HYBRID APPROACH FOR PREDICTIVE MAINTENANCE USING DIGITAL TWIN

Prof. Priya Jadhav Sneh Sonam Utkarsh Nath

Dept. Of Mechanical Engineering Dept. Of Mechanical Engineering Dept. Of Mechanical Engineering
Symbiosis Institute of Technology Symbiosis Institute of Technology Symbiosis Institute of Technology
Pune, India Pune, India Pune, India
[email protected] [email protected] [email protected]

Dr. Satish Kumar Sheron Sisodiya Mehr Billawaria

Dept. Of Mechanical Engineering Dep. of Mechanical Engineering Dept. Of Mechanical Engineering
Symbiosis Institute of Technology Symbiosis Institute of Technology Symbiosis Institute of Technology
Pune, India Pune, India Pune, India
[email protected] [email protected] [email protected]

computer simulation that can be analyzed, tested as well as

Abstract - This paper discusses the simulation and presented like a real machine is called the virtual prototype.
virtual modelling of the feed drive system of a complex of the machine tool. The product’s development cost as well
CNC machine tool. The first part of the project deals time is significantly decreased by making changes and
with the modelling of a structure of the feed drive which optimizations in a virtual model of the machine tool in
is very complex and has several parameters such as iterations as well as implementing design modifications,
position, speed, inertia, control loops, etc. All these throughout the design process, till the performance
parameters have a significant impact on the machine requirements have been achieved.
tool such as movement stability, positioning accuracy CNC machine tool’s performance and reliability is
and feed drive sensitivity. For the modelling of the determined by the quality of products it will process during
system the Matlab SIMULINK software was used. its working. Feed drives are the most significant
Simulations, tests and validations were done to components of each machine tool. The feed drives are very
understand the effects of varying and optimizing complex electro-mechanical systems for CNC machine
different feed drive parameters on the dynamic stiffness tools. The nonlinearity in the process along with the noise is
and positioning accuracy. what causes degradation in the performance of the
Computer Numerical Control machine tool. As a result, an
Key words: virtual modelling, simulation, CNCMT, DTMT error occurs in the prognosis.
In the life cycle of industrial products, different forms
ⅠIntroduction of engineering applications and digitalized machinery are
widely adopted. As a result, immense data of different sorts
Since the mid-1990s a change in perspective from is likewise being delivered. These data, however, are
'physical' to 'virtual' creation has brought about an increased hysteretic and separated from one another, resulting in poor
interest in research in the field. With the continuous productivity and limited use of such useful information. In
development in the field of computer technology and its early planning stage, simulation related to the theoretical
software developments, it has become very easy to simulate and static models was a conventional and effective method
a physical manufacturing system. The main purpose of a for verifying, validating, and integrating a system, but
virtual simulation is to understand and recreate or reproduce during system run-time little emphasis has been placed on
similar characteristics of the physical manufacturing system the simulation application. More data will be obtained with
on a computer system. This reduces the number of tests the advancement of new-generation intelligence and
required and the number of experiments conducted digitalization technology, and it is an opportunity to figure a
physically in the workshop. A virtual system results in low way to implement all this data in detail. As a response, the
material wastage and errors in operations of the machine. idea of the digital twin has generated a great deal of concern
These techniques are vital in reducing the cost of the and is emerging very fast.
product which are manufactured. Hybrid predictive maintenance can be implemented
Currently the process of testing as well as manufacturing effectively with the help of DT concept CPS are enabled by
of physical prototypes, in order to identify spots that are Digital Twins. Digital Twin allows the CPS method to be
weak and also optimize the design, are very cost intensive executed. It uses laws of physics and thus results in having
and time consuming. This is something that cannot be an extremely accurate digital model. This model can then
afforded by machine tool builders anymore. Instead, to cache historical data used while running to be used further.
decrease the cost and time taken to test the hardware and It can also procure and sense data in real time one the
make iterative optimizations to the physical prototype, equipment is utilized. To understand the hybrid predictive
modern design processes of machine tools utilize “virtual maintenance strategy however, it is important to develop
prototyping” technology. A physical item or product’s precise physical and mathematical models, obtain accurate
running data in real time, and render useful data.
 Ⅱ. Literature Review

