1
2019年9月19日
鄭芳田Fan-Tien Cheng
國立成功大學智慧製造研究中心Intelligent Manufacturing Research Center (iMRC)
National Cheng Kung University
2
Outline
Introduction to Intelligent Manufacturing Research Center (iMRC)
Lead to Industry 4.1 by iMRC
Automatic Virtual Metrology (AVM)
Intelligent Factory Automation ( ) Alliance
How the Cloud-based Operates
Main Features & Benefits of
System Platform Demos
Promotion & Business Model of the System
iFA
iFA
iFA
iFA
iFA
3
智慧製造研究中心簡介
Introduction to
Intelligent Manufacturing Research Center
(iMRC)
4
The Importance of Quality and
the Visions of Industry 4.1 Industry 4.0 values productivity, but overlooked the importance of quality
Industry 4.0 stresses highly on improving the productivity of production lines, but with less emphasis on quality. This
makes it impossible for the factories to achieve the goal of zero defects. The key reason is the lack of an affordable
and practical online real-time total inspection system.
Samsung Note 7 battery defects causing over 24 billion USD of loss
Take the flaws of the Samsung Note 7 cellphone battery production process for example, while demanding high
productivity from the production line, the quality of the products is relatively neglected. According to the
estimation of Bloomberg, Samsung lost 2 billion USD of revenue and the market value of its stock depreciated about
22 billion USD.
Zero Defects and the Visions of Industry 4.1
From the example mentioned above can we understand that it’s important not to overlook the quality of products
while pursuing productivity. By integrating the intelligent services such as Automatic Virtual Metrology (AVM),
Intelligent Predictive Maintenance (IPM), and Intelligent Yield Management (IYM) into the Advanced
Manufacturing Cloud of Things (AMCoT) platform, the goal of zero defects can be achieved. This is defined as
“Industry 4.1” by professor Fan-Tien Cheng. The mission of the Intelligent Manufacturing System based on
AMCoT developed by iMRC is to realize the visions of Industry 4.1.
5
Intelligent Manufacturing System FrameworkSingle Machine
ApplicationsPro
du
ctio
n L
ine
Ap
plic
atio
ns
Service Subscription
3D PrintingSolar
Energy
AerospaceWheel Rim
TFT-LCDStretch Blow
MoldingCNC Machine
Tool
Semiconductor
...
AMCoT Platform
Manufacturing Service
Automated
Construction Scheme
(MSACS)
Big Data Cloud Platform
NoSQL DBAVR DB3D DBRDB HDFS
Co
re
Tec
hn
olo
gie
s
AVM IPM IYM AEO ROT DMT `AM
Automatic Virtual Metrology (AVM)
Intelligent Predictive Maintenance
(IPM)
Intelligent Yield Management (IYM)
Advanced Manufacturing Cloud of
Things (AMCoT)
Cyber-Physical Agent (CPA)
Manufacturing Service Automated
Construction Scheme
(MSACS)
Fan-Tien Cheng
e-Manufacturing
Research Center
Metal Powder Processing Technology
Optimization Processing Parameters for
Additive Manufacture
Optimization Metallographic/Microstructure
for Additive Manufacture
On-line Measurement Temperature and Area
of Additive Manufacture Melting Pool
On-line Measurement Cladding Height for
Additive Manufacture
Yu-Long Luo/ You-Ren Cheng
Metal 3D Additive Manufacturing Research Center
Big Data Cloud Platform
Calculation
Machine learning and
intelligent management
Nondestructive Testing and
Intelligent Internet of Things
AR/ VR Human-Machine
Interaction
Hong-Jang Hsiao/ Ming-Chi Tsai
Intelligent Big Data Integrated
Solutions Center
Process Data Sensing (PDS)
Data Driven Modeling (DMT)
Robust Optimization (ROT)
Advanced Evolution
Optimization (AEO)
Hao-Ching Yang/ Jyh-Horng Chou
Performance and Quality Robust Optimization Research
Center
CPAP CPAP CPAP CPAECPAECPAE
BA/ML AR/VR
Sy
ste
m
fram
ew
ork
DA
On Demand
6
Various Application Scenarios
7
Missions of iMRC
Our Intelligent Manufacturing Research Center is dedicated to
assist various Manufacturing Industries to realize the visions of
Industry 4.1.
