[invited review paper] machine health management …...formance life until complete failure [1, 3,...

Journal of Mechanical Science and Technology 32 (3) (2018) 987~1009

www.springerlink.com/content/1738-494x(Print)/1976-3824(Online) DOI 10.1007/s12206-018-0201-1

[Invited Review Paper]

Machine health management in smart factory: A review† Gil-Yong Lee1, Mincheol Kim2, Ying-Jun Quan2, Min-Sik Kim2, Thomas Joon Young Kim2,

Hae-Sung Yoon3, Sangkee Min3, Dong-Hyeon Kim4, Jeong-Wook Mun2, Jin Woo Oh2, In Gyu Choi2, Chung-Soo Kim5, Won-Shik Chu1, Jinkyu Yang6, Binayak Bhandari7, Choon-Man Lee8,

Jeong-Beom Ihn9 and Sung-Hoon Ahn1,2,10,11* 1Institute of Advanced Machines and Design (IAMD), Seoul National University, Seoul 08826, Korea

2Department of Mechanical and Aerospace Engineering, Seoul National University, Seoul 08826, Korea 3Department of Mechanical Engineering, University of Wisconsin-Madison, Madison, WI 53706, USA

4BK21 Plus Transformative Training Program for Creative Mechanical and Aerospace Engineers, Seoul National University, Seoul 08826, Korea 5Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge, MA 02139, USA

6William E. Boeing Department of Aeronautics & Astronautics, University of Washington, Seattle, WA 98195, USA 7Department of Railroad Integrated System, Woosong University, Daejeon 34518, Korea

8School of Mechanical Engineering, Changwon National University, Changwon 51140, Korea 9Boeing Research and Technology, Boeing, Seattle, Washington 98108, USA

10Graduate School of Engineering Practice, Seoul National University, Seoul 08826, Korea 11Institute of Engineering Research, Seoul National University, Seoul 08826, Korea

(Manuscript Received June 23, 2017; Revised November 30, 2017; Accepted December 17, 2017)

----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Abstract In this paper, we present a review of machine health managements for the smart factory. As the Industry 4.0 leads current factory

automation and intelligent machines, the machine health management for diagnostic and prognostic purposes are essential, and their importance is getting more significant for the realization of the smart factory in the Industry 4.0. After brief introductions to important concepts and definitions composing smart factory and Industry 4.0, the developments in maintenance strategies towards Prognostics and health management (PHM) of machines are summarized. The review of machine health managements is followed, classifying the refer-ences by the monitoring components, types of measurements, as well as PHM tools and algorithms. 94 existing articles are reviewed and summarized in this regard. The implementations of machine health managements within the smart factory are discussed in terms of data connectivity, communications, Cyber-physical system (CPS) and virtual factory, relating them to Internet of things (IoT), cloud comput-ing, and big data management.

Keywords: Smart factory; Machine health management; Virtual factory; Cloud manufacturing ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 1. Introduction

The Prognostics and health managements (PHM) of ma-chines have been important research area for last several dec-ades [1-8]. Different types of machine maintenance strategies have been proposed and implemented for various types of machines for the PHM purpose [1, 9-15]. As Industry 4.0 and smart factory lead the intelligent machines and factory auto-mations with data communications in the manufacturing tech-nologies, health managements of local machines are gathering more significant importance for the purpose of machine diag-nostics and prognostics [16-20].

This work focuses on reviews, and summarizes of different types of machine health managements techniques, classifying them by the types of monitoring components, physical meas-urements for the machine condition monitoring, and various algorithms used for the health managements. In Sec. 2, impor-tant concepts, recent developments, research trends, and out-looks of the smart factory and Industry 4.0 are introduced. The developments in the machine maintenance strategies towards prognostics and health managements are discussed and sum-marized as well. In Sec. 3, 94 existing articles for the machine health managements are reviewed and summarized. We pre-sent the classifications of the references by the monitoring components, sensor (Measurement) types for machine moni-toring, and different types of algorithms and PHM tools used for the machine health managements. In Secs. 4 and 5, the

*Corresponding author. Tel.: +82 2 880 7110, Fax.: +82 2 888 9073 E-mail address: [email protected]

† Recommended by Editor Hyung Wook Park © KSME & Springer 2018

988 G.-Y. Lee et al. / Journal of Mechanical Science and Technology 32 (3) (2018) 987~1009

implementations of machine health managements within the smart factory are discussed in terms of data connectivity, communications, cyber-physical system, and virtual factory, including Internet of things (IoT), cloud computing, and big data management.

2. Smart factory and development of machine health

managements

The Industry 4.0 is represented by the smart factory con-nected based on the Information and communication technol-ogy (ICT), including IoT, cloud computing, and big data man-agements into the Cyber-physical system (CPS) [9, 21-29]. The CPS will not only be composed of the virtual components and virtual machines in the virtual factory, but also have the information of the virtual products to be produced by the physical factory [9, 17, 30-32].

As shown in Fig. 1, smart factories are monitored, managed, and controlled within the CPS, through the advanced networks, such as industrial Wide area network (WAN) or Wireless local area network (WLAN), and etc. The multiple types of users can access the factories through the CPS and virtual factory to get the information on numerous products [9, 17, 33-35]. The users can design, customize, and order their prod-ucts using this cyber environment and they can trace the entire production history of their products until they get the physical products. Recent developments in high speed networks, cloud computing based on big data managements, and a variety of security solutions will leverage the realization of this revolu-tion in the near future [21-29, 36, 37].

Another essential development that needs to be considered in the physical factory towards smart factory for the realiza-tion of Industry 4.0 is machine health monitoring and man-agement [38-42]. The machines in smart factories cooperate together in more autonomous and robust ways. They need to be carefully monitored and managed to be operated within

near zero breakdown, as well as to minimize the downtime caused by local machine or component fault. Thus, different types of measurements by the appropriate sensors capture the health status of machines and monitoring components in real time. The local communications and data acquisition tech-niques use Bluetooth, RFID (Radio-frequency identification), Zigbee, RS232, and etc., record these data and transmit to the factory level database via Wi-Fi, or other factory networks [43-46]. The collected data is monitored and analyzed by the control center, and appropriate maintenance strategies are selected and applied for the individual machine and compo-nent. Furthermore, because the machines and components systems are highly networked between more robust, flexible, multi-functional, and intelligent machines in smart factory as shown in Fig. 2, the significance of their reliable health man-agements should be more emphasized for the realization of Industry 4.0.

As Lee et al. discussed [10], in modern factories and ma-chines, their maintenance strategies are being transformed from the traditional fail-and-fix (Diagnostics) practices to-wards predict-and-prevent methodologies (Prognostics). This transformation has driven much research interest for the de-

Fig. 1. Conceptual illustration of communications between organizations in Industry 4.0.

Fig. 2. System structures of factory in Industry 3.0 and Industry 4.0 (F: Factory, M: Machine, C: Component).

G.-Y. Lee et al. / Journal of Mechanical Science and Technology 32 (3) (2018) 987~1009 989

velopment of PHM. A schematic illustration presented in Fig. 3 shows a comparison between diagnostics and prognostics approaches for health managements (Figure is redrawn from Ref. [10] with the permission). While diagnostics is investi-gated to evaluate and analyze the nature of conditions, or problems of faults, as well as to identify failure modes in the machine or system, prognostics is conducted to calculate and predict the future status of machines or components with the estimation of Remaining useful life (RUL) or remaining per-formance life until complete failure [1, 3, 10]. Generally, prognostics is for monitoring, and detecting the initial indica-tions of performance degradation of the component. Because diagnostics is the process of identifying the relationship be-tween the cause of failure (Machine conditions) and the failure results, PHM also includes diagnostics [10].

The development of PHM has been leveraged by the tradi-tional maintenance strategies, such as Reactive maintenance (RM), Condition-based maintenance (CBM), Reliability cen-tered maintenance (RCM), Preventive maintenance (PM) as well as the recently developed maintenance methodologies, e.g., eMaintenance based on ICT, self-maintenance, Engineer-ing immune system (EIS), and so on. Different maintenance strategies are well reviewed in other references. Here, we pre-sent a brief summary of different maintenance strategies, as shown in Table 1 [1, 9-15, 47-49].

3. Machine health managements - reviews

As the factories are getting highly networked and machines are getting integrated and smarter, the procedures for their health managements will be considered in more complicated approaches. This transformation leads to lots of concepts and studies on the factory level managements, automations, and maintenance methodologies [11, 14, 38, 47, 50]. Meanwhile, the health status of a local machine can be well evaluated and characterized by the conventional machine diagnostics and prognostics approaches [2, 4, 9, 10]. This procedure is pre-sented as a schematic illustration in Fig. 4. Once a monitoring component (Critical component) is identified, an appropriate measurement type for monitoring each component status needs to be determined (Monitoring component identification). Not only single but multiple components can also be moni-

Table 1. Comparison of different maintenance types.

Maintenance strategy [1, 9-15, 47-49]

Reactive maintenance Condition-based maintenance (CBM)

Reliability centered maintenance (RCM) eMaintenance Preventive maintenance

(PM)

• Corrective maintenance • Fixes when it breaks • Minimal important com-

ponent

• Predictive maintenance (PdM) • Machine/component con-

dition monitoring • Condition-based diagnos-

tics/prognostics • Continuous data acquisi-

tion, data processing, and decision making

• Analyze all the possible failure modes • Customize a maintenance

strategy for individual machine • High costs

• Integration of ICT within the maintenance • Web-based, tether-free,

wireless maintenance

• Scheduled maintenance • Periodic (Predetermined

intervals) inspection • Relying on historical data

or life cycle calculation

Fig. 3. Diagnostics and prognostics health managements (Figure is redrawn from Ref. [10] with the permission from Elsevier).

Fig. 4. Schematic illustration of machine health management in the smart factory.


tored with single or multiple types of sensors to get the com-ponent or machine conditions (Data acquisition). In smart factory, the data captured in the individual component is trans-mitted to the factory level data base towards the control center. In the control center, the data by different sensors is processed and extracted to represent the health status of the component and machine (Health status assessment). Appropriate algo-rithms and feature extraction modules are used for this step.

The maintenance or machine control decisions are made based on the health status assessment, and are then applied to the local machine. In smart factory, this control and mainte-nance will be conducted autonomously.

In this section, articles presenting the local machine and component health managements are reviewed. Among numer-ous studies on different types of machines’ or structures’ health monitoring and managements, this work focuses on the manu-facturing machines for the applications to the smart factory and Industry 4.0. Examples of different types of manufacturing machines, along with some of their critical components and representative measurement types, are shown in Fig. 5.

3.1 Critical components in manufacturing machines

The critical components of machines can be identified based on the component fault frequency with respect to the machine downtime caused by the component fault. An exam-ple of this is presented as a plot in Fig. 6 (Figure is redrawn from Ref. [10] with the permission). The components can be divided into four different groups on this plot. The compo-nents in the upper right group cause long average downtime, and their faults occur frequently. In modern machines, those components need to be perfectly omitted by robust design so that the machine does not have these components at all. The components in the upper left group have short downtime al-

though their failures occur frequently. The preventive mainte-nance methodology with a scheduled and periodic mainte-nance can be applied for the components in this group with an adequate number of spare parts prepared.