Production time studies in smaller businesses show that

each operation's machining and assembly activities
constitute just 5% of the entire length of the operation. Most
of the working time is often spent excessively on transport
and storage Nyhuis, P, 2001. A change in view to 'virtual’
from 'real output, as a result has led to an expansion in the
field of research since the early 1990s. Less material is lost
using a computer device and interruptions can be avoided in
the operation of an actual machine on the shop floor
Radharamanan R, 2002. Figure 1. Current predictive maintenance methods.
The target of current assembling advances is to deliver
the initial segment accurately and in the savviest path in the C. Digital Twin
briefest time of time. The development and testing of actual
models are significant bottlenecks for the effective and With the evolution of production towards intelligent
financially profitable creation of present-day machine 7 mode, automatic technology, convergence, the concern has
instruments, as the product complexities increment and the increased about DT along with its growth, ref Fig. 2.
serious item life process durations are diminished Weck, M Initially, the idea of "twin" was initially derived from the
et al., 2003. Apollo Project of NASA, when the aircraft was ``A real
The making of present-day programming tools that are physical system is a twin body Hochhalter J et al., 2014.
used for simulation of properties of the product have been Nowadays a twin body model can assist astronauts and
significantly improved lately, initiated primarily by the personnel make decisions in the context of circumstances of
automotive and aircraft industries Weck M. et al., 2003. emergency. Further, for this Digital Twin idea, a product
Before the first physical prototype is constructed, with the lifecycle is proposed for the formation of the 2002
help of software that are advanced and hardware systems Management Center Rosen R et al., 2015. Enabling the IoT
engineers test and optimize critical characteristics of the to help new intelligent services, the Digital Twin model
product with virtual prototypes. For the various design connects and interacts with a physical object Wang, P and
phases of a machine tool, a large range of software tools are Zuehlke D., 2015. These capabilities permit makers to
available Weule, H et al., 2003. perceive and drive new plans of action, including real-time
simulation and simulation issue forecast Tao F et al., 2014.
A. CNCMT (CNC Machine Tool)

As the industry's mother computer, the CNC machine

tool (CNCMT) is one type of significant production
equipment. The achievement of CNCMT's consistency and
reliability determines the quality of the goods it has stored.
CNCMT, however can unpredictably incur defects, which
low accuracy, stagnation of output or even significant loss
and loss can be caused. Casualties if there is no prompt Figure 2. Evolution of Digital Twin technology
troubleshooting. Therefore, to prevent any fault and to make
these faults more reliable in CNCMT, we require predictive Different facets of technology (cyber-physical
maintenance. T. Kjellberg et al., 2009. structures, IoT/IoS, AI, machine utilization (energy
As in today’s generation technologies such as Arduino productivity) and operations (cloud development, multiple
based sensing, monitoring, simulation in MATLAB and intelligence interaction & virtual- physical interlinkage)
ANSYS, big data and AI are being developed. A lot of constitute the characteristics of the digital twin model. In
savvy strategies have been applied to manufacturing, which today's world of machines with computer and physical
give a proficient method for prescient support fueled by the facilities, cyber-physical systems will create a strong
Data from manufacturing J. Wang et al., 2019. linkage between people & the world of digital information
Baheti R et al., 2011. The Internet of Things (IoT/IoS)
B. Current Predictive Maintenance Methods utilizes different sensors, (for example, radio recurrence
Currently, there are 3 strategies that are mostly utilized acknowledgment, infrared acceptance and worldwide
for predictive maintenance of complex hardware, in situating, and so forth to deal with huge volumes of
particular the strategy for unwavering quality information of various kinds to understand the elements of
measurements, the technique dependent on actual models danger ID, shortcoming area and wellbeing the board Kelly
and the strategy dependent on information/data, refer Fig. S et al. From a security point of view, there are several
1.all the methods have their own features and fields that resources (monitoring of the process, prognosis as well as
apply Mobley R K, 2002. the diagnosis of the faults, management of the alarms,
Shorten MTBF) in the latest put forward model of DT of the mechanical framework, etc., as introduced in Fig. The CNC
tool system. deciphers 17 G-codes, increasing speed and retardation
plans, and afterward dispatches the servo position orders &
axle speed orders to the servo and axle drive frameworks,
separately, at various testing times. To give better situating
exactness of the machine table, the servo drive framework
for the most part embraces a full-shut circle position control
by utilizing a direct scale to input the current situation of the
table. In the current segment, the model of the end-
processing cutting power also the model of servo drive is
introduced independently and afterward consolidated as an
incorporated framework model Chen-Jung L et al., (2019)