Phase 1:Accomplishing the goal of having Zero Defects of
all the deliverables.
Phase 2:Accomplishing the goal of having Zero Defects of
all the products.
(Big Data Analytics & Continuous Improvement)
9
Virtual Metrology (VM) is a method to conjecture manufacturing quality of a process tool based on
data sensed from the process tool and without physical metrology operation.
it 1it 2it 3it 4it jt 1jt 2jt 3jt 4jt
UCL
LCL
Process TimeVir
tua
l M
etr
olo
gy
Da
tai
y(
)^
OOC
PHASE 5
DEPT 6
DEPT 5
DEPT 4
DEPT 3
DEPT 2
DEPT 1
PHASE 4PHASE 3PHASE 2PHASE 1
DEPLOYMENT CHART
Metrology
Equipment
Virtual Metrology
System
Sampling Products
Products
Sensor Data ijx
Real -Time & On-Line
TTransportation &
Measurement Time
T
Tti Ttj it jt
UCL
LCL
Process Time
T
Rea
l M
etr
olo
gy
Da
taiy
()it
jt
Production
Equipment
IBM
VM can convert sampling inspections with metrology delay into
real-time and on-line total inspection.
Applying Virtual Metrology in Semiconductor and
TFT-LCD Industry
10
Introduction to AVM
11
Data
PreprocessingProcess
Data
Data
Preprocessing
VM
Only for
Training
& Tuning
√√√√√√√√√√√
√Metrology
Data
Conjecture Model
Traditional VM Framework
Promptness and accuracy of traditional VM may not be achieved simultaneously. When promptnessis emphasized, accuracy is poor; and when accuracy is emphasized, promptness cannot be achieved.
Traditional VM values are provided without the reliance indexes (RIs) so users don’t know whetherVM values are reliable or not. This phenomenon is attributed to the so-calledapplicability/manufacturability problem of VM.
The traditional VM scheme is not able to perform on-line and real-time quality evaluation of process-and-metrology data collected. As such, abnormalities in process data or metrology data cannot beexcluded and will be added to the model tuning or re-training processes, resulting in deteriorated VMaccuracy.
12
AVM Framework
Conjecture Model
RI Module
GSI
RI
Process
Data
GSI Module
VMI
VMII
Only for
Training
& Tuning
√√√√√√√√√√√
Metrology
Data√
Dual-Phase
Algorithm
Data Preprocessing
DQI Z-Scorey
DQIx Z-Score
Data Preprocessing
Promptness and accuracy can both be taken into consideration in the dual-phase algorithm. Phase
I emphasizes promptness to immediately calculate and output the Phase-I VM value (VMI); Phase II
improves accuracy to re-calculate (with the newly refreshed VM models) and output the Phase-II
VM values (VMII).
The AVM system generates the accompanying reliance index (RI) of each VMI and VMII. Users can
check the reliability of the VM prediction via its corresponding RI value.
The AVM system can ensure the quality of process data and actual metrology data on-line and real-
time, thus the quality of the outputted VM values can be further assured.
RI & GSI
Conjecture AlgorithmData Quality Evaluation
13
Server-based AVM System
VM
Client
SOAP
VM
Manager
Metrology
Equipment
SOAP
VM
Server-1
Metrology
Equipment
...
...
Model
Creation
Server
Central
Database
Process
EquipmentProcess
Equipment
VM
Client
VM
Server-n
14
The purpose is to integrate the functions of AVM with those of MES. The interfaces among AVM,
other MES components, and R2R (run-to-run) modules in the novel manufacturing system are
defined so that the total quality inspection system can be realized and the R2R capability can be
migrated from lot-to-lot control to wafer-to-wafer control.
