20.09.2012
Cyber-Physical Systems in Factory Automation Towards the 4th Industrial Revolution Dr.-Ing. Jochen Schlick Deputy Head of Department Innovative Factory Systems German Research Center for Artificial Intelligence, DFKI GmbH Scientific Coordinator Technology Initiative SmartFactoryKL e.V.
© Dr. Jochen Schlick, DFKI-IFS 2012-1
The Technology Initiative SmartFactoryKL e.V.
ULBS Sibiu
© Dr. Jochen Schlick, DFKI-IFS 2012-2
1
20.09.2012
Cyber-Physical Systems in Factory Automation
• Cyber-Physical Systems in the Domain of Production Automation • Demands to Factory Automation • Existing Approaches and Open Issues • Distributed Automation • Open Communication • Semantic Interoperability • Towards the 4th Industrial Revolution
© Dr. Jochen Schlick, DFKI-IFS 2012-3
Definitions of Cyber-Physical Systems From the Point of view of Informatics “Cyber-Physical Systems (CPS) are integrations of computation and physical processes. Embedded computers and networks monitor and control the physical processes, usually with feedback loops where physical processes affect computations and vice versa.” Lee, E.: Cyber Physical Systems: Design Challenges. Technical Report. Berkeley: University of California, 2008.
From the Point of view of Automation Technology “CPS are engineered systems whose operations are monitored and controlled by a computing and communication core embedded in objects and structures in the physical environment.” Karl Henrik Johansson, 2011
© Dr. Jochen Schlick, DFKI-IFS 2012-4
2
20.09.2012
Cyber-Physical Systems – The Internet of Things
Source: Youtube.com; TED Merrill, Shiftables
© Dr. Jochen Schlick, DFKI-IFS 2012-5
Cyber-Physical Systems – Smart and Communicating Objects
Source: Youtube.com; TED Merrill, Shiftables
© Dr. Jochen Schlick, DFKI-IFS 2012-6
3
20.09.2012
CPS in Production Automation – Main Topics
• Adaptivity and Autonomy – –
Component-based automation Deployment time configuration
• Communication and Distributed Functionality – – –
No strong adherence to hierarchical communication architectures Horizontal and vertical integration Wireless and cable-bound ad-hoc communication based on IP
• Context Aware Automation – – –
Components work independent of location Components can realize their role in the overall production line Dynamic adaption of process parameters and process structures according to environmental influences
• Integration of the Human Factor – –
Augmented operator Human as orchestrator, optimizer and conductor
© Dr. Jochen Schlick, DFKI-IFS 2012-7
Cyber-Physical Systems in Factory Automation
• Cyber-Physical Systems in the Domain of Production Automation • Demands to Factory Automation • Existing Approaches and Open Issues • Distributed Automation • Open Communication • Semantic Interoperability • Towards the 4th Industrial Revolution
© Dr. Jochen Schlick, DFKI-IFS 2012-8
4
20.09.2012
Challenges to Industrial Manufacturing
Volatile Markets Increasing Relevance of Value Networks
Product Individualization
Shortening Product Life Cycles
Rising Product Variety
© Dr. Jochen Schlick, DFKI-IFS 2012-9
Challenges to Factory Automation in the Lifecycle of Equipment Product Design Production Process Planning
Mechanical Engineering Putting into Service
Lifecycle Set Up
Control Engineering
Setting Up and Integration
© Dr. Jochen Schlick, DFKI-IFS 2012-10
5
20.09.2012
Maximum Productivity Lifecycle of Equipment
Product Design Minimal Engineering Time
Production
Continuous Process Planning Process Oriented Engineering Plug&Play Integration Mechanical Engineering
Production - Maximum Availability Putting into - Maximum Service Transparency - Lot Size 1
Lifecycle Control Engineering Setting Up and Integration © Dr. Jochen Schlick, DFKI-IFS 2012-11
Vision: from Cyber-Physical Systems to Smart Factories
Aggregation Level
smart factories
smart machines
smart objects
© Dr. Jochen Schlick, DFKI-IFS 2012-12
6
20.09.2012
Cyber-Physical Systems in Factory Automation
• Cyber-Physical Systems in the Domain of Production Automation • Demands to Factory Automation • Existing Approaches and Open Issues • Distributed Automation • Open Communication • Semantic Interoperability • Towards the 4th Industrial Revolution
© Dr. Jochen Schlick, DFKI-IFS 2012-13
Cyber-Physical Systems in Factory Automation
• Cyber-Physical Systems in the Domain of Production Automation • Demands to Factory Automation • Existing Approaches and Open Issues • Distributed Automation • Open Communication • Semantic Interoperability • Towards the 4th Industrial Revolution
© Dr. Jochen Schlick, DFKI-IFS 2012-14
7
20.09.2012
State of the Art from the Perspective of Automation Technology
Distributed Automation Coupling Sensors and Actuators via Fieldbuses • Data-collection of spatially distributed sites • Central PLC takes over the processing Component-Integrative Architectures • •
Simplifying the integration of PLCs in networked systems Example Profinet CBA
IEC 61499 •
• •
Development of control intelligence as decentralized software in embedded systems Executable specification of a distributed automation system Conceptual design to IEC 61131
© Dr. Jochen Schlick, DFKI-IFS 2012-15
Impulses from Information Technology Agent Service Service
Agent Service
Service Service
Agent
Service Agent
Service-Oriented Architectures
Agent Systems
• Modeling of business processes
• Concept from the research on distributed AI
• Encapsulation of methods and application software as reusable services
