Embed Size (px)
Citation preview
Duncan Lee, Principal Engineer, [email protected] Manufacturing IT – Factory Automation28th Mar 2019
Legal notices
This presentation is for informational purposes only. INTEL MAKES NO WARRANTIES, EXPRESS OR IMPLIED, IN THIS SUMMARY.
For more complete information about performance and benchmark results, visit www.intel.com/benchmarks
Intel and the Intel logo are trademarks of Intel Corporation in the U.S. and/or other countries.
* Other names and brands may be claimed as the property of others.
Copyright © 2019, Intel Corporation. All rights reserved.
2
Intel factory history and Today
Digital Transformation in Intel Manufacturing
Standardization Enhance Digital Transformation
IoT case studies
Leveraging data: process/ equipment source of defect identification
Leveraging images: process/ equipment source of defect identification
Summary
Agenda
1234567
3
4
Intel Factory Functions
Wafer Fab
Wafer Test
Die Prep
Packaging
Final Test
[[
]]
Intel Factory History
80’s Factory ▪ No robotic material
transport
▪ Run cards on wafer boxes
▪ Basic equipment standards
▪ Initial equipment control
▪ Initial manufacturing execution solution
5
90’s Factory ▪ Beginning robotic
material transport (FAB)
▪ Automated statistical process control
▪ Improved equipment control
▪ Improved inventory control and tracking
▪ Improved equipment automation standards
▪ Integrated manufacturing execution solutions
▪ Planning and supply chain integration
▪ Improved decision-making systems
Intel Factory Today
6
Today’s Factory ▪ Pervasive robotic material transport (FAB)
▪ Pervasive equipment standardization
▪ Advanced manufacturing execution solutions
▪ Real-time excursion control and excursion notification.
▪ Advanced process control and adjustment
▪ Predictive and adaptive maintenance
▪ Advanced inventory control and tracking
▪ Advanced rapid decision making
▪ World-class supply chain capabilities
▪ Big data repositories
▪ Greater Intel Atom®, Intel® Core™, Intel® Optane™ and Intel® Xeon® integration with factory tools .
7
Intel Manufacturing’s Digital TransformationA
uto
ma
tio
n C
ap
ab
ilit
y
Optimize
“Things” and networks are monitored, secure
and managed
Analyze, expose, and manage data to provide
business insights
Level 1-3
Level 6-8
Level 4-5
MESWIP and Equipment
Tracking
Tool Controllers
L5: Automated Tool Control
Paper Tracking
Advanced Dispatching
L6: Continuous Queue Processing
L7: Tool SelectionL8: Lot Selection
Pervasive AMHS
L4: Point to point
Remote Operations
Center
Advanced Process Control
Leverage Data Explosion
Advanced Analytics5B+ Sensor Data Points
Auto Quality Monitoring
Stop ProductionAuto Run Monitor Wafers
Return to Production
ProcessImprovements
Impossible without IT
Manual
Unit Level Traceability
CybersecurityData
Warehouse
1980’s 1990’s 2000’s 2010’s
Advanced TestingIT enabled, highly parallel
testing
Connect
Secure & Manage
Analyze & Expose
Predict
8
Standardization Enhances Digital Transformation
Wafer Fabrication
2) Wafer Test3) Die Prep4) Packaging5) Final test
Process Cycle Time
1) Fab
Standard Dual Data Center, IT Hardware, OS Suite, Tool Interface
Standard IT Support Hub, Process, Model and Metrics
Standard Architecture , Framework, Applications
Standard Material Handling
Standard Architecture and Framework, Apps
Central Planning
9
Intel Factory Automation System
OLTP = On Line Transaction ProcessingDSS = Decision Support SystemRFID = Radio Frequency Identification
Desktop PC
MiddlewareCluster
Factory Application Clusters
Manufacturing Execution
Statistical Process Control
Excursion Protection
Data Storage
OLTP DB
DSS DBsData
Replication
Web Cluster
Material Handling Control
Reporting & Offline Analytics
Production Tool
Industry Standard Protocols
Temperature, Pressure, Vacuum,
Carrier ID, Alarms, Etc....
RFID Read forLocation and
Product Check
Automated Material Handling
Yield Analysis
Advanced Process Control
Gateways to Supplement
10
Remote Operations Center
Pervasive “sameness” enables tremendous efficiencies,
like the Intel ROC
Why are we still interested in the Internet of Things and A.I. ?