Integrating AVM with
Manufacturing Execution System (MES)
Equipment i Equipment i+1
Process
Data
Material flowInformation flow
Metrology
)25 pcs
)
)1 pc
)
MES
R2R i R2R i+1
Alarm Manager
WIP Tracking
SPC Scheduler
Equipment Manager
Material Manager Reporting
)25 pcs
)
Equipment i Equipment i+1
Process
Data
Material flowInformation flow
Metrology
AVM
)25 pcs
)
)1 pc
)
MES
R2R i R2R i+1
Alarm Manager
WIP Tracking
SPC Scheduler
Equipment Manager
Material Manager Reporting
)25 pcs
)
VMI for FB
R2R Control
VMII for FF
R2R Control
15
A key challenge preventing effective utilization of VM in R2R control is the inability to take the reliance level
in the VM feedback loop of R2R control into consideration. The reason is that adopting an unreliable VM
value may be worse than if no VM is utilized. The AVM system possesses the RI of VM to gauge the
reliability of VM results [2], [A]. Therefore, this novelty is to invent a novel scheme of R2R control that
utilizes AVM with RI/GSI in the feedback loop.
Applying AVM for W2W APC
Metrology
Tgt
Material flowInformation flow
Process
1ku
1
Product
zy
: No. of RunskZ: No. of Actural Measurements
Sampled product
k
R2R
0 1 ku
EWMA Filter
Metrology
Tgt
Material flowInformation flow
Process
AVMˆ
kykX
1ku
1
Product
Process
Data
For Training or Tuning
zy
: No. of RunskZ: No. of Actural Measurements
Sampled product
VMI or VMII
k
2 /RI GSI2 ( ,f RI
R2R
Upstream Metrology Data
0 1 ku
EWMA Filter
1)GSI
16
AVM Deployment for the
Flat Panel Display Processes
17
(a) TFT Process Production Line
Process
Data
Process
Data
Metrology
CDMetrology
DepthProcess
Data
Metrology
WidthProcess
Data
Process
Data
MESMaterial flowInformation flow
Film
Deposition
(Stage I) Exposure Developing Etching Stripping
Positive
Photoresist
Coating
Film
Deposition
(Stage II)
Metrology
THK (G1)
(a)
Substrate (Glass)
IN
GG1
G2
Substrate (Glass)Substrate (Glass)
G1Gate Layer
Photoresist
Substrate (Glass) Substrate (Glass)
UV
Mask
Substrate (Glass) Substrate (Glass)
Metrology
THK (G) Process
Data
Process
Data
Metrology
CDMetrology
DepthProcess
Data
Metrology
WidthProcess
Data
Process
Data
VMI or VMII
MES
AVM
CD
AVM
THK (G1)
AVM
THK (G2)~
Material flowInformation flow
(G1)^
VMI )
)
VMII&
Film
Deposition
(Stage I)
Exposure Developing Etching Stripping
Positive
Photoresist
Coating
AVM
DepthAVM
Width
Film
Deposition
(Stage II)
VMI )
)
VMII&VMI )
)
VMII&VMI )
)
VMII&VMI )
)
VMII&
Metrology
THK (G1)
(b)
(a)
Substrate (Glass)
IN
GG1
G2
Substrate (Glass)Substrate (Glass)
G1Gate Layer
Photoresist
Substrate (Glass) Substrate (Glass)
UV
Mask
Substrate (Glass)
+-
Substrate (Glass)
Metrology
THK (G)
~G2 = G - G1
^
(b) Deployment of AVM Servers
Semiconductor Layer of the TFT Process Flow
with Deployment of AVM Servers
33
Intelligent Factory Automation ( ) Alliance
Intelligent
Manufacturing
Intelligent Manufacturing Solution Strategic Alliance
The Best Choice of Industry 4.1