The lower left group has the components for which a regu-lar maintenance will be enough for the machine operation because these components infrequently fail, and their failures can be fixed shortly. The most critical components are grouped in the lower right group. The components in this group have the most downtime although their failures occur infrequently, so the health managements need to be focused on these components [10]. In manufacturing machines, the examples of these components include machining tools, bear-ings, gears, electric motors, printing nozzles, and etc., as pre-sented in Figs. 5 and 6.

In this regard, the references are categorized and listed with respect to the monitoring (Critical) components in Tables 2-5. In this review, we classified the references into six different groups based on the monitoring components used, i.e., the rotary component, machining and micro machining tool or process, electric component, components in non-conventional manufacturing machines, and other components. Tables 2-5 provide the detailed information of the references, i.e., the monitoring components, the measurement types, the health status assessment algorithms used, as well as the applicable machines. Machining/micro machining, and non-conventional manufacturing/other are grouped together. The references listed in Tables 2-5 are also plotted with respect to the moni-toring components, as in Fig. 7.

Focusing on the manufacturing machines, the most fre-quently monitored types of components for their health man-agements, are the rotary components and the machining tools or processes, including micro machining. 78 % of the refer-ences considered these components as the most important. Other references that monitored the electrical components, as well as the components in non-conventional manufacturing processes, such as laser and plasma processes or 3D printing, are also included in this review. The references that studied the health management of manipulators, structural compo-nents, and screw threads were categorized as other examples of the monitoring components.

Fig. 5. Classifications of monitoring components and the sensor meas-urement types in different types of manufacturing machines.

Fig. 6. Average downtime caused by components faults (Figure is redrawn from Ref. [10] with the permission from Elsevier).


Table 2. Summary of the references - machine health managements with rotary components monitoring.

Ref. No. Country, year Monitoring components Measurements Health status assessment algorithms Applicable machines

[51] Canada, 2016 Bearing Acoustic emission Empirical mode decomposition (EMD), Wavelet transform (WT) Rotary machine

[52] Canada, 2016 Bearing Vibration EMD Rotary machine

[53] UK, 2012 Bearing Vibration Fourier transform (FT) Turbo generator, rotary machine

[54] UK, 2015 Bearing Acoustic emission FT Centrifugal compressor

[55] Hong Kong, 2009 Bearing, shaft Vibration EMD Rotary machine

[56] USA, 2007 Bearing Vibration Approximated entropy (ApEn) Rotary machine

[57] Sultanate of Oman, 2001 Bearing Vibration Artificial neural network (ANN) -

[58] USA, 2008 Bearing Vibration Time domain analysis, analytical model Rotary machine

[59] France, 2015 Bearing Vibration Hilbert-Hwang transform (HHT), Support vector machine (SVM)

Machining center, Rotary machine

[60] India, 2016 Bearing Vibration, rotating speed ANN, SVM Rotary machine [61] Canada, 2011 Bearing Vibration Proportional hazards model Rotary machine [62] Slovenia, 2015 Bearing Vibration FT Rotary machine [63] UK, 2016 Bearing Acoustic emission ANN Rotary machine

[64] China, 2016 Bearing, impeller Vibration EMD Rotary machine, centrifugal pump

[65] China, 2016 Bearing Vibration Principal component analysis (PCA), Neural network (NN) Rotary machine

[66] USA, 2006 Bearing Vibration HHT, EMD Rotary machine [67] USA, 1993 Bearing Vibration ANN Rotary machine [68] UK, 2009 Bearing Acoustic emission Time domain analysis Rotary machine [69] USA, 2000 Bearing Vibration NN, FT Rotary machine [70] UK, 2011 Bearing Acoustic emission, vibration Time domain analysis, FT Rotary machine [71] UK, 2008 Bearing Acoustic emission Time domain analysis Rotary machine [72] UK, 2007 Bearing Vibration, wear Gaussian mixture model (GMM) Rotary machine [73] Taiwan, 2016 Bearing, shaft Vibration NN Rotary machine [74] UK, 2007 Gear Acoustic emission, vibration, oil Time domain analysis Aircraft, helicopter

[75] Ireland, 2013 Gear, bearing Acoustic emission, electric voltage, power consumption FT Air compressor

[76] China, 2016 Gear, bearing, shaft Vibration Statistical pattern recognition (SPR), PCA Rotary machine

[77] Greece, 2009 Gear Acoustic emission, vibration WT Rotary machine [78] China, 2016 Gear, bearing Vibration ANN Electric machine

[79] UK, 2014 Gear, shaft, structure

Vibration, deformation, acoustic emission, oil Hidden Markov model (HMM) Wind turbine,

rotary machine [80] China, 2016 Gear, bearing Vibration EMD, PCA Rotary machine [81] China, 2016 Gear Vibration EMD, SVM Rotary machine [82] USA, 2012 Gear Acoustic emission EMD Rotary machine [83] Greece, 2011 Gear Vibration, acoustic emission, oil Time domain analysis, FT, PCA Rotary machine

[84] China, 2011 Hydraulic pump Vibration HMM Hydraulic system, rotary machine

[85] Taiwan, 2003 Shaft, turbine, generator

Flow rate, temperature, pressure, electric voltage, rotating speed Kalman filter, Petri nets Power plant, turbo

machine, generator

[86] USA, 2011 Shaft Temperature FT Turbine, engine, power plant

[87] China, 2014 Shaft Vibration SVM Hydropower, rotary machine


Table 3. Summary of the references - machine health managements with components of machining tools or processes including micro machining.

Ref. No. Country, year Monitoring components Measurements Health status

assessment algorithms Applicable machines

[88] Japan, 2008 Cutting Temperature Analytical model Machining center [89] China, 2012 Cutting Power consumption Experimental model Machining center

[90] France, 2014 Cutting, bearing Vibration Extreme learning model (ELM), WT Machining center, rotary machine

[91] USA, 2008 Cutting Temperature Experimental model Machining center [92] Myanmar, 2011 Drilling Force, torque, acoustic emission Fuzzy logic Machining center [93] South Korea, 2002 Drilling Power consumption, torque Time domain analysis Machining center

[94] Spain, 2004 Milling Force, acoustic emission, vibration, wear Time domain analysis, FT Machining center

[95] South Korea, 2008 Milling, drilling Force, vibration Numerical model, FT Machining center

[96] Japan, 1998 Milling, drilling, grinding Acoustic emission ANN Machining center

[97] UK, 2006 Milling, turning Vision, wear SPR Machining center [98] Turkey, 2011 Milling Force, torque ANN Machining center

[99] USA, 2010 Milling, drilling, turning Power consumption - Machining center

[100] UK, 2013 Milling Power consumption Analytical model Machining center [101] Singapore, 2006 Milling Acoustic emission, force, vibration Time domain analysis, FT, WT Machining center [102] Italy, 1996 Milling Power consumption Experimental model Machining center [103] Germany, 2014 Milling Force Time domain analysis Machining center [104] USA, 2009 Milling Force, power consumption Experimental model Machining center [105] USA, 2012 Milling Force, deformation Analytical model, FT Machining center [106] Australia, 2011 Milling, turning Power consumption Experimental model Machining center [107] India, 2015 Turning Vision SPR Machining center, lathe [108] China, 2011 Turning Vibration WT, FT, Logistic regression (LR) Machining center [109] India, 2016 Turning Force Experimental model Machining center [110] Ireland, 2013 Turning Deformation Time domain analysis, FT Machining center

[111] Malaysia, 2010 Turning Temperature, deformation, surface roughness Experimental model Machining center

[112] USA, 2013 Welding Temperature Time domain analysis Ultrasonic joining, metal welding

[113] India, 2017 Micro drilling Force, wear ANN Micro machining, machining center

[114] UK, 2017 Micro milling Acoustic emission Short time Fourier transform (STFT), regression trees, NN Machining center

[115] Taiwan, 2010 Micro milling Sound pressure Fisher linear discriminant (FLD), FT Micro machining, machining center

[116] Germany, 2012 Micro EDM, laser Acoustic emission FT Micro machining, micro EDM, laser ablation

[117] Belgium, 2017 Micro EDM Electric voltage, electric current, EDM pulse Time domain analysis Micro machining, micro

EDM [118] Poland, 2016 Micro milling Vision, wear SPR Micro machining [119] China, 2017 Micro grinding Deformation, wear grains SPR Machining center [120] Spain, 2016 Micro drilling Force, vibration Time domain analysis Micro machining

[121] Italy, 2015 Micro milling, micro EDM

Power consumption, electric voltage, electric current Time domain analysis Machining center

[122] India, 2016 Micro turning Deformation Bayesian network Micro machining, machining center

[123] South Korea, 2016 Micro milling Force, torque WT, FLD, HMM Micro machining, machining center

[124] China, 2017 Micro milling Vision PCA Micro machining, machining center


While most of the works in the references focused on a sin-gle type of monitoring component based on our classification, in some references, they monitored two or more types of com-ponents for the machine health managements. For example, when the electrical motor is monitored, the bearing, or shaft is usually monitored together to assess the health status of the machine. 68 references studied single monitoring component, 17 references monitored two monitoring components, and 9 references considered three monitoring components in the 94 references reviewed in this work.

Based on the classification shown in Tables 2-5, we plot the

frequency of different types of monitoring components with more details in Fig. 8. It can be seen that gears, bearings, and shafts represent the most frequently monitored component for the machine health assessment in the rotary components group. Among various machining tools, milling and turning tools are studied with an important consideration. In the electric com-ponents group, the electric motor is considered as the most critical component for the monitoring. This qualitative evalua-tion of the critical components for the machine health man-agements gives us a good reference for the identification of the critical components within different machines.

Table 4. Summary of the references - machine health managements with electric components monitoring.

Ref. No. Country, year Monitoring components Measurements Health status assessment algorithms Applicable

machines

[125] India, 2016 Generator Electric current EMD, ANN, PCA Wind turbine

[126] South Korea, 2004

Electric motor, bearing, shaft Vibration Condition based reasoning (CBR), Petri nets Rotary machine,

induction motor

[127] India, 2003 Electric motor, bearing

Electric voltage, electric current, temperature, rotating speed,

vibration STFT, FT Rotary machine,

[128] South Korea, 2007 Electric motor Vibration NN Rotary machine

[129] Sweden, 1989 Electric motor, milling, turning

Power consumption, electric current Experimental model Machining center

[130] Norway, 2012 Electric motor, shaft, blower Vibration WT, ANN Pressure blower,

rotary machine


Electric motor, shaft Vibration Regression trees, time-series forecasting

techniques Compressor


Electric motor, shaft Vibration Regression trees, fuzzy logic Compressor

[133] Norway, 2014 Electric motor, bearing Vibration ANN Rotary machine

[134] USA, 2010 Electric motor, shaft Vibration Time domain analysis, FT Rotary machine

[135] Switzerland, 2012

Electric motor, spindle, cutting Force, torque, electric current Analytical model Machining center,

rotary machine

Table 5. Summary of the references - machine health managements with other monitoring components including components in non-conventional manufacturing.

Ref. No. Country, year Monitoring components Measurements Health status assessment algorithms Applicable

machines

[136] Germany, 2006 Laser Temperature Experimental model Laser cladding

[137] Germany, 2007 Plasma Acoustic emission, temperature Time domain analysis, FT Plasma process, high temperature process

[138] USA, 2016 3D printing nozzle Acoustic emission HMM 3D printer, FDM

[139] Japan, 2017 3D printing nozzle Acoustic emission FT 3D printer, FDM

[140] Germany, 2016 Manipulator Force, power consumption, electric current Time domain analysis Robot, factory

[141] USA, 2010 Manipulator Force SVM, PCA Assembly, robot

[142] Saudi Arabia, 2017 Screw thread Vision SPR Screw thread

measurement, other

[143] USA, 2001 Structure Vibration Auto regression (AR) Other

[144] USA, 2015 Structure Stress Numerical model Aircraft


The examples of machine health managements for different monitoring components are introduced and summarized in the following subsections.

