0.01
0.005
0
Error (mm)

-0.005 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
No Comp
-0.01 Comp

-0.015
-0.02
-0.025
-0.03
Measurement point

(c)

Figure 10.
Figure 10.Straightness
Straightness error
error measurement
measurement data distribution.
data distribution. (a) Slot 1 (a) Slot 1 error
straightness straightness error
measurement;
measurement;
(b) (b) Sloterror
Slot 2 straightness 2 straightness error (c)
measurement; measurement; (c) Sloterror
Slot 3 straightness 3 straightness error measurement.
measurement.

0.01

0.005

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
-0.005
Error (mm)

No Comp
-0.01
Comp
-0.015
 -0.03
Measurement point

(c)

Figure 10. Straightness error measurement data distribution. (a) Slot 1 straightness error
measurement;
J. Manuf. Mater. Process.(b) Slot2, 253straightness error measurement; (c) Slot 3 straightness error measurement. 12 of 15
2018,

0.01

0.005

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
-0.005

Error (mm)
No Comp
-0.01
Comp
-0.015

-0.02

-0.025

-0.03
Measurement point
J. Manuf. Mater. Process. 2018, 2, x FOR PEER REVIEW 12 of 15
(a)

0.005

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
-0.005

-0.01
Error (mm)

No Comp
-0.015
Comp
-0.02

-0.025

-0.03

-0.035
Measurement point

(b)
0.01
0.005
0
-0.005 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Error (mm)

-0.01 No Comp
-0.015 Comp

-0.02
-0.025
-0.03
-0.035
Measurement point

(c)

Figure 11. Height error measurement data distribution. (a) Slot 1 height error measurement; (b) Slot 2
Figure 11. Height error measurement data distribution. (a) Slot 1 height error measurement; (b) Slot 2
height error
height error measurement;
measurement; (c)
(c) Slot
Slot 33 height
height error
error measurement.
measurement.

Table 2. Comparison of results of cutting tests.

5. Conclusions
Slot Number 1 2 3
This study proposed the use of ILC to adjust the compensated tool path for the 21 volumetric
Straightness error root-mean-square value comparison (unit: mm)
errors in three-axis CNC milling machine tools, to increase the error compensation accuracy of the
Without compensation 0.0179 0.0186 0.0160
modified part program. The main conclusions of this study
With ILC compensation 0.0067
are as follows.0.0056
0.0039
• The volumetric error Rate of working
of the improvement (%) of the three-axis
volume 62.57 79.03 65.00 machine tool strongly
CNC milling
Heightdimension
influences the geometric error root-mean-square value
accuracy of the comparison
workpiece. (unit:
The mm)
volumetric error compensation
therefore becomes an Without compensation
issue that 0.0177 parts
high-precision mechanical 0.0187manufacturers
0.0193 must consider.
With ILC compensation 0.0029 0.0037 0.0035
• Existing part-program modification methods only consider the compensation of the position of a
Rate of improvement (%) 83.62 80.21 81.87
single block of a part program but neglect the movement bias induced by volumetric error in the
tool moving process, which leads to limited volumetric error compensation effect.
5. Conclusions
• In this study, the position of a single block of a part program was segmented, and volumetric error
This study proposed
compensation the use of ILC
was implemented to adjust
for the the compensated
tool moving tool path
position in each for the
segment 21 volumetric
to generate a part
errors in three-axis CNC milling machine tools, to increase the error compensation accuracy of the
modified part program. The main conclusions of this study are as follows.
● The volumetric error of the working volume of the three-axis CNC milling machine tool
strongly influences the geometric dimension accuracy of the workpiece. The volumetric error
compensation therefore becomes an issue that high-precision mechanical parts manufacturers
J. Manuf. Mater. Process. 2018, 2, 53 13 of 15