Ⅲ. Experimentation

A. Modelling and simulation of virtual CNC machine feed

drive system with MATLAB/Simulink

Feed drives are the most significant components of each

machine tool. The feed drives are very complex electro-
mechanical systems for CNC machine tools. A good control
Figure 3. Characteristics, service, advantage and enabling dynamic, capability of the acceleration, precision of the
factors defining tool system digital twin model. position, the load stiffness and velocity of the process must
be displayed. The various mistakes of each feed drive pivot
D. Feed Drives for machine devices should be explored to achieve high
Feed drives are the most significant components of each precision machining. A machine instrument comprises servo
machine tool. The feed drives are very complex electro- engines, spindles, structures, etc., and a complex interaction
mechanical systems for CNC machine tools. In particular a of a lot of characteristics like servo dynamics, friction,
good control dynamic, capability of the acceleration, structural vibration, the thermal effects, & the cutting
precision of the position, the load stiffness and velocity of dynamics is its dynamic response.
the process must be displayed Gross H (2006) and Pandilov Typically, an actuator (rotary or linear), elements for
Z (2014). mechanical transmission and a cascade control of three
It is conceivable the traditional feed drive has simplistic loops (PI current control loop, PI velocity control loop, P
linear dynamics, ref Fig.4 Here, the control signal u '(V) is position control loop) collectively constitute the feed drive
applied to the current amplifier input, denoted by the Ka system. The machine segments that stand in the power
(A/V) gain. Amplifier then induces the current (A) in the transmission flow (torque) between the workpiece/tool and
armature of the motor, which gives us the output in the Tm the actuator make up the elements of mechanical
(Nm) motor torque, that is linearly proportional to I wherein transmission. The given mechanical transmission elements
the Kt (Nm/A) motor torque constant. The engine shaft is are used most often in different design variants: clutches,
additionally exposed to a disturbance force Td (Nm) rack and pinion units, gearboxes, gears, ball lead screw and
notwithstanding the current produced force going through nut units, bearings.
armature, which remembers the impact of grating for the
guideways, ball screws & bearings, just like the cutting
forces Y. Altintas et al., 2005.

Figure 4 Linear dynamics of a feed drive

E. Integrated models of servo feed drive and cutting force

Milling is a method of cutting that separates material

from a workpiece using a milling cutter. The design of a
CNC machine instrument, which is electromechanical,
incorporates a CNC regulator, a servo drive framework,
 .
Figure 7. SIMULINK block representation of CAD model

The next step in modelling is to create a Feed Drive

System for the X and Y axis. Every axis requires a separate
feed drive system. We create a SIMULINK one mass model
for all the axes to check the influence of changing several
parameters related to the drive system and the compliance
of the drive system in the frequency domain.

Figure 5. Flowchart of Modelling and Simulation

The first step is to create a virtual three-dimensional

CAD model of one CNC portal milling machine. The model
has a direct lead screw feed drive for each axis. A vertical
milling machine is designed in Solidwork. The PX 10 series
of vertical machining centers has been designed, it offers Figure 8. Simulink one mass model for Y-axis feed drive
high cutting rigidity and low vibration for heavy cutting model
loads and provides for an efficient chip flow. The tool
progresses perpendicularly to the workpiece on a dedicated
axis to conduct multiple special, tiny cuts in the shape of
chips to extract material from a surface.