Suitable for All Kinds of Machine and Product Systems
iFA
34
Server-based FrameworkiFA
35
Cloud-based FrameworkiFA
40
How Does the Cloud-based OperateiFA
41
Main Features & Benefits of
Achieving ZERO DEFECTs of All Products for Industry 4.1
• Accomplishing Big-Data-Collection Infrastructure
[w/ CPA – Internet of Things and Edge computing]
• Accomplishing Cloud-Computing Infrastructure
[w/ AMCoT – Cloud Manufacturing]
• Establishing internet security and monitoring system
[w/ IoT Security Operation System – Information Security]
• Managing the utilization of the factory and monitoring the Overall Equipment
Effectiveness (OEE)
[w/ EMS – Equipment management ]
• Accomplishing Workpieces Total Quality Inspection
[w/ AVM – Total Quality Inspection for Achieving Zero Defects]
• Improving Production-Tools’ Availability
[w/ IPM – RUL Prediction Preventing Unscheduled Down]
• Real-time monitoring the tool status and provide the adequate timing for tool
replacing
[w/TLM – RUL Prediction of the tool status ]
• Constructing Root-Causes Searching Infrastructure of Yield Loss
[w/ IYM – Root-Causes Identification of Yield Loss for Continuous Improvement]
iFA
42
System Platform Demos
Bottle Industry
PET Stretch-Blow Molding Machine(2018.08.15 2018 TaipeiPlas CHUM POWER)
Semiconductor Industry
Bumping Process(2018.08.01 ASE Group)
Automobile Industry
Wheel Machining(2018.05.09 2018 iMTDuo FEMCO)
iFA
43
System Platform Demos
Semiconductor Industry
Bumping Process
(2018.08.01 ASE Group)
iFA
44
Bumping Process
Exposure Cu Plating
Positive Photoresist
Coating
Sputtering Deposition
Stripping Developing
ProcessData
ProcessData
ProcessData
Etching
ProcessData
ProcessData THKCu
THKTi
CD
CD
THK
Delivery after Total Inspection
ReflowFlux
Clean
MetrologyExterior
Defects Check
MetrologyBall Height/
DiameterMetrologyBall Sheer
ProcessData
UV
UBM
Material flowInformation flow
SubstrateCutout
Dipositive Protection Layer
(PB01)In
Come Wafer
RDL
SubstrateCutout
Exposure Cu Plating
Positive Photoresist
Coating
Sputtering Deposition
Stripping Developing
ProcessData
ProcessData
ProcessData
Etching
ProcessData
ProcessData THKCu
THKTi
CD
CD
THK
UV
Dipositive
Protection Layer
(PB02)
SubstrateCutout
Ball mount
ProcessData
RDL: Re-distribution Layer
UBM: Under Bump Metallurgy
45
AMCoT Framework &Intelligent Manufacturing Deployment Procedure
CPA CPA CPA CPA CPA CPA CPA CPACPACPA CPACPA
Material flow
Information flow
Take UMP as example
Exposure Cu
Plating
Positive
Photoresist
Coating
Sputtering
Deposition Stripping Developing Etching
THKCu THKTi CD CD THK
Delivery after
Total InspectionReflow
Flux
Clean
Metrology
Exterior Defects
Check
Metrology
Ball
Height/DiameterMetrology
Ball Sheer
Ball
mount
UV Mask
Advanced Manufacturing Cloud of Things
Equipment
Prognosis
Predictive
Maintenance
Simulation
IYM
...Model
Creation
Virtual
Metrology
AVM
...
R2R
Yield
EnhancementYield
Management
IPM
Control
AlgorithmStrategy...
Service Broker
...