3.1.1 Rotary component

Rotary components are one of the most critical components of current and future manufacturing machines. The machine is composed of large amounts of bearings, gears, shafts, and other types of rotary elements. Sinha et al. detected faults of bearings in Turbo-generator (TG). They also proposed a method to reduce the number of sensors by enhancing the computational effort in signal processing [53]. Li et al. pre-sented a methodology for rotational machine health monitor-ing and fault detection using EMD-based AE feature quantifi-cation [82]. Elforjani presented experimental and analytical studies focused on the estimation of remaining useful life for bearings whilst in operation [63], and Zarei et al. proposed a method based on Artificial neural networks (ANNs) to detect bearing defects of induction motors [133].

3.1.2 Machining tool and process

The machine tool (in cutting, milling, drilling, etc.) is an-other critical component in the machining processes. There have been tremendous efforts to monitor the machine tools for the enhancement of their life cycle and energy consumption,

as well as failure predictions. Inasaki presented the application of acoustic emission sensor for monitoring the cutting process with single point and multi point cutting tools, as well as the grinding process [96]. Tun et al. also presented a process monitoring system in drilling process on a high speed machin-ing center [92]. Lanz et al. used the measured energy as an indicator for sustainability of the manufacturing in drilling and turning process [99]. Tristo et al. developed an Arduino-based framework for a real-time energy consumption monitoring of the micro Electric discharge machining (EDM) process [121]. They measured the voltage and current signals and then calcu-lated the power consumption during the process. As we will discuss in the following section, various types of signals were used to specify the machine tools’ status, i.e., force/torque, deformation, vibration, acoustic emission, and temperatures.

3.1.3 Electric component

Because many machines are driven by electric motors, it is essential to predict their failures by proper techniques with prognostic tools. The mechanical vibration measured from the motors is one of the most popular techniques for this purpose. Su et al. proposed an analytical redundancy method using neural network modeling of the induction motor in the vibra-tion spectra for machine fault detection and diagnosis [128]. Singh et al. proposed a machine monitoring system with an

Fig. 7. Different types of monitoring (Critical) components in the reference.

Fig. 8. Frequency of different types of monitoring components in the reference.


online analysis from vibration signals using c++, and offline analysis with MATLAB [127]. Four different types of ap-proaches were compared: Narrow band analysis, variable band harmonic analysis, band frequency measurement, and spectrum masking. Suratsavadee et al. proposed a novel health monitoring system of electric machine based on wireless sen-sor network (ZigBeeTM/IEEE802.15.4 Standard) [134]. The experimental results of the proposed severity detection tech-nique of rotor vibration under different levels of imbalance conditions were investigated. Another practical technique for electric motors’ health monitoring is to use electric signals such as current, voltage or power consumption. We will dis-cuss further on this topic in the following sections.

3.1.4 Component in other manufacturing machine

Recent advances in the manufacturing technology enable us to fabricate various engineering parts by 3D printing. Numer-ous engineering materials are being used as printing materials through different types of 3D printing techniques. Obviously, the future cloud manufacturing should be driven by the 3D printer. The printing nozzle and printing material feeding (i.e., Filament) are examples of the critical components to be moni-tored in 3D printer. Wu et al. proposed a method to monitor a real-time lightweight Additive manufacturing (AM) machine condition [138]. A Fused deposition method (FDM) was em-ployed in their experiment. Lotrakul et al. also presented an alternative to monitor filament feeding process in FDM using an acoustic emission sensor [139]. In addition to these, other types of manufacturing machines were also studied for their health managements. Rodriguez et al. presented analytical and experimental studies on failure detection in automated parts assembly using the force signature captured during the contact phase of the assembly process [141]. They used a less expen-sive sensor (A single-axis load cell instead of a six-axis force/torque sensor) to provide enough information to detect failure. Zhao et al. developed a transient temperature monitor-ing technique for ultrasonic joining processes [112]. They

fabricated micro Thin film thermocouples (TFTCs) on thin silicon substrates, which were then inserted in the welding anvil as a permanent feature so that the sensors were always located about 100 mm directly under the welding spot during joining of multilayer Ni-coated Cu thin sheets for battery as-sembly. Courtois et al. presented a thermal method (Lock-in thermography) to detect defect propagation during shutdown of plasma facing components for modern steady state fusion devices [137].

3.2 Data acquisition with different measurement types

While various types of measurements are used for the health managements of different components, a proper selection of sensor type which measures and represents the status of ma-chine or component, needs to be addressed. As summarized in Tables 2-5, single measurement type is frequently used for the machine and component health monitoring, however, multiple types of measurements are also used simultaneously. In sum-mary, 63 references used single measurement, 17 references used two measurement types, 14 references used three or more measurement types for the machine health monitoring, in the 94 references reviewed in this work. The measurement types used in the references (Tables 2-5), are also sorted into 20 different measurements. Examples of the types of measure-ments include vibration, acoustic emission, and electric signal (Voltage, current, or power consumption). We extracted the frequencies of each measurement type used in the references and plotted them on Fig. 9.

Since vibration signals well represent the dynamics of ma-chines, it has been used most frequently for the machine health managements. Other important measurement types are acoustic emissions, force/torque, temperature, and electric signals. To investigate which type of signal is measured within different types of monitoring components, the meas-urements extracted from the references are grouped by six types of monitoring components, i.e., the rotary, machining,

Fig. 9. Frequency of different types of measurements in the reference.


micro machining, electric, non-conventional manufacturing, and other components. They are then plotted in Fig. 10 based on the frequency of different types of measurements used within different types of monitoring components. For the ro-tary components, vibration and acoustic emission signals are preferred for their health monitoring and managements, as shown in Fig. 10. Since the gears and bearings usually consist of oils for the lubrication, investigation of oil condition is an-other useful measurement type for those components.

Because the cutting force or torque provides important in-formation on the various machining conditions in the machin-ing process, force and torque signals are mostly used with vibration and acoustic emission signals for monitoring ma-chining tools and processes. Another important measurement type in the machining process is power consumption for esti-mating the cutting load indirectly, or for optimizing the energy consumed during the process. Although the energy saving issues in the smart factory is beyond the scope of this review, it is another essential field to be studied for the realization of the Industry 4.0 and to make the factory smarter. Vibration, electric signals, and rotating speed are useful for the health monitoring and managements of the electric components, and recent studies presented the feasibility of using acoustic emis-sion signal for the health managements of non-conventional manufacturing processes, such as laser cladding, plasma proc-ess, and 3D printing systems. As the factory and the machines will be more complicated and robust, the studies on the other types of measurements with the developments of novel and cost-effective sensor systems need to be followed in the era of Industry 4.0. We introduce and summarize the examples of different measurement types used for the machine health monitoring in the following subsections.

3.2.1 Vibration signals

Because the mechanical vibration gives us lots of informa-tion about the machines’ health status, it is one of the most practical measurement for the health monitoring purposes. Yan et al. presented a new approach to machine health moni-toring based on the Approximate entropy (ApEn), which is a statistical measurement that quantifies the regularity of a time series, such as vibration signals measured from an electrical

motor or a rolling bearing [56]. A numerical simulation of an analytic signal is presented and verified experimentally. Chen et al. presented analytical and experimental studies on ma-chine tool wear monitoring for the cutting tools based on lo-gistic regression model by using vibration signals [108]. Soualhi et al. presented a new approach that combines the Hilbert-Huang transform (HHT), the Support vector machine (SVM), and the Support vector regression (SVR) for the monitoring of ball bearings [59]. The proposed approach uses the HHT to extract new heath indicators from station-ary/nonstationary vibration signals in order to track the degra-dation of the critical components of bearings.

3.2.2 Acoustic emission sensors

Acoustic emission is another important practical measure-ment for the machine health monitoring. Mba reported the application of AE technique for identifying the size of a defect on a radially loaded bearing [71]. Wang et al. introduced a multi-source energy-harvesting-powered wireless acoustic emission monitoring system for the applications of rotating machine fault detections [75]. Elforjani et al. reported the AE measurements to detect and monitor the growth of defect as well as to locate natural defect initiation and propagation in a conventional rolling element bearing [68]. They also investi-gated the source characterization of AE signals associated with a defective bearing during operation. Acoustic emission was monitored continuously in order to investigate natural life degradation of gears.

3.2.3 Force and torques

The cutting force or torque is a useful data for a real-time diagnostic and prognostic of machine tools. Saurabh Aggar-wal, et, al. presented an enhanced methodology for cutting torque prediction from the spindle motor current, readily available in modern machine tool controllers [135]. The pre-dicted cutting torque was further used to accurately predict the cutting forces and chatter-free regions for milling process planning purpose. Kaya et al. presented evaluation of cutting conditions on tool life and proposed an effective hybrid strat-egy to estimate tool flank wear while milling of Inconel 718 super alloy [98].

Fig. 10. Frequency of different types of measurements used within different monitoring components.


3.2.4 Deformation or strains Deformation or strain is another important physical data

useful for the machine health monitoring. Ma et al. developed sensor module utilizing a thin film Polyvinylidene fluoride (PVDF) piezoelectric strain sensor with an in-situ data logging platform for monitoring the feed and transverse forces in the peripheral end milling process [105]. Denkena et al. devel-oped a sensory z-slide that “Feels” the machining process, equipped with several strain sensors to the spindle sets of a 5-axis machining center [103]. By measuring the deformation of each point in the machine tool sets, they evaluated the process forces during milling. Stoney et al. presented multiple sensor networks for monitoring tool condition [110]. Strain-sensitive wireless passive SAW sensor system and Kistler 9232A strain sensor attached to KORLOY MGEVR tool holder was used.

3.2.5 Temperature

Arrazola et al. presented an experimental study on micro-scale thermal imaging to identify the effects of changing the machinability rating of the material on the temperature for two grades of AISI 4140 [91]. Hayashi et al. also developed a mi-cro-sized in-process thermometry type sensor which was used for monitoring the thermal behavior near the cutting point [88]. Rodriguez et al. presented a resonant inductive-capacitive (L-C) circuit based wireless temperature sensor suitable for work-ing in harsh environments to monitor the temperature of rotat-ing components [86]. Bi et al. presented an experimental study on temperature signal emitted from the melt pool in laser cladding process [136]. Different measurement systems, in-cluding a photodiode, pyrometer, and CCD camera with dif-ferent functional wavelengths, were used to detect the tem-perature radiation. Suhail et al. presented parameter optimiza-tion studies of workpiece surface temperature and surface roughness in cutting process [111].