Figure 9. Simulink one mass model for X-axis feed drive

model

The model is designed to show the entire princess of

milling carried out along the X and Y axis using
mathematical relations. For this parameter such as specific
cutting force, load inertia, Laplace operator/sampling
period, etc. are considered.
Nominal values of the parameters of the Y-axis feed
Figure 6. CAD model of vertical milling machine drive of the virtual milling machine are given as follows:
Kv=80 - position loop gain [1/s], Te=0.0007 - time constant
On completion of the virtual model in Solidworks, the of the current loop [s], Kp=0.017 - proportional gain of the
solid model is imported into Simulink. Simscape Multibody velocity controller [As/rad], Tn=0.0028 -integral time of the
or SimMechanics enables us to work with three dimensional velocity controller [s], Ts=0.002 - sampling period [s], Km=
mechanical systems, like robots, suspensions of the vehicles 1.63 - torque constant of the rotary motor [Nm/A], m=60 -
by providing a multibody simulation environment. It models mass of the moving elements [kg], J=0.001966 - Load
multibody systems using blocks which represent bodies, inertia [Kg/m2],Kg= 1256 - constant for the transformation
joints, constraints, force elements, and sensors. of linear to rotary movement [rad/m]. For the X asis model
minor changes in values are observed: J=0.019606 - Load
inertia [Kg/m2]
Measure of influence of the disturbance force F (torque
T) on the output position Xo (angular position) deviation in
the transient period is called the dynamic stiffness of the
feed drive system. The compliance of the feed drive system
to the disturbance forces is indicated by the results.
Sensitivity to load disturbances is one of the most crucial
requirements about the feed drive system concerns. Feed
drive stiffness is the qualitative measure of the given
sensitivity. The dynamic stiffness or the feed drive Figure 13. Influence of position gain Kv and proportional
sensitivity is a reciprocal value of compliance. gain Kp on the compliance of the velocity loop

The graphs of position loop (Figure 7) and velocity loop

(Figure 8) shows the effect of Kv and Kp on the compliance
of the closed loops. The compliance decreased over a period
of time. This indicates that the feed drive sensitivity to
disturbance force, or the dynamic stiffness increases with
time.
. B. simulation and multibody modelling of virtual CNC
machine feed drive system with Matlab SimScape Toolbox

Figure 10. Magnitude of actual feed rate v/s time for feed
drive system

Figure 14. Multibody Simscape model of Y-axis feed drive

of virtual CNC portal milling machine

Figure 11. Magnitude of actual cutting force vs time for

feed drive system

The actual cutting force value should not exceed the

specific cutting force value that is allowed. The Specific
Cutting force for the model is Kc=1760 N/mm2 for the X
and Y axis.

Figure 15. Simscape model of mechanical elements of Y-

axis feed drive of virtual CNC portal machine tool
(subsystem)