Regardless of the industry, we all share these same challenges and goals.
With billions invested in semiconductor process equipment, the Internet of Things (IoT) and data
analytics( traditional and A.I.) is leveraged to:
11
Reduce Capital Cost
IncreaseQuality
Improve Time to market
Monitor environmental conditions (i.e., temperature,
humidity, power usage) at a machine
Use of IoT in the factory
Enable older, unconnected, and/or stand-
alone tools to be connected and
smarter
Enable external sensors to be
connected where needed
Tap into internal machine sensors
and circuits
Detect defects via image analytics
12
▪ Eliminate the need for humans to manually monitor and collect readings
▪ Reduce unexpected tool downtime when chemical empties faster than expected or chemical temperature fluctuates unexpectedly
13
Enable older, unconnected stand-alone toolsChemcab: Monitor chemical use and temperature
%
Intel® IoT Gateway
Chemical Flow
TouchlessOCR Reader
USB to Gateway
Analytics Consumption
Devices
WirelessAccess Point
Processing, Storage, and
Analytics
Wireless Access PointWireless Access Point
Relative Humidity of
Chamber
Wirelessly to Gateway
Ambient Temperature of Chamber
USB to Gateway
Flow Temperature
USB to Gateway
Weight of Chemicals
USB to Gateway
12 KG
Legacy Industrial Support Equipment with Sensors IoT System
14
Tap into internal machine sensors and circuitsPreventive detection of potential oven coil failure
Intel® Architecture-based Gateway
Heating Oven
USB Hub
USBI/O 1
USBI/O 5
USB I/O 2
USB I/O 3
USB I/O 4
12
6
3
7
45
2829
33
30
34
3132
Solid State Relay 1Solid State Relay 2
Solid State Relay 6
Solid State Relay 3
Solid State Relay 7
Solid State Relay 4Solid State Relay 5
Solid State Relay 28Solid State Relay 29
Solid State Relay 33
Solid State Relay 30
Solid State Relay 34
Solid State Relay 31Solid State Relay 32
He
ate
r C
oil
s 1
-34
▪ Predict potential oven coil failure by analyzing the energy usage through monitoring relay performance
▪ This enables engineers to fix the faulty coils before they are further processed, causing a product reject downstream
IoT System
Intel® Secure Network
Intel® Architecture-based Server and Storage
Predictive maintenance: Identifying exceptions
15
▪ Keep track of all the equipment and entities that process a product.
▪ Detection of defect signals is followed by commonality analysis to identify the potentially faulty equipment.
16
Tracing causes of defect via structured data (numbers)
E1 E2
E4
E5 E6 E7Test/Metrology Equipment
E3
Defect signals collected by E7. Then, commonality analysis performed, resulting in identification of the faulty equipment as E5, E6 or E7.
Defect occurred
on E2
E5 sends Defect signals
17
Tracing causes of defect via images collected at the equipment
E1 E2
E4
E5 E6 E7
E3Detect & DispositionDetect & Disposition
▪ Some equipment is equipped with cameras that take images for positioning and alignment purposes.
▪ These images can be easily repurposed for defect detection (i.e., stains) using image analysis with Python* or other programming language. In several instances, A.I. are used.
*Other names and brands may be claimed as the property of others.
Data analytics
Different Scenario, Different Analytics, Lots of Data Needed
D2D1IN
OU
T
Material
OU
T
IN
Material
DX
IN
OU
T
Material
D1 DX
Data analysis types▪ Various data▪ Within a machine▪ Between machines within same family▪ Between different types of machines
Scenario types▪ Predictive maintenance▪ Product quality
Data Types▪ Equipment▪ Process▪ Quality▪ Human
Key learning: Data Fuels Digital Transformation
Data First Trace the equipment Power of AI + StatisticsData fuels quality assurance and
predictive maintenance.
Mastering data starts from digitization and computerization of
the factory via extensive standardization.
Track all the material, process, machines and components that
manufacture the product.
Downstream defect signals can help pinpoint the defect source
equipment/component from upstream processes.
Unstructured images analysis enables defect identification and
predictive maintenance.
The combination of conventional image processing and artificial
intelligence plus statistics tremendously enhances defect
sources identification.
!
19
IT@INTEL: Sharing Intel IT Best Practices With the World
Learn more about Intel IT’s initiatives at: www.intel.com/IT20