3.2.6 Electric signals

Mannan et al. investigated the feasibility of measuring the power and current signal, phase angle between current and voltage, and frequency shift of certain spectral components during the machining process (i.e. Turning, milling and drill-ing), in order to monitor tool wear and breakage [129]. For monitoring the tool wear, they established the correlation be-tween power/current and torque/force. Kim et al. adapted Wil-liams’ drilling model to represent drill wear-torque model for the estimation of real-time drill wear. They also developed the spindle motor power-torque model by calculating the motor power from the motor current and load in the machining cen-ter [93]. Some of the results regarding electric signal from machine tools have been conducted for greener processes, such as energy assessment, power demand in machining, or energy transfer. Using the power meter, Magini measured the power consumption during machining to investigate and con-firm the governing rule for the energy transfer in machining process [102]. Other study found high-speed cutting as ‘Greener process’ with less energy consumed than the conven-

tional cutting speeds [104]. They measured power demand in milling process using a WattNode Modbus power meter via an MTConnect monitoring system, which provides a standard for data exchanges and allows the access to manufacturing data with standard interfaces. Gontarz et al. introduced a mul-tichannel measurement system, which provides an energy consumption of the machine tool, calculated from the voltage by 3-phase direct voltage measurement with Hall effect cur-rent transformer [145].

3.2.7 Multiple measurements

Multiple types of measurements are used for the health managements of machines by implementing different types of sensors simultaneously. Loutas et al. reported experimental results concluded by long (~50 hours) experiments for a de-fected gear system to simulate the tooth crack [77]. The AE and vibration signals were recorded. Eftekharnejad et al. also provided an investigation that compared the applicability of AE and vibration technologies in monitoring a naturally de-graded roller bearing [70]. Rodolfo et al. reported an investi-gation of tool-wear monitoring in a high-speed machining process using a dynamometer, accelerometer, and acoustic emission sensor [94]. The cutting force was measured by a three-component platform featuring a Kistler 9257A dyna-mometer. Zeng et al. presented analytical and experimental studies on Tool condition monitoring (TCM) in high speed machining [101]. They implemented a multi-modal sensing, which includes an accelerometer, acoustic emission sensor, and a dynamometer, along with an advanced signal processing to monitor the high-speed milling process. A Kistler quartz 3-component platform dynamometer was mounted between the workpiece and machining table to measure the cutting force.

3.3 Machine health management algorithms

For the health managements of the components or machines, the measured data by the sensors are processed and extracted to represent their health status with the variety of algorithms and tools. The distribution of various algorithms in the refer-ences, sorted by their types, is shown in Fig. 11. The examples include the analytical, numerical, or experimental model based feature extraction, time domain analysis, Fourier transform (FT), Empirical mode decomposition (EMD), Neural network (NN), Wavelet transform (WT), and etc. In other review arti-cles, commonly used algorithms are well summarized with their theoretical backgrounds and the characteristic of each methodology. In this review, the algorithms used in the refer-ences within Tables 2-5, are classified and summarized to present which type of algorithm is used for different monitor-ing components and measured signals. To present this, the extracted frequency of different algorithms with respect to the types of monitoring components is summarized in Fig. 12.

For the rotary components, while most of the algorithms are investigated for their health managements, FT, NN, or EMD based algorithms are frequently used. Because there have been


numerous efforts for the analytical or numerical modeling of the machining processes, since the last several decades ago, the health status of machining tools or processes are fre-quently managed based on the analytical, or numerical model, i.e., the cutting mechanics.

The time domain analysis is another useful and preferable tool for the health managements of various types of machines because it directly uses the waveform to compare the signals. However, the time domain waveform may not provide enough information, such as the features in the frequency domain. Because Fourier transform can represent the waveform in the frequency domain (for stationary signal) with good spectrum resolution, it is also frequently used for different types of ma-chines.

As shown in Fig. 13, where the sorted algorithms with re-spect to the measurements types, time domain analysis, FT, NN, EMD and WT methods are commonly used for the health managements with vibration and acoustic emission signals. Because the force or torque signals are preferred for the ma-chining tool (Process) monitoring as well as the managements, they are frequently processed by the analytical and numerical based model, or time domain analysis, based on the cutting mechanics and so on. The statistical pattern recognition is dominantly used for the vision data processing, based on the image process. It is found that power consumption is com-monly evaluated via the analytical model and time domain analysis.

Several examples of commonly used algorithms are briefly

introduced and summarized in the following subsections.

3.3.1 Empirical mode decomposition EMD method, first introduced by Huang et al. is an analysis

technique used for processing nonlinear and nonstationary data by decomposing the localized data into a few Intrinsic mode functions (IMFs), of which the frequencies can be cal-culated from their Hilbert transforms [146]. Lei et al. summa-rized and reviewed the recent research and developments on Empirical mode decomposition (EMD) methods for fault di-agnosis of rotating machineries, further providing information about the problems and solutions, as well as their applications [5]. Various modified EMD methods can be observed. For instance, Imaouchen et al. presented an improved EMD method for bearing fault detection by using Complementary ensemble empirical mode decomposition (CEEMD) to de-compose vibration signals into IMFs with Frequency-weighted energy operator (FWEO) to demodulate the signals without the need of a pre-filtering process [52].

3.3.2 Neural network

NN analysis is the technique to simulate the systems and their functions with biological neural networks. It is useful for modeling complex systems with non-linear behaviors. Samanta et al. presented a fault diagnosis of rolling element bearings based on Artificial neural network (ANN) based procedures that directly use the features obtained from time-domain vibration signals [57]. Chang et al. developed an online fault diagnosis system for a rotating machine by using fractal theory to obtain fractal dimension and lacunarity from the features of the shaft orbital data; back-propagation net-work algorithm was used to analyze the data [73]. Jia et al. proposed a Deep neural network (DNN) based method for fault diagnosis, which can mine fault characteristics from a given frequency spectrum for various diagnosis that can by classified into different health conditions [78]. Recently, Patel and Upadhyay compared the Artificial neural network (ANN) method and the Support vector machine (SVM) method for predicting bearing faults and found that the SVM produced better results than the ANN method [60].

3.3.3 Statistical method

By analyzing previous data using statistical methods and approaches, the Remaining useful life (RUL) can be estimated and thus contribute to machine prognostics. Si et al. reviewed the statistical data driven methods for estimating the remain-ing useful life, a main component of Condition-based mainte-nance (CBM) and prognostics [40]. The different models dis-cussed include wiener processes, gamma processes, regres-sion-based models, Markovian-based models, filtering-based models, and hazard models. However, different models have also been presented by other authors. Yan and Gao presented a new approach to machine health monitoring by quantifying the regularity of a time series using a statistical method called Approximate entropy (ApEn), which can effectively charac-

Fig. 11. Frequency of different types of health managements algo-rithms used in the references.


terize the severity of structural defects [56].

3.3.4 Spectral kurtosis Spectral kurtosis is a statistical tool that measures non-

Gaussian components in a signal, as well as its location in the frequency domain, to determine the impulsiveness of a signal. Wang et al. summarized and reviewed the theory and applica-tions of Spectral kurtosis (SK), one of the powerful techniques

for vibration signal analysis, in fault detection, diagnosis, and prognosis of rotating machines [7]. New variations of SK are presented, covering the related publications from 1980 until now.

3.3.5 Support vector machine (SVM)

SVM is a computational learning model that is based on the Structural risk minimization (SRM), which allows for better

Fig. 12. Frequency of different types of health managements algorithms sorted by the types of monitoring components.

Fig. 13. Frequency of different types of health managements algorithms sorted by the types of measurements.


classification and regression analysis of the data. Widodo and Yang summarized and reviewed the classification methods for machine condition monitoring and fault diagnosis using Sup-port vector machine (SVM) [13]. The different types of SVMs are discussed, along with their applications and limitations.

3.3.6 Hybrid

Hybrid models involve the combination of different meth-ods, such as the ones mentioned above. Wang et al. presents a fault diagnosis for centrifugal pumps by integrating Comple-mentary ensemble empirical mode decomposition (CEEMD), a noise assisted method based EMD, Sample entropy (SampEn), a method for quantifying the degree of complexity in different time scales, and Random forest (RF), a classifier for the different fault modes [64]. Zhao et al. presented a gear-box fault diagnosis method by using Complementary ensem-ble empirical mode decomposition (CEEMD) to obtain Intrin-sic mode functions (IMF) and of which the Permutation en-tropy (PE) values are used as input values for a Support vector machine (SVM) [81]. Niu et al. presented an advanced Condi-tion-based maintenance (CBM) method for health assessment and prognostics by combining Reliability-centered mainte-nance (RCM) for optimizing maintenance costs and data fu-sion strategy for improving the maintenance accuracy [11]. Loutas et al. presented a more effective diagnostic scheme for gears by combining vibration and acoustic emission re-cordings with Oil debris monitoring (ODM) data; Principal

component analysis (PCA) and Independent component analysis (ICA) were used to analyze the data [83].

4. Implementation of machine health management

within the smart factory

Smart factory is one of the underlying aspects of Industry 4.0 as a prerequisite to the ultimate interface for cyber-physical system. Since a smart factory needs to put all the parts together, connecting machines, conveyors, and products with information systems, building its framework is necessary to realize and fulfill the goals of Industry 4.0 [147-149]. An example of its framework is illustrated in Fig. 14 showing the connectivity and collaboration of machines, monitoring de-vices, diagnosis tools, big data analysis, IoT, and etc. Detail reviews related to the above issues are followed in the subsec-tion. The framework also needs to be developed to fulfill more advanced form of smart factory addressing the energy effi-ciency, costs, sustainability, and so on. In Sec. 5, we summa-rize these issues with discussing the applications and future trends of smart factory.

4.1 Machine-to-machine communication

In order to overcome the limits of a conventional factory, the smart factory initiative was conducted through the collabo-ration of industry and research institutes. According to

Fig. 14. Schematic illustration of an example framework with the implementation of machine health monitoring within the smart factory.


Zuehlke et al. [24] the initiative experiment underlines a few key requirements for building a smart factory, such as strict modularization of factory components, which can be easily adapted to the environment, creation of standards at each level of manufacturing automation to minimize planning cost, and fabrication of interfaces and technologies for humans.

As Weyrich et al. highlighted in their work [150], the con-cept of smart factory is reaching out further to form its reality where spontaneous interaction among devices, products, ma-chines, and humans is made. The choice to apply either ‘Wired’ or ‘Wireless’ communication to the construction of smart factory revolves around the issue of energy-consumption and security for the data transfer within the net-work. While ‘Wired’ connection gives concrete network secu-rity along with stability, ‘Wireless’ network could form ubiq-uity, which allows data transfer between machines and devices regardless of their positions within the site [151].

4.2 Wireless and sensor networks (IoT)

Consolidation of language among factory components, such as machines, sensors, devices, and human, is a necessary step to build a digital factory. As many protocols for network communication diverse in its level and type of transmitted data, founding a base model, which may suggest a template for a digital factory regardless of its structures and type of components, is essential. This is when the concept of IoT be-comes useful [152]. Since IoT framework makes use of a cloud server that processes real time data and interacts with users, its functionality could be fully manifested when applied to integrating the swarm of data within a digital factory. Shari-atzadeh et al. describes that in order to achieve interoperability between a smart factory (Real time data) and a digital factory, three layers must be considered – data transfer protocols, data representation & presentation, and semantics & understanding of data [34].

4.3 Machine collaboration

The underlying requirement for sensors for smart factory machines is the capability of sustaining communication through the swarm of data. While sensor nodes conventionally make access to a designated network, the element of smart factory highlights the strict modularization. Without complex installation processes, the parts are to be easily adapted to the surrounding work conditions. Along with strict modulariza-tion of the components, the sensors are to connect to the net-work of the factory hub without specific configuration for its topology. One concept that could be understood under the concept of machine collaboration is ‘Opportunistic network-ing,’ which refers to the capability of the sensor to make selec-tive connection and relay data either to a hub or a nearby de-vice [153].