• Assignment of clearly defined sub-tasks to an agent
• Integration and invocation of services independent of platform and implementation
• High importance of communication among the various agents
• Service discovery
• Self-organization for solving complex tasks
• Late binding
Ggf. Animiere Service/Agent Mechatronic D Component er
© Dr. Jochen Schlick, DFKI-IFS 2012-16
8
20.09.2012
Multi Agent Approach for the Control of Factory Systems High Business Level Supervisor Process
Supervisor MES
Supervisor MES
Supervisor MES
Agent
Agent
Agent
Agent
Agent
Agent
Agent Agent Agent Agent
Agent
© Dr. Jochen Schlick, DFKI-IFS 2012-17
Distributed Intelligent Paradigms Level of influence Between Organizations
Holonic Manufacturing Systems
Internet of Things Within Organizations
Multi Agent System
Within Operation Boundaries Individual Element (machine, product, etc.) [Reference: McFarlane, D.C.: Distributed Intelligence in Manufacturing & Service Environments, Presentation at SOHOMA Workshop, Paris, 06/2011]
RFID
Data Capturing
SOA
Smart Object
Modeling and Insight
Information Management
Intelligent Product
Decision Support
Influencing/ Controlling
Task level
© Dr. Jochen Schlick, DFKI-IFS 2012-18
9
20.09.2012
Service-Oriented Control Architectures MES/ERP system
Web Service
Properties • Functional encapsulation of mechatronic functions in basic services • Clear defined service interfaces (e.g. described with WSDL) • Standardized communication interfaces (e.g. SOAP,DPWS)
Service d1
Service invocation with SOAPWebservices
Service e
Process control
Service d3
Application benefits • Interoperability of automation components • Minimal integration effort • Process control programs are easy to customize and change
Web Service
Service d2
Service a
Service b
Service c
© Dr. Jochen Schlick, DFKI-IFS 2012-19
SoA-AT Demonstration system (1) Objective of the mobile Demonstrator from the SmartFactoryKL: • Implementation of a service-oriented control architecture at the field device and control level • Evaluation of standard Web technologies in automation technology
© Dr. Jochen Schlick, DFKI-IFS 2012-20
10
20.09.2012
SoA-AT Demonstration system (2) Hardware and service structure of the „quality control“ process
Demonstrator characteristics: • The services concerning the quality control are implemented based on „Service-Gateways“ • The services concerning the filling process are implemented on a PC
Services: stopper: hold, release, check carrier µC
• Testing and evaluation of different service-technologies (e.g. SOAP-Web Services, DPWS, UPnP) • Implementation of the control logic with different technologies (e.g. BPEL, Statechart/Matlab, JGrafchart, OWL-S)
• Implementation of the service discovery with Semantic Web technologies (e.g. OWL, SAWSDL)
Services: sensor: check RFID: read, write µC
Gateway
Services: camera: count pills
Gateway
Inductive Stopper sensor Stopping unit
Ultrasonic sensor
µC Gateway
RFID-WriteRead-Device
Camera system
Ethernet Bits and bytes
© Dr. Jochen Schlick, DFKI-IFS 2012-21
Research Topics within Service-Oriented Control Architectures
Linking the factory planning with the principles of service-oriented control architectures for implementing an integrated process-oriented planning of automated systems Concept view Challenge
Business process modeling
Skills
Business Top-Down Approach
Application
Finance
abstract service level
Businessprocess
specific service level
Technological Bottom-Up Approach
Performance & Scalability Standards
Infrastructure Service Design
Safety
Detail View
© Dr. Jochen Schlick, DFKI-IFS 2012-22
11
20.09.2012
Cyber-Physical Systems in Factory Automation
• Cyber-Physical Systems in the Domain of Production Automation • Demands to Factory Automation • Existing Approaches and Open Issues • Distributed Automation • Open Communication • Semantic Interoperability • Towards the 4th Industrial Revolution
© Dr. Jochen Schlick, DFKI-IFS 2012-23
IT Infrastructure – Areas of Automation and IT Networks Major Application Areas of Automation •
Motion Control & Safety Systems – Hard real-time requirements – Challenge Development of distributed systems that meet hard real-time requirements
•
Factory Automation – Typically no hard-real time requirements – Challenge Mastering heterogeneous data transmission on one bus system
•
Batch and Continuous Flow Processes – Timing requirements typically not in major scope – Challenge Gaining transparency with a large number of network participants (>2000)
DFKI Study: Requirements of Production Processes for Automation Networks, 2010 EU project IoT@Work, 2011
© Dr. Jochen Schlick, DFKI-IFS 2012-24
12
20.09.2012
Extension of Wired Industrial Communication by Wireless Networks … Field Device Field Device PLC 1 2
Field Device n+1
Trend to hybrid networks: Extension of the wired networks by wireless segments Challenges of wireless networks
Opportunities of wireless networks
•
• • •
Wireless and hybrid networks differ significantly from traditional automation networks Wireless networks are never 100% reliable Real time communication underlies restrictions
• •
Mobility of systems and users Ad-hoc communication Ad-hoc configuration of field devices
© Dr. Jochen Schlick, DFKI-IFS 2012-25
Implications for the Development of Wireless Automation Solutions • Main question: How should the wireless automation solutions be designed, to enhance the wireless technology potential benefits of mobility and cable replacement?