Ⅳ. Future Works
The next stage of the study was to use the feed drive
models created in Matlab Simulink and Simscape in the VR
system generated with the EON Studio software along with
Figure 12. Influence of position gain Kv and proportional the machine tool CAD models in order to visualize the
gain Kp on the compliance of the position loop entire machine tool. The quality of visualization was
remarkably enhanced due to the EON studio software. We
can also integrate the Z- Axis feed drive system in the
simulink model.
 [14] Li, C. J., Tseng, H. C., Tsai, M. S., & Cheng, C. C. (2020). Novel
ⅴ. Conclusion servo-feed-drive model considering cutting force and structural effects in
milling to predict servo dynamic behaviors. The International Journal of
Advanced Manufacturing Technology, 106(3), 1441-1451.
The paper proposes a method of maintenance where a [15] Wang, J., Yan, J., Li, C., Gao, R. X., & Zhao, R. (2019). Deep
DT drives a hybrid predictive maintenance method for a heterogeneous GRU model for predictive analytics in smart manufacturing:
Computer Numerical Control Machine Tool (CNCMT). The application to tool wear prediction. Computers in Industry, 111, 1-14.
paper uses an approach to create a hybrid algorithm using [16] Duan, L., Xie, M., Bai, T., & Wang, J. (2016, June). Support vector
software such as MATLAB and SIMULINK for fault data description
for machinery multi-fault classification with unbalanced datasets In 2016
detection and maintenance. In the given method, a multi- IEEE
domain DT model is created using physical, electrical, International Conference on Prognostics and Health Management (ICPHM)
hydraulic, and thermodynamic information to reflect the (pp.
working conditions in their actual form. When compared 1-6). IEEE.
with the predictions of the traditional model, we found [17] Hanachi, H., Yu, W., Kim, I. Y., Liu, J., & Mechefske, C. K. (2019).
Hybrid data driven physics-based model fusion framework for tool wear
through experimental outcomes that the accuracy of the prediction. The International Journal of Advanced Manufacturing
tracking error estimates was higher when the transmission Technology, 101(9-12), 2861- 2872.
path was not taken into account. [18] Liao, L., & Köttig, F. (2016). A hybrid framework combining data-
Engineering evaluation systems and virtual modelling driven and model-based methods for system remaining useful life
can assist product or machine tool designers and prediction. Applied Soft Computing, 44, 191-199.
manufacturing engineers in improving the precision and [19] Mobley, R. K. (2002). An introduction to predictive maintenance.
Elsevier.
reducing the lead time of the machining process. This entire [20] Shimada, J., & Sakajo, S. (2016, July). A statistical approach to reduce
study could change the current expensive and time- failure facilities based on predictive maintenance. In 2016 International
consuming processes of making real, physical prototypes of Joint Conference on Neural Networks (IJCNN) (pp. 5156-5160). IEEE.
CNC MTs to developing virtual prototypes of them; Thus, [21] Li, C., Zhang, Y., & Xu, M. (2012). Reliability-based maintenance
proving to be very useful to machine tool producers. optimization under imperfect predictive maintenance. Chinese Journal of
Mechanical Engineering, 25(1), 160-165.
[22] Ortiz, J., & Carrasco, R. A. (2016, July). Model-based fault detection
References and diagnosis in ALMA subsystems. In Observatory Operations:
Strategies, Processes, and Systems VI (Vol. 9910, p. 99102S). International
[1] Nyhuis, P., & Wiendahl, H. P. (2008). Fundamentals of production Society for Optics and Photonics.
logistics: theory, tools and applications. Springer Science & Business [23] Lei, Y., Li, N., Gontarz, S., Lin, J., Radkowski, S., & Dybala, J.