Cloud of Things Services ( SOAP / REST )
Big Data Analytics Apps
HDFSRDB
Manufacturing Services Automated Construction
Scheme (MSACS)
46
【Cloud-based AVM】Real-time prediction of the product
processing quality
• Collect the process & metrology data from the tools via CPA
• Upload to cloud-based AVM service to predict the product quality,
and then send notification to GUI
• Users can check the product processing quality in real time
47
AVM System Display Framework
48
Cloud-based AVM Demo for CD of Photo Process (ASE)
52
【Cloud based IPM Service】Predict the device health status and
service life
• Collect the process data from the target device via CPA
• Upload to cloud-based IPM service to build the remaining useful life (RUL)
prediction model, and then download the RUL model to the CPA for
monitoring the device health status and predicting the RUL
• Users can check the device health status and remaining useful life (RUL) in
real time
53
SOAP
……
C&H
Creation
Server
(CCS)
IPM
Manager
IPM
Client
……
Central
DB
IPM
Server nCA
Process
Equipment n
IPM
Server 1
CPA
Device 1 Device 2 Device n
……
Process
Data
Each Device
Related
Process Data
CA
CPA
BPM
CPA
BPM
Process
Equipment 1
BPM
CPA
Device 1 Device 2 Device n
……
Process
Data
Each Device
Related
Process Data
CPA
BPM
CPA
BPMBPM
SOAPPrivate Cloud
Private Cloud
Cloud-based IPM Service (1/2)
Intelligent Predictive Maintenance (IPM) System:
Systematic management flow via 3-layered framework
Cyber-Physical Agent (CPA) is pluggable, users can select different predictive maintenance
algorithm according to needs, and establish Predictive Maintenance Algorithms Library for
implementing different types of target devices.
55
IPM System Display Framework
60
【Cloud-based IYM System】Find out the key devices and
parameters which affects the yield
Phase I: Search for the key devices which affects the yield
Phase II: Search for the key parameters which affects the yield
• Finding out the root cause which affect the yield via the 2 Phase
process effectively
• Reduce the trouble shooting time and improve the yield
63
KSA Input Data
Take Bumping Process as an Example
Exposure Cu Plating
Positive Photoresist
Coating
Sputtering Deposition
Stripping Developing
ProcessData
ProcessData
ProcessData
Etching
ProcessData
ProcessData THKCu
THKTi
CD
CD
THK
Delivery after Total Inspection
ReflowFlux
Clean
MetrologyExterior
Defects Check
MetrologyBall Height/
DiameterMetrologyBall Sheer
ProcessData
UV
UBM
Material flowInformation flow
SubstrateCutout
Dipositive Protection Layer
(PB01)In
Come Wafer
RDL
SubstrateCutout
Exposure Cu Plating
Positive Photoresist
Coating
Sputtering Deposition
Stripping Developing
ProcessData
ProcessData
ProcessData
Etching
ProcessData
ProcessData THKCu
THKTi
CD
CD
THK
UV
Dipositive
Protection Layer
(PB02)
SubstrateCutout
Ball mount
ProcessData
Final Inspetion (Y)Inline Metrology (y)
Production Route
(XR)
Process Data (XP)
WaferID:
AB0001
65
IYM System Operation
78
System Platform Demos
Automobile Industry
Wheel Machining
(2018.05.09 2018 iMTDuo FEMCO)
iFA
80
2018 iMTDuo System Demo --Wheel Machining (FEMCO)
iFA
81
Prof. Fan-Tien Cheng Presenting FEMCO’s AMCoT System to President Tsai, Ying-Wen at 2018 iMTduo (2018.5.9) (1/2)
82
System Platform DemosiFA
Bottle Industry
PET Stretch-Blow Molding
(2018.08.15 2018 TaipeiPlas CHUM POWER)
84
iFA2018 TaipeiPlas System Demo --PET Stretch-Blow Molding
85
Vice President Chen, Chien-Jen Visited Chum Power AMCoT at 2018 TaipeiPlas (2018.8.15)
86
TAIPEI PLAS 2018 – Smart Machinery Award
2018.08.22 經濟日報
87
Promotion & Business Model of
The SystemiFA
88
Server-based System Cloud-based SystemiFA iFA
89
Business Model
90
Intelligent Manufacturing Cloud
Promotion and Operation
Mission Task Item Company in Charge
Marketing &
Promotion
Categorization of Manufacturing
Industries for Promotion
Sales Support
System
Planning &
Implementation
IoT platform, Internet, Data
Security, and CPA
System Integration, Monitoring
Facilities, MMI, etc.
System
Warranty &
Maintenance
Tier-1 Maintenance
(Call or On-Site Service)
Tier-2 Maintenance
Tier-3 Maintenance
iFA
91
Invitation
Welcome to the joint booth of iMRC & FS-Tech at K3071
to learn more about , AVM, IPM, IYM, and FDC!
Stage
We are here
Entrance