In a similar context, the feasibility of deploying a monitor-ing robot into a smart factory has been proposed by the re-

searchers from Leibniz University. KUKA youBot was put into a field test to see its potential to move, monitor, commu-nicate within the network and the results suggests a promising outlook for collecting data variables from the deployed youBots inside the smart factory for assessing the overall fac-tory status. The ultimate goal is to enhance the overall quality of the integrated industry by using data analysis [140].

4.4 Real-time diagnosis

Conventional manufacturing processes have been managed by manufacturing execution systems. Yet, this type of surveil-lance system fails to detect invisible abnormality and notify the work manager in charge. To cope with such issues, the researchers from Stuttgart University have suggested the im-plementation of context-aware workflows to attain a greater level of intelligence within a factory, as a way of managing machine failure in factories. The researchers categorized types of “Smart work flows” according to different tasks and they proposed a context-aware workflow system. When a failure is induced, these Smart flows utilize Context integration proc-esses (CIP) to properly handle and approach the malfunction-ing point throughout the manufacturing line. It gathers all necessary and relevant records of similar failures to make its own judgment as to which tools and method to apply [38].

The interaction between humans and robots in an industrial factory is never a negligible aspect, along with other concep-tual expectations for a smart factory such as automation, self-organization, and inter/intra connectivity between the ma-chines. As humans still collaborate with robots in some sec-tions of a factory, the coordination of movements between them greatly affects the productivity. Safety of humans work-ing with the robots is a significant issue as well. For this mat-ter, a research has been conducted to analyze and improve the capability of the industrial robots to adapt to and synchronize with humans’ body movements. The user experience (UX) has been thoroughly investigated and the workers have been interviewed. Test parameters, including individual rhythm, speed, and steps, for adaptability, were used to verify the ro-bot’s collaboration performance with the workers [154].

4.5 Big data

As the era of the fourth industrial revolution comes nearby, the effort to integrate the manufacturing process and data born at each level is rising up to the surface. One of the key tools to collect and analyze the massive amounts of data is big data technology, essential for achieving cost-efficiency and feed-back control system, which minimizes the malfunctioning of the smart factory. This could also help shape the whole factory to adapt to manufacturing new products that are suitable for the changing needs of the customers. Yin et al. delineates the five aspects of the big data by 5 simplified Vs: Volume, vari-ety, veracity, velocity, and value. The terms, ‘Volume’ and ‘Variety’, refers to the amount of data that needs to be proc-


essed by the hardware and software. Other keywords, such as ‘Veracity’ and ‘Velocity,’ are related with real time process-ing and filtering of the data collected in order to detect any system failure before it actually happens. Finally, the term, ‘Value’, is giving the assessment on how valid the extracted data is [27].

Wang et al. propose a system architecture comprised of four layers: Physical resource layer, industrial network layer, cloud layer, and supervisory control terminal layer. Each layer communicates with one another, relaying the data from the bottom to the top layer. Utilizing the framework expedites the data collection process in the cloud and opens up negotiating opportunities among the layers. This self-organizing network, backed up by massive data in its cloud, creates flexible pro-duction and manufacturing processes [22].

5. Applications and future trends of smart factory

As reviewed above, numerous efforts have been made in industries for practical applications of PHM. In overall, the IoT technology market is expected to be an 8.9 trillion USD by 2020, and system health monitoring has been widely adopted by various industries [35]. In particular, development of capability for various sensors and their communications significantly have contributed to the implementation of health monitoring for more and more complex systems.

Following the technology development, view on health monitoring/maintenance has evolved from ‘Technical matter’ or ‘Necessary evil’ to ‘Partnership’ [15]. PHM system largely influences not only the maintenance cost, but also the avail-ability, product quality, and even safety requirements, etc. Basically, structural health monitoring can save additional credit in operation and can reduce potential risk due to severe usage [6]. In addition to this, it can also contribute to man-agement improvement of the whole manufacturing facility, since downtime of a single component/machine influences the entire manufacturing system. The importance of health moni-toring and maintenance significantly increases as the system complexity/uncertainty increases. Structural health monitoring data is now included in a life-cycle cost analysis [155].

One remarkable expansion of the role of health monitoring is sustainable manufacturing. By construction of collaborative network between machine and systems, health monitoring technologies are expected to contribute not only to energy use reduction but also to recover and remanufacture. Sensor em-bedding and structural health monitoring can contribute to decision-making in remanufacturing by reducing uncertainties in product and service with improved quality. Hence, a nu-merous study has investigated sustainable manufacturing by adopting health monitoring in the smart factory. More easily, redundant repairs and maintenance of machines have been generally considered as a burden to the environment, since they produce negative environmental impacts, i.e., defective parts and used lubricants. Ajukumar and Gandhi evaluated and compared several design alternatives from the perspectives of

maintenance at the design stage [50]. In addition, dynamic collaboration of machines could contribute to effective use of resources, as well as reduced uncertainties. Following the trend, Smart manufacturing leadership act (SMLA) was intro-duced in 2015 in order to seek to establish a national smart manufacturing plan, which will result in saving resources [156].

As one application in health monitoring in smart factory, eMaintenance, which consists of remote control, collaborative maintenance, predictive maintenance, etc., has evolved and been applied to industries. eMaintenance might refer to effi-cient-maintenance, effective-maintenance, and/or enterprise-maintenance, but also includes the meaning of excellent main-tenance [47]. Muller et al. conceptually reviewed the concept of eMaintenance, and defined the eMaintenance concept, which supports proactive decision making process by means of not only e-technologies, but also e-monitoring, e-diagnosis, e-prognosis, etc [47]. They claimed that eMaintenance re-quires systematic models and methodologies more than tech-nologies, in order to efficiently manage multi-sources data/knowledge and to maximize the effect of eMaintenance. Han and Yang applied the eMaintenance concept to induction motor industries, and they developed new smart sensors and methodologies by common standards in order to minimize the downtime [15].

As another example, Lee et al. introduced Engineering im-mune system (EIS) in order to responds to the complexity and uncertainty of machines additional to self-maintenance system, [12]. The concept of EIS includes the ability to assess and predict system performance and capability of surviving any disruptions, except for serious failures. Compared to the e-maintenance concept, the EIS is more emphasized on the ca-pability of surviving probabilistic conditions, and it can be achieved by comprehensive machine database, online learning tools, cloud computing, autonomic computing, etc. With vari-ous efforts, U.S. National Institute of Standards and Technol-ogy (NIST) estimated smart manufacturing infrastructure would save 57.4 billion USD annually, of which the ‘En-hanced sensing and monitoring,’ and ‘Determining required action and implementing action’ occupied about 37.1 % among the total [23]. From the perspectives of not only the efficient manufacturing, but also the synergetic manufacturing, health monitoring of smart factory is expected to keep con-tributing significantly to the manufacturing system.

Challenges in structural health monitoring can be consid-ered from the perspectives of sensors, prognosis algorithms, and system integration, respectively. Among them, develop-ment of integrated system with systematic model was consid-ered as the most important aspect in smart manufacturing since it is harder to integrate sensing hardware and develop robust data interrogation software rather than developing sin-gle component and system-level monitoring system [157].

One of the recent advancements is a deployment of Cyber-physical systems (CPS), which monitors and synchronizes on computers with all related perspectives in real physical fac-


tory; therefore, information can be more efficiently, collabora-tively, and resiliently used [9]. Importance of systematic and integrated model has been emphasized here again, following Muller et al. above [47].

In addition to the integrated system, sensors and wiring should be optimized without interfering current machine op-eration. For the prognosis algorithm, the time-consuming learning and implementation time has been an issue, but there have also been efforts to implement network-based health monitoring system for adaptive and efficient ways of improv-ing processing time rather than centralized PC-based solution [2].

As stated before, Industry 4.0 factory evolved from a com-ponent level with precision attributes to production system with productivity attributes [9]. The manufacturing system is getting more and more complex with higher uncertainties. Various technologies are required in terms of relevant level, i.e., self-optimized for disturbance at configuration level, and sensor network at smart connection level. Nevertheless, with more smart sensors and smart machines, manufacturing indus-tries are expected to continue on dynamic collaboration, en-terprise mobility, real-world connectivity, and manufacturing intelligence [14].

6. Conclusions

In conclusion, this work presents an overall review of ma-chine health managements for the smart factory. With the importance of the PHM of machines, and the developments in the machine maintenance strategies for variety types of ma-chines, we review and summarize different types of machine health managements techniques, classifying them by the types of monitoring components, physical measurements for the machine condition monitoring, and various algorithms used for the health managements. 94 existing articles for the ma-chine health managements are reviewed by the classifications of the references by the monitoring components, sensor types for machine monitoring, as well as the different types of algo-rithms and PHM tools used for the machine health manage-ments. As Industry 4.0 and smart factory lead the factory automation, the intelligent machines, and the production lines, the significance of machines’ health managements for their diagnostics and prognostics will be more emphasized. The advanced network technologies, ICT, IoT, cloud computing, and big data managements will also leverage the development of the machine health managements towards future smart factory, as well as the realization of Industry 4.0. In this regard, this work tries to give an overall summary and perspective of the machine health managements in the smart factory and Industry 4.0.

Acknowledgment

This work was supported by the SNU-Hojeon Garment Smart Factory Research Center funded by Hojeon Ltd. This

research was also supported by Next Generation Hybrid Manufacturing Consortium of Institute of Advanced Machines and Design at Seoul National University, the Korea Basic Science Institute (KBSI) Creative Convergence Research Project (CAP-PN2017003) funded by the National Research Council of Science and Technology (NST), International S&T Cooperation Program through the National Research Founda-tion of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (MSIP) (NRF-2017K1A3A9A04013801), and SNU-Boeing collaborative research project funded by the Boeing Research & Technology (BR&T), Boeing, USA.

References

[1] A. Cubillo, S. Perinpanayagam and M. Esperon-Miguez, A review of physics-based models in prognostics: Application to gears and bearings of rotating machinery, Advances in Mechanical Engineering, 8 (8) (2016).

[2] K. Goh, B. Tjahjono, T. Baines and S. Subramaniam, A review of research in manufacturing prognostics, 2006 IEEE International Conference on Industrial Informatics (2006) 417-422.

[3] A. K. Jardine, D. Lin and D. Banjevic, A review on machinery diagnostics and prognostics implementing condition-based maintenance, Mechanical Systems and Signal Processing, 20 (7) (2006) 1483-1510.

[4] M. S. Kan, A. C. Tan and J. Mathew, A review on prognostic techniques for non-stationary and non-linear rotating systems, Mechanical Systems and Signal Processing, 62 (2015) 1-20.

[5] Y. Lei, J. Lin, Z. He and M. J. Zuo, A review on empirical mode decomposition in fault diagnosis of rotating machinery, Mechanical Systems and Signal Processing, 35 (1) (2013) 108-126.

[6] C. Stolz and M. Neumair, Structural health monitoring, in-service experience, benefit and way ahead, Structural Health Monitoring, 9 (3) (2010) 209-217.

[7] Y. Wang, J. Xiang, R. Markert and M. Liang, Spectral kurtosis for fault detection, diagnosis and prognostics of rotating machines: A review with applications, Mechanical Systems and Signal Processing, 66 (2016) 679-698.