Media (2016). A model-based method for remaining useful life prediction of
[2] Radharamanan, R. (2002). Virtual manufacturing: an emerging machinery. IEEE Transactions on Reliability, 65(3), 1314-1326.
technology. age, 7, 1. [24] Iung, B., Monnin, M., Voisin, A., Cocheteux, P., & Levrat, E. (2008).
[3] Onosato, M., & Iwata, K. (1993). Development of a virtual Degradation state model-based prognosis for proactively maintaining
manufacturing system by integrating product models and factory models. product performance. CIRP annals, 57(1), 49-52.
CIRP annals, 42(1), 475-478 [25] Baptista, M., Sankararaman, S., de Medeiros, I. P., Nascimento Jr, C.,
[4] Mwinuka, T. E., & Mgwatu, M. I. (2015). Tool selection for rough and Prendinger, H., & Henriques, E. M. (2018). Forecasting fault events for
finish CNC milling operations based on tool-path generation and predictive maintenance using data-driven techniques and ARMA modeling.
machining optimisation. Advances in Production Engineering & Computers & Industrial Engineering, 115, 41-53.
Management, 10(1). [26] Daily, J., & Peterson, J. (2017). Predictive maintenance: How big data
[5] Debevec, M., Simic, M., & Herakovic, N. (2014). Virtual factory as an analysis can improve maintenance. In Supply Chain Integration Challenges
advanced approach for production process optimization. International in Commercial Aerospace (pp. 267-278). Springer, Cham.
journal of simulation modelling, 13(1), 66-78. [27] Canizo, M., Onieva, E., Conde, A., Charramendieta, S., & Trujillo, S.
[6] Cus, F., & Zuperl, U. (2015). Surface roughness control simulation of (2017, June). Real-time predictive maintenance for wind turbines using Big
turning processes/Simulator sistema vodenja hrapavosti povrsine pri Data frameworks. In 2017 IEEE International Conference on Prognostics
procesu struzenja. Strojniski vestnik-Journal of Mechanical Engineering, and Health Management (ICPHM) (pp. 70-77). IEEE.
61(4), 245-255. [28] Wang, J., Fu, P., Zhang, L., Gao, R. X., & Zhao, R. (2019). Multilevel
[7] Hrelja, M., Klancnik, S., Irgolic, T., Paulic, M., Jurkovic, Z., Balic, J., information fusion for induction motor fault diagnosis. IEEE/ASME
& Brezocnik, M. (2014). Particle swarm optimization approach for Transactions on Mechatronics, 24(5), 2139-2150.
modelling a turning process. Advances in Production Engineering & [29] Wang, J., Yan, J., Li, C., Gao, R. X., & Zhao, R. (2019). Deep
Management, 9(1), 21. heterogeneous GRU model for predictive analytics in smart manufacturing:
[8]Todorović, P., Vukelic, D., Tadić, B., Veljkovic, D., Budak, I., Macuzić, application to tool wear prediction. Computers in Industry, 111, 1-14.
I., & Lalic, B. (2014). Modelling of dynamic compliance of [30] Wang, J., Zheng, Y., Wang, P., & Gao, R. X. (2017). A virtual sensing
fixture/workpiece interface. International Journal of Simulation Modelling. based augmented particle filter for tool condition prognosis. Journal of
[9] Weck, M., Queins, M., & Witt, S. (2003). Effektive Entwicklung von Manufacturing Processes, 28, 472-478.
Werkzeugmaschinen. VDI-Z, 145(10), 32-36. [31] J. Wang, J. Xie, R. Zhao, et al., Multisensory fusion based virtual tool
[10] Altintas, Y., Brecher, C., Weck, M., & Witt, S. (2005). Virtual wear sensing for ubiquitous manufacturing, Robot. Comput. Integr. Manuf.
machine tool. CIRP annals, 54(2), 115-138. 45 (2017) 47–58.
[11] Weule, H., Fleischer, J., Neithardt, W., Emmrich, D., & Just, D. [32] Liao, L., & Köttig, F. (2016). A hybrid framework combining data-
(2003, June). Structural Optimization of Machine Tools including the static driven and model-based methods for system remaining useful life
and dynamic Workspace Behavior. In The 36th cirp-international seminar prediction. Applied Soft Computing, 44, 191-199.