program with multiple compensated blocks. The proposed compensation method can effectively
compensate for volumetric errors in the tool moving process.
• In addition, this study used the ILC concept and referred to the present compensated tool path
and volumetric error induced by the compensated tool path to calculate the next compensated tool
path. Therefore, the optimum compensated tool path can be obtained by the iterative computation
of the compensated tool path.
• To validate the effect of the proposed segmented volumetric error compensation, the measured
and established table of volumetric errors was used for path simulation and cutting tests.
Both the simulation and experimental results prove the feasibility of using the ILC method
for the segmented modification of the volumetric error-compensated tool path.
• Several cutting tests were performed on a three-axis CNC milling machine tool. The experimental
results showed that the average rate of improvement of the straightness error of the proposed
volumetric error compensation method is higher than 60% while the rate of improvement of the
height error is higher than 80%.

Author Contributions: Investigation, Y.-C.L.; Investigation, Supervision, S.-S.Y.

Funding: This research was funded in part by the Ministry of Science and Technology, Taiwan, under Contract
MOST104-2221-E-027-132 and MOST103-2218-E-009-027-MY2.
Acknowledgments: The authors would like to thank representatives from the 3DFAMILY Technology for their
beneficial discussions with the project team. The authors would also like to thank Chen-Chou Chung (Ford Lio Ho
Motor Company) for his valuable comments and suggestions on volumetric error compensation of CNC milling
machine tools.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Zou, X.; Zhao, X.; Li, G.; Li, Z.; Sun, T. Sensitivity analysis using a variance-based method for a three-axis
diamond turning machine. Int. J. Adv. Manuf. Technol. 2017, 92, 4429–4443. [CrossRef]
2. Wang, Z.; Maropoulos, P.G. Real-time laser tracker compensation of a 3-axis positioning system—Dynamic
accuracy characterization. Int. J. Adv. Manuf. Technol. 2016, 84, 1413–1420. [CrossRef]
3. Uekita, M.; Takaya, Y. On-machine dimensional measurement of large parts by compensating for volumetric
errors of machine tools. Precis. Eng. 2016, 43, 200–210. [CrossRef]
4. Wang, K.; Sheng, X.; Kang, R. Volumetric error modelling, measurement, and compensation for an integrated
measurement-processing machine tool. Proc. Inst. Mech. Eng. Part C J. Mech. Eng. Sci. 2010, 224, 2477–2486.
[CrossRef]
5. Fletcher, S.; Postlethwaite, S.R.; Ford, D.G. Volumetric compensation through the machine controller.
Laser Metrol. Mach. Perform. V 2001, 34, 321–330.
6. Li, D.; Feng, P.; Zhang, J.; Yu, D.; Wu, Z. An identification method for key geometric errors of machine tool
based on matrix differential and experimental test. Proc. Inst. Mech. Eng. Part C J. Mech. Eng. Sci. 2014, 228,
3141–3155. [CrossRef]
7. Okafor, A.C.; Ertekin, Y.M. Derivation of machine tool error models and error compensation procedure
for three axes vertical machining center using rigid body kinematics. Int. J. Mach. Tools Manuf. 2000, 40,
1199–1213. [CrossRef]
8. Chen, J.S.; Kou, T.W.; Chiou, S.H. Geometric error calibration of multi-axis machines using an auto-alignment
laser interferometer. Precis. Eng. 1999, 23, 243–252. [CrossRef]
9. Knapp, W.; Matthias, E. Test of the three-dimensional uncertainty of machine tools and measuring machines
and its relation to the machine errors. CIRP Ann. Manuf. Technol. 1983, 32, 459–464. [CrossRef]
10. Ni, J.; Wu, S.M. On-line measurement technique for machine volumetric error compensation. J. Eng. Ind.
Trans. ASME 1993, 115, 85–92. [CrossRef]
11. Ibaraki, S.; Knapp, W. Indirect measurement of volumetric accuracy for three-axis and five-axis machine
tools: A review. Int. J. Autom. Technol. 2012, 6, 110–124. [CrossRef]
J. Manuf. Mater. Process. 2018, 2, 53 14 of 15