[8] R. Ahmad and S. Kamaruddin, An overview of time-based and condition-based maintenance in industrial application, Computers & Industrial Engineering, 63 (1) (2012) 135-149.

[9] J. Lee, B. Bagheri and H.-A. Kao, A cyber-physical systems architecture for Industry 4.0-based manufacturing systems, Manufacturing Letters, 3 (2015) 18-23.

[10] J. Lee, F. Wu, W. Zhao, M. Ghaffari, L. Liao and D. Siegel, Prognostics and health management design for rotary machinery systems—Reviews, methodology and applications, Mechanical Systems and Signal Processing, 42 (1) (2014) 314-334.

[11] G. Niu, B.-S. Yang and M. Pecht, Development of an optimized condition-based maintenance system by data fusion and reliability-centered maintenance, Reliability


Engineering & System Safety, 95 (7) (2010) 786-796. [12] J. Lee, M. Ghaffari and S. Elmeligy, Self-maintenance and

engineering immune systems: Towards smarter machines and manufacturing systems, Annual Reviews in Control, 35 (1) (2011) 111-122.

[13] A. Widodo and B.-S. Yang, Support vector machine in machine condition monitoring and fault diagnosis, Mechanical Systems and Signal Processing, 21 (6) (2007) 2560-2574.

[14] A. M. Alexandru, A. De Mauro, M. Fiasché, F. G. Sisca, M. Taisch, L. Fasanotti and P. Grasseni, A smart web-based maintenance system for a smart manufacturing environment, 2015 IEEE 1st International Forum on Research and Technologies for Society and Industry Leveraging a Better Tomorrow (RTSI) (2015) 579-584.

[15] T. Han and B.-S. Yang, Development of an e-maintenance system integrating advanced techniques, Computers in Industry, 57 (6) (2006) 569-580.

[16] M. Papazoglou, W.-J. van den Heuvel and J. Mascolo, Reference architecture and knowledge-based structures for smart manufacturing networks, IEEE Software (2015).

[17] S. Choi, B. H. Kim and S. D. Noh, A diagnosis and evaluation method for strategic planning and systematic design of a virtual factory in smart manufacturing systems, International Journal of Precision Engineering and Manufacturing, 16 (6) (2015) 1107-1115.

[18] Y. Zhang, T. Qu, O. K. Ho and G. Q. Huang, Agent-based smart gateway for RFID-enabled real-time wireless manufacturing, International Journal of Production Research, 49 (5) (2011) 1337-1352.

[19] Y. Zhang, G. Q. Huang, T. Qu, O. Ho and S. Sun, Agent-based smart objects management system for real-time ubiquitous manufacturing, Robotics and Computer-Integrated Manufacturing, 27 (3) (2011) 538-549.

[20] A. M. Ghalayini, J. S. Noble and T. J. Crowe, An integrated dynamic performance measurement system for improving manufacturing competitiveness, International Journal of Production Economics, 48 (3) (1997) 207-225.

[21] H. Ramamurthy, B. Prabhu, R. Gadh and A. M. Madni, Wireless industrial monitoring and control using a smart sensor platform, IEEE Sensors Journal, 7 (5) (2007) 611-618.

[22] S. Wang, J. Wan, D. Zhang, D. Li and C. Zhang, Towards smart factory for Industry 4.0: A self-organized multi-agent system with big data based feedback and coordination, Computer Networks, 101 (2016) 158-168.

[23] G. Anderson, The economic impact of technology infrastructure for smart manufacturing, NIST, Gaithersburg, NIST Economic Analysis Briefs, 4 (2016).

[24] D. Zuehlke, Smart factory—towards a factory-of-things, Annual Reviews in Control, 34 (1) (2010) 129-138.

[25] Y. Lu and J. Cecil, An Internet of things (IoT)-based collaborative framework for advanced manufacturing, International Journal of Advanced Manufacturing Technology, 84 (2016).

[26] H.-S. Park and N.-H. Tran, Autonomy for smart manufacturing, Journal of the Korean Society for Precision Engineering, 31 (4) (2014) 287-295.

[27] S. Yin and O. Kaynak, Big data for modern industry: challenges and trends [Point of view], Proceedings of the IEEE, 103 (2) (2015) 143-146.

[28] G. Hwang, J. Lee, J. Park and T.-W. Chang, Developing performance measurement system for Internet of Things and smart factory environment, International Journal of Production Research, 55 (9) (2017) 2590-2602.

[29] X. Xu, From cloud computing to cloud manufacturing, Robotics and Computer-integrated Manufacturing, 28 (1) (2012) 75-86.

[30] J. Delaram and O. F. Valilai, Development of a novel solution to enable integration and interoperability for cloud manufacturing, Procedia CIRP, 52 (2016) 6-11.

[31] D. Wu, D. Schaefer and D. W. Rosen, Cloud-based design and manufacturing systems: A social network analysis, International Conference on Engineering Design (ICED 2013) (2013).

[32] S. Wang, J. Wan, D. Li and C. Zhang, Implementing smart factory of industrie 4.0: an outlook, International Journal of Distributed Sensor Networks (2016).

[33] L. Monostori, Cyber-physical production systems: Roots, expectations and R&D challenges, Procedia CIRP, 17 (2014) 9-13.

[34] N. Shariatzadeh, T. Lundholm, L. Lindberg and G. Sivard, Integration of digital factory with smart factory based on Internet of things, Procedia CIRP, 50 (2016) 512-517.

[35] B. K. Paul, R. Panat, C. Mastrangelo, D. Kim and D. Johnson, Manufacturing of smart goods: Current state, future potential, and research recommendations, Journal of Micro and Nano-Manufacturing, 4 (4) (2016) 044001.

[36] M. Adnan and H. Zen, ICT Convergence in internet of things-the birth of smart factories (A technical note), International Journal of Computer Science and Information Security, 14 (4) (2016) 93.

[37] J. Lee, B. Bagheri and H.-A. Kao, Recent advances and trends of cyber-physical systems and big data analytics in industrial informatics, International Proceeding of Int Conference on Industrial Informatics (INDIN) (2014) 1-6.

[38] M. Wieland, F. Leymann, M. Schäfer, D. Lucke, C. Constantinescu and E. Westkämper, Using context-aware workflows for failure management in a smart factory, Proceedings of Fourth International Conference on Mobile Ubiquitous Computing, Systems, Services and Technologies UBICOMM (2010) 379-384.

[39] K. Le Son, M. Fouladirad, A. Barros, E. Levrat and B. Iung, Remaining useful life estimation based on stochastic deterioration models: A comparative study, Reliability Engineering & System Safety, 112 (2013) 165-175.

[40] X.-S. Si, W. Wang, C.-H. Hu and D.-H. Zhou, Remaining useful life estimation–A review on the statistical data driven approaches, European Journal of Operational Research, 213 (1) (2011) 1-14.


[41] N. M. Vichare and M. G. Pecht, Prognostics and health management of electronics, IEEE Transactions on Components and Packaging Technologies, 29 (1) (2006) 222-229.

[42] J. Sikorska, M. Hodkiewicz and L. Ma, Prognostic modelling options for remaining useful life estimation by industry, Mechanical Systems and Signal Processing, 25 (5) (2011) 1803-1836.

[43] S. Cheng, M. H. Azarian and M. G. Pecht, Sensor systems for prognostics and health management, Sensors, 10 (6) (2010) 5774-5797.

[44] H. Ramamurthy, B. Prabhu, R. Gadh and A. M. Madni, Smart sensor platform for industrial monitoring and control, Sensors, 2005 IEEE (2005) 4.

[45] C. Zhu, Y. Huo, V. C. Leung and L. T. Yang, Sensor-cloud and power line communication: Recent developments and integration, Dependable, Autonomic and Secure Computing, 14th Intl Conf on Pervasive Intelligence and Computing, 2nd Intl Conf on Big Data Intelligence and Computing and Cyber Science and Technology Congress, 2016 IEEE 14th Intl C (2016) 302-308.

[46] S. C. Lee, T. G. Jeon, H.-S. Hwang and C.-S. Kim, Design and implementation of wireless sensor based-monitoring system for smart factory, International Conference on Computational Science and its Applications (2007) 584-592.

[47] A. Muller, A. C. Marquez and B. Iung, On the concept of e-maintenance: Review and current research, Reliability Engineering & System Safety, 93 (8) (2008) 1165-1187.

[48] W. Zhou, T. G. Habetler and R. G. Harley, Bearing condition monitoring methods for electric machines: A general review, Diagnostics for Electric Machines, Power Electronics and Drives, 2007. SDEMPED 2007. IEEE International Symposium on (2007) 3-6.

[49] A. Jardine, D. Banjevic, N. Montgomery and A. Pak, Repairable system reliability: Recent developments in CBM optimization, International Journal of Performability Engineering, 4 (3) (2008) 205-214.

[50] V. Ajukumar and O. Gandhi, Evaluation of green maintenance initiatives in design and development of mechanical systems using an integrated approach, Journal of Cleaner Production, 51 (2013) 34-46.

[51] M. Kedadouche, M. Thomas and A. Tahan, A comparative study between empirical wavelet transforms and empirical mode decomposition methods: Application to bearing defect diagnosis, Mechanical Systems and Signal Processing, 81 (2016) 88-107.

[52] Y. Imaouchen, M. Kedadouche, R. Alkama and M. Thomas, A frequency-weighted energy operator and complementary ensemble empirical mode decomposition for bearing fault detection, Mechanical Systems and Signal Processing, 82 (2017) 103-116.

[53] J. K. Sinha and K. Elbhbah, A future possibility of vibration based condition monitoring of rotating machines, Mechanical Systems and Signal Processing, 34 (1) (2013) 231-240.

[54] S. Gowid, R. Dixon and S. Ghani, A novel robust automated FFT-based segmentation and features selection algorithm for acoustic emission condition based monitoring systems, Applied Acoustics, 88 (2015) 66-74.

[55] J. C. Chan and W. T. Peter, A novel, fast, reliable data transmission algorithm for wireless machine health monitoring, IEEE Transactions on Reliability, 58 (2) (2009) 295-304.

[56] R. Yan and R. X. Gao, Approximate entropy as a diagnostic tool for machine health monitoring, Mechanical Systems and Signal Processing, 21 (2) (2007) 824-839.

[57] B. Samanta and K. Al-Balushi, Artificial neural network based fault diagnostics of rolling element bearings using time-domain features, Mechanical Systems and Signal Processing, 17 (2) (2003) 317-328.

[58] S. Janjarasjitt, H. Ocak and K. Loparo, Bearing condition diagnosis and prognosis using applied nonlinear dynamical analysis of machine vibration signal, Journal of Sound and Vibration, 317 (1) (2008) 112-126.

[59] A. Soualhi, K. Medjaher and N. Zerhouni, Bearing health monitoring based on Hilbert–Huang transform, support vector machine, and regression, IEEE Transactions on Instrumentation and Measurement, 64 (1) (2015) 52-62.

[60] J. Patel and S. Upadhyay, Comparison between artificial neural network and support vector method for a fault diagnostics in rolling element bearings, Procedia Engi- neering, 144 (2016) 390-397.

[61] Z. Tian and H. Liao, Condition based maintenance optimization for multi-component systems using proportional hazards model, Reliability Engineering & System Safety, 96 (5) (2011) 581-589.

[62] B. Dolenc, P. Boškoski and Đ. Juričić, Distributed bearing fault diagnosis based on vibration analysis, Mechanical Systems and Signal Processing, 66 (2016) 521-532.