on manufacturing systems (Vol. 12, pp. 56-60). [33] Wang, J., Wang, P., & Gao, R. X. (2015). Enhanced particle filter for
[12] Brecher, C., Esser, M., & Witt, S. (2009). Interaction of tool wear prediction. Journal of Manufacturing Systems, 36, 35-45.
manufacturing process and machine tool. CIRP annals, 58(2), 588-607. [34] Wang, J., Liang, Y., Zheng, Y., Gao, R. X., & Zhang, F. (2020). An
[13] Kjellberg, T., von Euler-Chelpin, A., Hedlind, M., Lundgren, M., integrated fault diagnosis and prognosis approach for predictive
Sivard, G., & Chen, D. (2009). The machine tool model—A core part of maintenance of wind turbine bearing with limited samples. Renewable
the digital factory. CIRP annals, 58(1), 425-428. Energy, 145, 642-650.
[35] Hochhalter, J., Leser, W. P., Newman, J. A., Gupta, V. K., Yamakov, [58] Ruijun, L., Wenhua, Y., Zhang, H. H., & Qifan, Y. (2012). The
V., Cornell, S. R., ... & Heber, G. (2014). Coupling Damage-Sensing thermal error optimization models for CNC machine tools. The
Particles to the Digital Twin Concept. International Journal of Advanced Manufacturing Technology, 63(9-12),
[36] Rosen, R., Von Wichert, G., Lo, G., & Bettenhausen, K. D. (2015). 1167-1176.
About the importance of autonomy and digital twins for the future of [59] Muhammad, B. B., Wan, M., Feng, J., & Zhang, W. H. (2017).
manufacturing. IFACPapersOnLine, 48(3), 567-572. Dynamic damping of machining vibration: a review. The International
[37] Wang, P., Gao, R. X., & Fan, Z. (2015). Cloud computing for cloud Journal of Advanced Manufacturing Technology, 89(9-12), 2935-2952.
manufacturing: benefits and limitations. Journal of Manufacturing Science [60] Altintas, Y., & Ber, A. A. (2001). Manufacturing automation: metal
and Engineering, 137(4). cutting mechanics, machine tool vibrations, and CNC design. Appl. Mech.
[38] Zuehlke, D. (2010). SmartFactory—Towards a factory-of-things. Rev., 54(5), B84-B84.
Annual reviews in control, 34(1), 129-138. [61] Zhang, G. P., Huang, Y. M., Shi, W. H., & Fu, W. P. (2003).
[39] Tao, F., Cheng, Y., Da Xu, L., Zhang, L., & Li, B. H. (2014). CCIoT- Predicting dynamic behaviours of a whole machine tool structure based on
CMfg: cloud computing and internet of things-based cloud manufacturing computer-aided 47 engineering. International Journal of Machine Tools
service system. IEEE transactions on industrial informatics, 10(2), 1435- and Manufacture, 43(7), 699- 706.
1442. [62] Law, M., Altintas, Y., & Phani, A. S. (2013). Rapid evaluation and
[40] Rosen, R., Von Wichert, G., Lo, G., & Bettenhausen, K. D. (2015). optimization of machine tools with position-dependent stability.
About the importance of autonomy and digital twins for the future of International Journal of Machine Tools and Manufacture, 68, 81-90.
manufacturing. IFACPapersOnLine, 48(3), 567-572. [63] Wu, J., Han, Y., Xiong, Z., & Ding, H. (2017). Servo performance
[41] Luo, W., Hu, T., Zhu, W., & Tao, F. (2018, March). Digital twin improvement through iterative tuning feedforward controller with
modeling method for CNC machine tool. In 2018 IEEE 15th International disturbance compensator. International Journal of Machine Tools and
Conference on Networking, Sensing and Control (ICNSC) (pp. 1-4). IEEE. Manufacture, 117, 1-10.
[42] Guo, J., Zhao, N., Sun, L., & Zhang, S. (2019). Modular based flexible [64] Ohishi, K., Miyazaki, T., Inomata, K., Yanagisawa, H., Koide, D., &
digital twin for factory design. Journal of Ambient Intelligence and Tokumaru, H. (2006). Robust tracking servo system considering force
Humanized Computing, 10(3), 1189-1200. disturbance for the optical disk recording system. IEEE Transactions on
[43] Zhang, H., Liu, Q., Chen, X., Zhang, D., & Leng, J. (2017). A digital Industrial electronics, 53(3), 838- 847.