Architecture
• • •
Components
Requirements
•
Minimal communication over the network Open standards and interfaces Semantic data description
• • •
Decentralized intelligent components Local decision process Model-based and semantic description Autonomy & Autarky
© Dr. Jochen Schlick, DFKI-IFS 2012-26
13
20.09.2012
Meaning of Interoperability concerning the Vertical Integration
Semantic interoperability
WLAN
Service repository UDDI
ERP-Level Corporate management level IP 183.77.19.0
Ethernet Communication layer
Parallel systems
integration VerticalVertical integration means - Quality control
MES-Level - Data collection • Establishing a data link between thecontrol business the automated - Lot-tracking room process software Control and cockpit level IP - (Remote) Diagnostics equipment 183.77.xx.x - Maintenance -…
EtherCAT
Syntactic • Enhancing signals with an explicit meaning interoperability
Communication layer
Control Level
• Mastering different packet sizes and period times
PLC, NC EtherCAT, Sercos III
UMTS
Interoperability signal level
Communication layer
power
Field Level Sensor-Actuator
UMTS
Logistics Processes Maintenance
Manufacturing
© Dr. Jochen Schlick, DFKI-IFS 2012-28
Implementation of the Vertical Integration with OPC-UA WLAN
Service repository UDDI
ERP-Level Corporate management level IP 183.77.19.0
Vertical integration within existing networks
MES-Level Control cockpit
control room IP 183.77.xx.x
Standard Middleware e.g. OPC-UA
Control Level PLC, NC
power
UMTS
Field Level Sensor-Actuator
UMTS
Logistic Processes Maintenance
Manufacturing
© Dr. Jochen Schlick, DFKI-IFS 2012-29
14
20.09.2012
Open-questions regarding the Integration of Data and Functions
Vertical integration
WLAN
Service repository UDDI
ERP-Level Corporate management level
IP 183.77.19.0 – Integration allows random access for all information generated in the factory
MES-Level
control room Control cockpit – But: Dynamic compilation of information as a basis for decision IP 183.77.xx.x requires knowledge of theStandard underlying technical system
Middleware e.g. OPC-UA
Control Level PLC, NC
Which are the necessary elements of a generic factory-information model, that explicitly represent the available knowledge? power
Field Level
UMTS
Sensor-Actuator UMTS
Logistic Processes Maintenance
Manufacturing
© Dr. Jochen Schlick, DFKI-IFS 2012-30
IT Security leaks in today‘s Automation Technology PC
PC
PC
PC PC PC
PC
PC PC PC
(PC)
(PC) PC
(PC)
(PC)
Scenario 2020 Malware Caused Downtime is the No 1 Loss Factor in Future Factories
[Reference: DIN EN 62443-2-1 (ISA 996), p. 65, 08/2009]
© Dr. Jochen Schlick, DFKI-IFS 2012-31
15
20.09.2012
Cyber-Physical Systems in Factory Automation
• Cyber-Physical Systems in the Domain of Production Automation • Demands to Factory Automation • Existing Approaches and Open Issues • Distributed Automation • Open Communication • Semantic Interoperability • Towards the 4th Industrial Revolution
© Dr. Jochen Schlick, DFKI-IFS 2012-32
Application Example: The Product as an Automation Component The product stores the abstract description of its production process in its product memory. It orchestrates the individual production process proactively
Individual filling
Individual assembly
Product data are for the client available
Quality control
Logistic chain monitoring
© Dr. Jochen Schlick, DFKI-IFS 2012-33
16
20.09.2012
Execution of an abstract described process in the product memory (1) Main question: How will the invoked functions be described?