12. Majda, P. Modeling of geometric errors of linear guideway and their influence on joint kinematic error in
machine tools. Precis. Eng. 2012, 36, 369–378. [CrossRef]
13. He, Z.; Fu, J.; Yao, X.; Qian, W. Volumetric error identification for CNC machine tool based on multi-body
system and vector diagonal measurement. In Proceedings of the 2010 International Conference on Mechanic
Automation and Control Engineering, Wuhan, China, 26–28 June 2010; pp. 3345–3349.
14. Jung, J.H.; Choi, J.P.; Lee, S.J. Machining accuracy enhancement by compensating for volumetric errors of a
machine tool and on-machine measurement. J. Mater. Process. Technol. 2006, 174, 56–66. [CrossRef]
15. Ahn, K.G.; Min, B.K.; Pasek, Z.J. Modeling and compensation of geometric errors in simultaneous cutting
using a multi-spindle machine tool. Int. J. Adv. Manuf. Technol. 2006, 29, 929–939. [CrossRef]
16. Yang, S.H.; Kim, K.H.; Park, Y.K.; Lee, S.G. Error analysis and compensation for the volumetric errors of
a vertical machining centre using a hemispherical helix ball bar test. Int. J. Adv. Manuf. Technol. 2004, 23,
495–500. [CrossRef]
17. Xiang, S.; Yang, J.; Fan, K.; Lu, H. Multi-machine tools volumetric error generalized modeling and
ethernetbased compensation technique. Proc. Inst. Mech. Eng. Part B J. Eng. Manuf. 2016, 230, 870–882.
[CrossRef]
18. Holub, M.; Blecha, P.; Bradac, F.; Kana, R. Volumetric compensation of three-axis vertical machining centre.
MM Sci. J. 2015, 2015, 677–681. [CrossRef]
19. Holub, M.; Vetiska, J.; Bradac, F.; Vala, M. Application on-the-fly measurement of CNC machine tools.
MM Sci. J. 2017, 2017, 2085–2089. [CrossRef]
20. Gomez-Acedo, E.; Olarra, A.; Orive, J.; Lopez De La Calle, L.N. Methodology for the design of a thermal
distortion compensation for large machine tools based in state-space representation with Kalman filter. Int. J.
Mach. Tools Manuf. 2013, 75, 100–108. [CrossRef]
21. Gomez-Acedo, E.; Olarra, A.; Zubieta, M.; Kortaberria, G.; Ariznabarreta, E.; López de Lacalle, L.N. Method
for measuring thermal distortion in large machine tools by means of laser multilateration. Int. J. Adv.
Manuf. Technol. 2015, 80, 523–534. [CrossRef]
22. López De Lacalle, L.N.; Lamikiz, A.; Ocerin, O.; Díez, D.; Maidagan, E. The Denavit and Hartenberg approach
applied to evaluate the consequences in the tool tip position of geometrical errors in five-axis milling centres.
Int. J. Adv. Manuf. Technol. 2008, 37, 122–139. [CrossRef]
23. Díaz-Tena, E.; Ugalde, U.; López De Lacalle, L.N.; De La Iglesia, A.; Calleja, A.; Campa, F.J. Propagation of
assembly errors in multitasking machines by the homogenous matrix method. Int. J. Adv. Manuf. Technol.
2013, 68, 149–164. [CrossRef]
24. Shen, H.; Fu, J.; He, Y.; Yao, X. On-line asynchronous compensation methods for static/quasi-static error
implemented on CNC machine tools. Int. J. Mach. Tools Manuf. 2012, 60, 14–26. [CrossRef]
25. Wang, J.; Guo, J. Research on volumetric error compensation for NC machine tool based on laser tracker
measurement. Sci. China Technol. Sci. 2012, 55, 3000–3009. [CrossRef]
26. Vahebi Nojedeh, M.; Habibi, M.; Arezoo, B. Tool path accuracy enhancement through geometrical error
compensation. Int. J. Mach. Tools Manuf. 2011, 51, 439–449. [CrossRef]
27. Wang, S.M.; Lin, J.J. On-machine volumetric-error measurement and compensation methods for micro
machine tools. Int. J. Precis. Eng. Manuf. 2013, 14, 989–994. [CrossRef]
28. Eskandari, S.; Arezoo, B.; Abdullah, A. Positional, geometrical, and thermal errors compensation by tool path
modification using three methods of regression, neural networks, and fuzzy logic. Int. J. Adv. Manuf. Technol.
2013, 65, 1635–1649. [CrossRef]
29. Sun, S.H.; Lee, E.S.; Sohn, J.W. Enhancement of geometric accuracy via an intermediate geometrical feedback
scheme. J. Manuf. Syst. 1999, 18, 12–21.
30. Lee, E.S.; Suh, S.H.; Shon, J.W. A comprehensive method for calibration of volumetric positioning accuracy
of CNC-machines. Int. J. Adv. Manuf. Technol. 1998, 14, 43–49. [CrossRef]
31. Kiridena, V.S.B.; Ferreira, P.M. Computational approaches to compensating quasistatic errors of three-axis
machining centers. Int. J. Mach. Tools Manuf. 1994, 34, 127–145. [CrossRef]
32. Zhu, W.; Wang, Z.; Yamazaki, K. Machine tool component error extraction and error compensation by
incorporating statistical analysis. Int. J. Mach. Tools Manuf. 2010, 50, 798–806. [CrossRef]
33. Lee, J.H.; Liu, Y.; Yang, S.H. Accuracy improvement of miniaturized machine tool: Geometric error modeling
and compensation. Int. J. Mach. Tools Manuf. 2006, 46, 1508–1516. [CrossRef]
J. Manuf. Mater. Process. 2018, 2, 53 15 of 15