[63] M. Elforjani, Estimation of remaining useful life of slow speed bearings using acoustic emission signals, Journal of Nondestructive Evaluation, 35 (4) (2016) 62.

[64] Y. Wang, C. Lu, H. Liu and Y. Wang, Fault diagnosis for centrifugal pumps based on complementary ensemble empirical mode decomposition, sample entropy and random forest, 2016 12th World Congress on Intelligent Control and Automation (WCICA) (2016) 1317-1320.

[65] B. Zhou and Y. Cheng, Fault diagnosis for rolling bearing under variable conditions based on image recognition, Shock and Vibration, 2016 (2016).

[66] R. Yan and R. X. Gao, Hilbert–Huang transform-based vibration signal analysis for machine health monitoring, IEEE Transactions on Instrumentation and Measurement, 55 (6) (2006) 2320-2329.

[67] I. E. Alguindigue, A. Loskiewicz-Buczak and R. E. Uhrig, Monitoring and diagnosis of rolling element bearings using artificial neural networks, IEEE Transactions on Industrial Electronics, 40 (2) (1993) 209-217.

[68] M. Elforjani and D. Mba, Natural mechanical degradation measurements in slow speed bearings, Engineering Failure


Analysis, 16 (1) (2009) 521-532. [69] B. Li, M.-Y. Chow, Y. Tipsuwan and J. C. Hung, Neural-

network-based motor rolling bearing fault diagnosis, IEEE Transactions on Industrial Electronics, 47 (5) (2000) 1060-1069.

[70] B. Eftekharnejad, M. Carrasco, B. Charnley and D. Mba, The application of spectral kurtosis on acoustic emission and vibrations from a defective bearing, Mechanical Systems and Signal Processing, 25 (1) (2011) 266-284.

[71] D. Mba, The use of acoustic emission for estimation of bearing defect size, Journal of Failure Analysis and Prevention, 8 (2) (2008) 188-192.

[72] S. Chen, M. Craig, R. Callan, H. Powrie and R. Wood, Use of artificial intelligence methods for advanced bearing health diagnostics and prognostics, Aerospace Conference, 2008 IEEE (2008) 1-9.

[73] H.-C. Chang, S.-C. Lin, C.-C. Kuo, C.-Y. Lin and C.-F. Hsieh, Using neural network based on the shaft orbit feature for online rotating machinery fault diagnosis, 2016 International Conference on System Science and Engineering (ICSSE) (2016) 1-4.

[74] C. K. Tan, P. Irving and D. Mba, A comparative experimental study on the diagnostic and prognostic capabilities of acoustics emission, vibration and spectrometric oil analysis for spur gears, Mechanical Systems and Signal Processing, 21 (1) (2007) 208-233.

[75] W. Wang, A. Vinco, N. Pavlov, N. Wang, M. Hayes and C. O'Mathuna, A rotating machine acoustic emission monitoring system powered by multi-source energy harvester, Proceedings of the 1st International Workshop on Energy Neutral Sensing Systems (2013) 5.

[76] Y. Wang and Y. Cheng, An approach to fault diagnosis for gearbox based on image processing, Shock and Vibration, 2016 (2016).

[77] T. Loutas, G. Sotiriades, I. Kalaitzoglou and V. Kostopoulos, Condition monitoring of a single-stage gearbox with artificially induced gear cracks utilizing on-line vibration and acoustic emission measurements, Applied Acoustics, 70 (9) (2009) 1148-1159.

[78] F. Jia, Y. Lei, J. Lin, X. Zhou and N. Lu, Deep neural networks: A promising tool for fault characteristic mining and intelligent diagnosis of rotating machinery with massive data, Mechanical Systems and Signal Processing, 72 (2016) 303-315.

[79] A. May, D. McMillan and S. Thöns, Economic analysis of condition monitoring systems for offshore wind turbine sub-systems, IET Renewable Power Generation, 9 (8) (2015) 900-907.

[80] H. Liu, J. Zhang, Y. Cheng and C. Lu, Fault diagnosis of gearbox using empirical mode decomposition and multi-fractal detrended cross-correlation analysis, Journal of Sound and Vibration, 385 (2016) 350-371.

[81] L. Zhao, W. Yu and R. Yan, Gearbox fault diagnosis using complementary ensemble empirical mode decomposition and permutation entropy, Shock and Vibration, 2016 (2016).

[82] R. Li and D. He, Rotational machine health monitoring and fault detection using EMD-based acoustic emission feature quantification, IEEE Transactions on Instrumentation and Measurement, 61 (4) (2012) 990-1001.

[83] T. Loutas, D. Roulias, E. Pauly and V. Kostopoulos, The combined use of vibration, acoustic emission and oil debris on-line monitoring towards a more effective condition monitoring of rotating machinery, Mechanical Systems and Signal Processing, 25 (4) (2011) 1339-1352.

[84] Y. Peng and M. Dong, A prognosis method using age-dependent hidden semi-Markov model for equipment health prediction, Mechanical Systems and Signal Processing, 25 (1) (2011) 237-252.

[85] S. Yang, A condition-based failure-prediction and processing- scheme for preventive maintenance, IEEE Transactions on Reliability, 52 (3) (2003) 373-383.

[86] R. I. Rodriguez and Y. Jia, A wireless inductive-capacitive (LC) sensor for rotating component temperature monitoring, International Journal on Smart Sensing and Intelligent Systems, 4 (2) (2011) 325-337.

[87] J. Bao, Z. Zhu, H. Tang, T. Lu and Q. Zhang, Apply low-level image feature representation and classification method to identifying shaft orbit of hydropower unit, 2014 Sixth International Conference on Intelligent Human-Machine Systems and Cybernetics (IHMSC) (2014) 165-168.

[88] M. Hayashi, H. Yoshioka and H. Shinno, An adaptive control of ultraprecision machining with an in-process micro-sensor, Journal of Advanced Mechanical Design, Systems and Manufacturing, 2 (3) (2008) 322-331.

[89] S. Hu, F. Liu, Y. He and T. Hu, An on-line approach for energy efficiency monitoring of machine tools, Journal of Cleaner Production, 27 (2012) 133-140.

[90] K. Javed, R. Gouriveau, N. Zerhouni and P. Nectoux, Enabling health monitoring approach based on vibration data for accurate prognostics, IEEE Transactions on Industrial Electronics, 62 (1) (2015) 647-656.

[91] P. Arrazola, I. Arriola, M. Davies, A. Cooke and B. Dutterer, The effect of machinability on thermal fields in orthogonal cutting of AISI 4140 steel, CIRP Annals-Manufacturing Technology, 57 (1) (2008) 65-68.

[92] H. M. Tun, M. Kyaw and Z. M. Naing, Development of process monitoring system in drilling process using fuzzy rules, International Journal of System Assurance Engineering and Management, 2 (1) (2011) 78-83.

[93] H. Kim, J. Ahn, S. Kim and S. Takata, Real-time drill wear estimation based on spindle motor power, Journal of Materials Processing Technology, 124 (3) (2002) 267-273.

[94] R. E. Haber, J. E. Jiménez, C. R. Peres and J. R. Alique, An investigation of tool-wear monitoring in a high-speed machining process, Sensors and Actuators A: Physical, 116 (3) (2004) 539-545.

[95] S.-G. Kim, S.-H. Jang, H.-Y. Hwang, Y.-H. Choi and J.-S. Ha, Analysis of dynamic characteristics and evaluation of dynamic stiffness of a 5-axis multi-tasking machine tool by using FEM and exciter test, International Conference on


Smart Manufacturing Application, ICSMA 2008 (2008) 565-569.

[96] I. Inasaki, Application of acoustic emission sensor for monitoring machining processes, Ultrasonics, 36 (1-5) (1998) 273-281.

[97] D. Kerr, J. Pengilley and R. Garwood, Assessment and visualisation of machine tool wear using computer vision, The International Journal of Advanced Manufacturing Technology, 28 (7) (2006) 781-791.

[98] B. Kaya, C. Oysu and H. M. Ertunc, Force-torque based on-line tool wear estimation system for CNC milling of Inconel 718 using neural networks, Advances in Engineering Software, 42 (3) (2011) 76-84.

[99] M. Lanz, M. Mani, S. Leong, K. Lyons, A. Ranta, K. Ikkala and N. Bengtsson, Impact of energy measurements in machining operations, ASME 2010 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference (2010) 867-873.

[100] V. A. Balogun and P. T. Mativenga, Modelling of direct energy requirements in mechanical machining processes, Journal of Cleaner Production, 41 (2013) 179-186.

[101] H. Zeng, T. B. Thoe, X. Li and J. Zhou, Multi-modal sensing for machine health monitoring in high speed machining, 2006 IEEE International Conference on Industrial Informatics (2006) 1217-1222.

[102] A. Iasonna and M. Magini, Power measurements during mechanical milling. An experimental way to investigate the energy transfer phenomena, Acta Materialia, 44 (3) (1996) 1109-1117.

[103] B. Denkena, K. M. Litwinski and H. Boujnah, Process monitoring with a force sensitive axis-slide for machine tools, Procedia Technology, 15 (2014) 416-423.

[104] N. Diaz, M. Helu, A. Jarvis, S. Tönissen, D. Dornfeld and R. Schlosser, Strategies for minimum energy operation for precision machining, Laboratory for Manufacturing and Sustainability (2009).

[105] L. Ma, S. N. Melkote, J. B. Morehouse, J. B. Castle, J. W. Fonda and M. A. Johnson, Thin-film PVDF sensor-based monitoring of cutting forces in peripheral end milling, Journal of Dynamic Systems, Measurement, and Control, 134 (5) (2012) 051014.

[106] S. Kara and W. Li, Unit process energy consumption models for material removal processes, CIRP Annals-Manufacturing Technology, 60 (1) (2011) 37-40.

[107] Y. Chethan, H. Ravindra and S. B. Kumar, Machine vision for tool status monitoring in turning Inconel 718 using blob analysis, Materials Today: Proceedings, 2 (4-5) (2015) 1841-1848.

[108] B. Chen, X. Chen, B. Li, Z. He, H. Cao and G. Cai, Reliability estimation for cutting tools based on logistic regression model using vibration signals, Mechanical Systems and Signal Processing, 25 (7) (2011) 2526-2537.

[109] R. Jain, J. Rathore and V. Gorana, Design, development and testing of a three component lathe tool dynamometer using resistance strain gauges, CAD/CAM, robotics and

factories of the future, Springer (2016) 13-21. [110] R. Stoney, G. E. O’Donnell and D. Geraghty, Dynamic

wireless passive strain measurement in CNC turning using surface acoustic wave sensors, The International Journal of Advanced Manufacturing Technology, 69 (5-8) (2013) 1421-1430.

[111] A. H. Suhail, N. Ismail, S. V. Wong and N. A. Jalil, Optimization of cutting parameters based on surface roughness and assistance of workpiece surface temperature in turning process, American Journal of Engineering and Applied Sciences, 3 (1) (2010) 102-108.

[112] J. Zhao, H. Li, H. Choi, W. Cai, J. A. Abell and X. Li, Insertable thin film thermocouples for in situ transient temperature monitoring in ultrasonic metal welding of battery tabs, Journal of Manufacturing Processes, 15 (1) (2013) 136-140.