twin-based approach for designing and decoupling hollow glass production [65] Erkorkmaz, K., & Altintas, Y. (2001). High speed CNC system
lines. IEEE Access, 5, 26901-26911. design. Part II: modeling and identification of feed drives. International
[44] Liu, Q., Zhang, H., Leng, J., & Chen, X. (2019). Digital twin-driven Journal of Machine Tools and Manufacture, 41(10), 1487-1509.
rapid individualised designing of automated flow-shop manufacturing [66] Chen, Y. Y., Huang, P. Y., & Yen, J. Y. (2002). Frequency-domain
systems. International Journal of Production Research, 57(12), 3903-3919. identification algorithms for servo systems with friction. IEEE transactions
[45] Ruiz, E., Osmani, V., Sucar, L. E., & Mayora, O. (2019). Detecting on control systems technology, 10(5), 654-665.
dressing failures using temporal–relational visual grammars. Journal of [67] Olsson, H., Åström, K. J., De Wit, C. C., Gäfvert, M., & Lischinsky,
Ambient Intelligence and Humanized Computing, 10(7), 2757-2770. P. (1998). Friction models and friction compensation. Eur. J. Control, 4(3),
[46] Wang, J., Ma, Y., Zhang, L., Gao, R. X., & Wu, D. (2018). Deep 176-195.
learning for smart manufacturing: Methods and applications. Journal of [68] Do, N. B., Ferri, A. A., & Bauchau, O. A. (2007). Efficient simulation
Manufacturing Systems, 48, 144-156. of a dynamic system with LuGre friction.
[47] Lu, S., Zhou, P., Wang, X., Liu, Y., Liu, F., & Zhao, J. (2018). [69] Li, C. J., Tseng, H. C., Tsai, M. S., & Cheng, C. C. (2020). Novel
Condition monitoring and fault diagnosis of motor bearings using servo-feed-drive model considering cutting force and structural effects in
undersampled vibration signals from a wireless sensor network. Journal of milling to predict servo dynamic behaviors. The International Journal of
Sound and Vibration, 414, 81-96. Advanced Manufacturing Technology, 106(3), 1441-1451.
[48] Uematsu, T., & Mohri, N. (1986). Prediction and detection of cutting
tool failure by modified group method of data handling. International
Journal of Machine Tool Design and Research, 26(1), 69-80.
[49] El Wardany, T. I., & Elbestawi, M. A. (1997). Prediction of tool
failure rate in turning hardened steels. The International Journal of
Advanced Manufacturing Technology, 13(1), 1-16.
[50] Elshenawy, L. M., & Mahmoud, T. A. (2018). Fault diagnosis of time-
varying processes using modified reconstruction-based contributions.
Journal of Process Control, 70, 12-23.
[51] Kang, J., Kang, N., Feng, C. J., & Hu, H. P. (2007, November).
Research on tool failure prediction and wear monitoring based hmm
pattern recognition theory. In 2007 International Conference on Wavelet
Analysis and Pattern Recognition (Vol. 3, pp. 1167-1172). IEEE.
[52] Park, S., Kapoor, S. G., & DeVor, R. E. (2006). Microstructure-level
model for the prediction of tool failure in WC-Co cutting tool materials.
[53] Sanislav, T., & Miclea, L. (2012). Cyber-physical systems-concept,
challenges, and research areas. Journal of Control Engineering and Applied
Informatics, 14(2), 28- 33.
[54] Kelly, S. D. T., Suryadevara, N. K., & Mukhopadhyay, S. C. (2013).
Towards the implementation of IoT for environmental condition
monitoring in homes. IEEE sensors journal, 13(10), 3846-3853.
[55] Pandilov, Z., Milecki, A., Nowak, A., Górski, F., Grajewski, D.,
Ciglar, D., ... & Mulc, T. (2015). Virtual modelling and simulation of a
CNC machine feed drive system. Transactions of FAMENA, 39(4), 37-54.
[56] Pandilov, Z., & Dukovski, V. (2014). Analytical Determination of the
CNC Machines High-Speed Feed Drives Position Loop Gain. In Applied
Mechanics and Materials (Vol. 555, pp. 505-510). Trans Tech Publications
Ltd.
[57] Ramesh, R., Mannan, M. A., & Poo, A. N. (2005). Tracking and
contour error control in CNC servo systems. International Journal of
Machine Tools and Manufacture, 45(3), 301-326.