34. Wang, C. Laser vector measurement technique for the determination and compensation of volumetric
positioning errors. Part I: Basic theory. Rev. Sci. Instrum. 2000, 71, 3933–3937. [CrossRef]
35. Janeczko, J.; Griffin, B.; Wang, C. Laser vector measurement technique for the determination and
compensation of volumetric position errors. Part II: Experimental verification. Rev. Sci. Instrum. 2000, 71,
3938–3941. [CrossRef]
36. Fanuc. Fanuc User’s Manual: Common to Lathe System/Machining Center System; Fanuc Corporation: Yamanashi
Prefecture, Japan, 2004; Volume B-63944EN/02.
37. Spong, M.K.; Hutchinson, S.; Vidyasagar, M. Robot Modeling and Control; John Wiley & Sons: Hoboken, NJ,
USA, 2006.
38. Uchiyama, M. Formation of high speed motion pattern of mechanical arm by trial. Trans. Soc. Instrum.
Control Eng. 1978, 14, 706–712. [CrossRef]
39. Kim, D.-I.; Kim, S. On iterative learning control algorithm for industrial robots and CNC machine tools.
In Proceedings of the IECON’93 19th Annual Conference of IEEE Industrial Electronics, Maui, HI, USA,
15–19 November 1993; pp. 601–606.
40. Kim, D.I.; Kim, S. An iterative learning control method with application for CNC machine tools. IEEE Trans.
Ind. Appl. 1996, 32, 66–72. [CrossRef]
41. Arimoto, S.; Kawamura, S.; Miyazaki, F. Bettering operation of robots by learning. J. Robot. Syst. 1984, 1,
123–140. [CrossRef]
42. Arimoto, S.; Naniwa, T.; Suzuki, H. Robustness of P-type learning control with a forgetting factor for robotic
motions. In Proceedings of the IEEE Conference on Decision and Control, Honolulu, HI, USA, 5–7 December
1990; pp. 2640–2645.
43. Bien, Z.; Huh, K.M. Higher-order iterative learning control algorithm. IEEE Proc. D Control Theory Appl.
1989, 136, 105–112. [CrossRef]

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).