[113] K. Patra, A. Jha, T. Szalay, J. Ranjan and L. Monostori, Artificial neural network based tool condition monitoring in micro mechanical peck drilling using thrust force signals, Precision Engineering, 48 (2017) 279-291.

[114] J. M. Griffin, F. Diaz, E. Geerling, M. Clasing, V. Ponce, C. Taylor, S. Turner, E. A. Michael, F. P. Mena and L. Bronfman, Control of deviations and prediction of surface roughness from micro machining of THz waveguides using acoustic emission signals, Mechanical Systems and Signal Processing, 85 (2017) 1020-1034.

[115] M.-H. Lee, M.-C. Lu and J.-C. Tsai, Development of sound based tool wear monitoring system in micro-milling, ASME 2010 International Manufacturing Science and Engineering Conference (2010) 427-434.

[116] V. Schulze, P. Weber and C. Ruhs, Increase of process reliability in the micro-machining processes EDM-milling and laser ablation using on-machine sensors, Journal of Materials Processing Technology, 212 (3) (2012) 625-632.

[117] J. Wang, J. Qian, E. Ferraris and D. Reynaerts, In-situ process monitoring and adaptive control for precision micro-EDM cavity milling, Precision Engineering, 47 (2017) 261-275.

[118] M. Szydłowski, B. Powałka, M. Matuszak and P. Kochmański, Machine vision micro-milling tool wear inspection by image reconstruction and light reflectance, Precision Engineering, 44 (2016) 236-244.

[119] X. Wen and Y. Gong, Modeling and prediction research on wear of electroplated diamond micro-grinding tool in soda lime glass grinding, The International Journal of Advanced Manufacturing Technology (2017) 1-13.

[120] F. Castaño, R. M. del Toro, R. E. Haber and G. Beruvides, Monitoring tool usage on the basis of sensory information in micro-drilling operations, 2016 IEEE International Confer- ence on Industrial Technology (ICIT) (2016) 667-672.

[121] G. Tristo, G. Bissacco, A. Lebar and J. Valentinčič, Real time power consumption monitoring for energy efficiency analysis in micro EDM milling, The International Journal of Advanced Manufacturing Technology, 78 (9-12) (2015) 1511-1521.


[122] S. Mandal, V. K. Sharma, A. Pal and Nagahanumaiah, Tool strain–based wear estimation in micro turning using Bayesian networks, Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture, 230 (10) (2016) 1952-1960.

[123] Y.-S. Hong, H.-S. Yoon, J.-S. Moon, Y.-M. Cho and S.-H. Ahn, Tool-wear monitoring during micro-end milling using wavelet packet transform and Fisher’s linear discriminant, International Journal of Precision Engineering and Manufacturing, 17 (7) (2016) 845-855.

[124] K. Zhu and X. Yu, The monitoring of micro milling tool wear conditions by wear area estimation, Mechanical Systems and Signal Processing, 93 (2017) 80-91.

[125] H. Malik and S. Mishra, Artificial neural network and empirical mode decomposition based imbalance fault diagnosis of wind turbine using TurbSim, FAST and Simulink, IET Renewable Power Generation (2016).

[126] B.-S. Yang, S. K. Jeong, Y.-M. Oh and A. C. C. Tan, Case-based reasoning system with Petri nets for induction motor fault diagnosis, Expert Systems with Applications, 27 (2) (2004) 301-311.

[127] G. Singh, Experimental investigations on induction machine condition monitoring and fault diagnosis using digital signal processing techniques, Electric Power Systems Research, 65 (3) (2003) 197-221.

[128] H. Su and K. T. Chong, Induction machine condition monitoring using neural network modeling, IEEE Transactions on Industrial Electronics, 54 (1) (2007) 241-249.

[129] M. Mannan, S. Broms and B. Lindström, Monitoring and adaptive control of cutting process by means of motor power and current measurements, CIRP Annals-Manufacturing Technology, 38 (1) (1989) 347-350.

[130] Z. Zhang, Y. Wang and K. Wang, Fault diagnosis and prognosis using wavelet packet decomposition, Fourier transform and artificial neural network, Journal of Intelligent Manufacturing, 24 (6) (2013) 1213-1227.

[131] B.-S. Yang, M.-S. Oh and A. C. C. Tan, Machine condition prognosis based on regression trees and one-step-ahead prediction, Mechanical Systems and Signal Processing, 22 (5) (2008) 1179-1193.

[132] B.-S. Yang and A. C. C. Tan, Multi-step ahead direct prediction for the machine condition prognosis using regression trees and neuro-fuzzy systems, Expert Systems with Applications, 36 (5) (2009) 9378-9387.

[133] J. Zarei, M. A. Tajeddini and H. R. Karimi, Vibration analysis for bearing fault detection and classification using an intelligent filter, Mechatronics, 24 (2) (2014) 151-157.

[134] S. Korkua, H. Jain, W.-J. Lee and C. Kwan, Wireless health monitoring system for vibration detection of induction motors, Industrial and Commercial Power Systems Technical Conference (I&CPS), 2010 IEEE (2010) 1-6.

[135] S. Aggarwal, N. Nešić and P. Xirouchakis, Cutting torque and tangential cutting force coefficient identification from spindle motor current, The International Journal of

Advanced Manufacturing Technology, 65 (1-4) (2013) 81-95. [136] G. Bi, A. Gasser, K. Wissenbach, A. Drenker and R.

Poprawe, Identification and qualification of temperature signal for monitoring and control in laser cladding, Optics and Lasers in Engineering, 44 (12) (2006) 1348-1359.

[137] X. Courtois, A. Durocher, F. Escourbiac, J. Schlosser, R. Mitteau, M. Merola and R. Tivey, In-situ monitoring of actively cooled plasma facing components using acoustic and thermal methods, Physica Scripta, 2007 (T128) (2007) 189.

[138] H. Wu, Z. Yu and Y. Wang, Real-time FDM machine condition monitoring and diagnosis based on acoustic emission and hidden semi-Markov model, The International Journal of Advanced Manufacturing Technology, 90 (5-8) (2017) 2027-2036.

[139] P. Lotrakul, W. San-Um and M. Takahashi, The monitoring of three-dimensional printer filament feeding process using an acoustic emission sensor, in sustainability through innovation in product life cycle design, Springer (2017) 499-511.

[140] J. Carstensen, T. Carstensen, M. Pabst, F. Schulz, J. Friederichs, S. Aden, D. Kaczor, J. Kotlarski and T. Ortmaier, Condition monitoring and cloud-based energy analysis for autonomous mobile manipulation-smart factory concept with LUHbots, Procedia Technology, 26 (2016) 560-569.

[141] A. Rodriguez, D. Bourne, M. Mason, G. F. Rossano and J. Wang, Failure detection in assembly: Force signature analysis, 2010 IEEE Conference on Automation Science and Engineering (CASE) (2010) 210-215.

[142] E. Gadelmawla, Computer vision algorithms for measurement and inspection of external screw threads, Measurement, 100 (2017) 36-49.

[143] H. Sohn and C. R. Farrar, Damage diagnosis using time series analysis of vibration signals, Smart Materials and Structures, 10 (3) (2001) 446.

[144] M. J. Pais and N. H. Kim, Predicting fatigue crack growth under variable amplitude loadings with usage monitoring data, Advances in Mechanical Engineering, 7 (12) (2015) 1687814015619135.

[145] A. Gontarz, L. Weiss and K. Wegener, Energy consumption measurement with a multichannel measurement system on a machine tool (2010).

[146] N. E. Huang, Z. Shen, S. R. Long, M. C. Wu, H. H. Shih, Q. Zheng, N.-C. Yen, C. C. Tung and H. H. Liu, The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis, Proceedings of the Royal Society of London A: Mathematical, Physical and Engineering Sciences (1998) 903-995.

[147] W.-S. Chu, M.-S. Kim, K.-H. Jang, J.-H. Song, H. Rodrigue, D.-M. Chun, Y. T. Cho, S. H. Ko, K.-J. Cho and S. W. Cha, From Design for manufacturing (DFM) to Manufacturing for design (MFD) via hybrid manufacturing and smart factory: A review and perspective of paradigm


shift, International Journal of Precision Engineering and Manufacturing-Green Technology, 3 (2) (2016) 209-222.

[148] H. S. Kang, J. Y. Lee, S. Choi, H. Kim, J. H. Park, J. Y. Son, B. H. Kim and S. Do Noh, Smart manufacturing: Past research, present findings, and future directions, International Journal of Precision Engineering and Manufacturing-Green Technology, 3 (1) (2016) 111-128.

[149] K. Liu, P. Zhong, Q. Zeng, D. Li and S. Li, Application modes of cloud manufacturing and program analysis, Journal of Mechanical Science and Technology, 31 (1) (2017) 157-164.

[150] M. Weyrich, J.-P. Schmidt and C. Ebert, Machine-to-machine communication, IEEE Software, 31 (4) (2014) 19-23.

[151] P. Stenumgaard, J. Chilo, J. Ferrer-Coll and P. Angskog, Challenges and conditions for wireless machine-to-machine communications in industrial environments, IEEE Communications Magazine, 51 (6) (2013) 187-192.

[152] S. Gusmeroli, S. Piccione and D. Rotondi, IoT@ Work automation middleware system design and architecture, 2012 IEEE 17th Conference on Emerging Technologies & Factory Automation (ETFA) (2012) 1-8.

[153] L. Pelusi, A. Passarella and M. Conti, Opportunistic networking: data forwarding in disconnected mobile ad hoc networks, IEEE Communications Magazine, 44 (11) (2006).

[154] A. Weiss and A. Huber, User experience of a smart factory robot: Assembly line workers demand adaptive robots, arXiv preprint arXiv:1606.03846 (2016).

[155] A. D. Orcesi and D. M. Frangopol, Optimization of bridge maintenance strategies based on structural health monitoring information, Structural Safety, 33 (1) (2011) 26-41.

[156] J. A. N. Malik, US expects energy savings through smart manufacturing, MRS Bulletin, 41 (1) (2016) 10.

[157] C. R. Farrar and N. A. Lieven, Damage prognosis: The

future of structural health monitoring, Philosophical Trans-actions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences, 365 (1851) (2007) 623-632.

Gil-Yong Lee received the B.S., M.S. and Ph.D. degrees from Seoul National University, Seoul, Korea in 2006, 2008, and 2013, respectively. After gradua-tions, he conducted his research at the University of Washington, Seattle, WA from 2013 to 2016. Currently he is a Research Assistant Professor in the In-

stitute of Advanced Machines and Design (IAMD) at the Seoul National University, Seoul, Korea. His research inter-ests are in direct printing, nanoparticle printer, rapid prototyp-ing, micro/nano fabrication, soft actuators and sensors, energy devices, composites, and acoustic metamaterials.

Sung-Hoon Ahn received the B.S. de-gree from University of Michigan, Ann Arbor, MI, USA in 1992, and M.S. and Ph.D. degrees from the Stanford Univer-sity, CA, USA in 1994 and 1997, re-spectively. Currently, he is a Professor in the Dept. of Mechanical and Aerospace Engineering and Associate

Dean in Graduate School of Engineering Practice at the Seoul National University, Seoul, Korea. His research interests in-clude 3D Printing, Smart Soft Composite Materials, mi-cro/nano-scale fabrication (Laser, focused ion beam, milling drilling, and Nano Particle Deposition System), green manu-facturing, energy device, and appropriate technology.

[invited review paper] machine health management …...formance life until complete failure [1, 3,...

Documents