Syntactic Approach without Service-Repository Service-Name, Service-Parameter and Address must emerge from the products’ memory: Example: Invocation Stopper-Services Host: 192.168.178.30:10221 Service-Name: /StopperService Operation: hold Parameter: none Disadvantages • Addresses, ports, service names, operations and parameters restrict flexibility • Cross-manufacturer orchestration practically impossible
© Dr. Jochen Schlick, DFKI-IFS 2012-34
Execution of an abstract described process in the product memory (2) Syntactic Approach with Service-Repository The service name is used to find the WSDL service definition from the service repository Example: 1) Request for the „StopperService“ service -> Returns the WSDL-service definition 2) Reading the host address, the possible operations, the parameter name, etc. from the WSDL file 3) The human selects the appropriate operations and parameterizes the service call Disadvantages • WSDL is purely a syntactic definition of the interface • Operation meaning is not formal described • Cross-vendor automated orchestration practically impossible
© Dr. Jochen Schlick, DFKI-IFS 2012-35
17
20.09.2012
Semantic Service Discovery for Production Processes Main tasks: search, locate, select, and orchestrate the appropriate functionalities Service-Ontology Semantic approach Definition of a service ontology -> i.e. Description of a service by including • Device Information • Functional information • Context information
Manufacturer Device type
Serial number Device information
Service Discovery with semantic Technologies • Search for services that directly meet the requirements of the service description • Search for services that have equivalent properties to that of the requirements
Service
Functional information
Context information
Operation
Device location
Has_Subclass Is_descriebed_through
© Dr. Jochen Schlick, DFKI-IFS 2012-36
Vision: Dynamic Orchestration as a basis for Context-Sensitive Automation Delivery Resource conservation Reliability
X Abstract process description
Conveyor1.transport (fullSpeed) (lowSpeed)
Pick&Place.insertBottom (AssemblyPlace3) (AssemblyPlace1) (AssemblyPlace2)
Pick&Place.insertBoard (AssemblyPlace1) (AssemblyPlace3) (AssemblyPlace2)
Pick&Place.insertCap (AssemblyPlace2) (AssemblyPlace1) (AssemblyPlace3)
AssemblyPlace2.compress AssemblyPlace3.compress AssemblyPlace1.compress
Camera.controlQuality
© Dr. Jochen Schlick, DFKI-IFS 2012-37
18
20.09.2012
Questions - Semantic Technologies and Context-Sensitive Automation Semantic Service Discovery •
Which are the necessary service properties for an unambiguous description regarding the Service Discovery?
Manufacturer Device type
Serial number Device information
Context-sensitive automation and dynamic orchestration •
System architecture based on existing standards and concepts
•
Evaluation criteria regarding the quality and resilience of dynamically orchestrated process chains Conveyor1.transport (lowSpeed)
Pick&Place.insertBottom (AssemblyPlace3)
Pick&Place.insertBoard (AssemblyPlace3)
Pick&Place.insertCap (AssemblyPlace3)
Service
Functional information
Context information
Operation
Device location
AssemblyPlace3.compress
© Dr. Jochen Schlick, DFKI-IFS 2012-38
Cyber-Physical Systems in Factory Automation
• Cyber-Physical Systems in the Domain of Production Automation • Demands to Factory Automation • Existing Approaches and Open Issues • Distributed Automation • Open Communication • Semantic Interoperability • Towards the 4th Industrial Revolution
© Dr. Jochen Schlick, DFKI-IFS 2012-39
19
20.09.2012
The Use of Cyber-Physical Production Systems leads to Smart Factories Borders dissipation - Global planning, local production - Situation-based extension of capacities - Synergies usage
Master surprises - Optimum production - Complete transparency of operating conditions and product status
Virtual Production
Knowledge storage
Global Facility
-
Knows history/ Current State Communicates with the environment Knows ways to achieve its goal
Smart Product
Smart Factory Social Machine
Augmented Operator Reduction of limitations
Knowledge sharing
- Allows „Telepresence“ - Makes knowledge omnipresent - Error avoidance
- Compensation for failure - Balancing of loads - Flexible adaptation
© Dr. Jochen Schlick, DFKI-IFS 2012-40
Towards the Cyber-Physical Production System Productivity
Cyber-Physical Production System Computer-Integrated Manufacturing Taylorism Industrial Production
Time
© Dr. Jochen Schlick, DFKI-IFS 2012-41
20
20.09.2012
Closing remarks Cyber Physical Production Systems leading to Smart Factories have the potential to trigger the 4th Industrial Revolution. Automation Technology has the approaches to be part of this revolution. However, Automation Technology has to open up to the points of view of IT. Automation Technology should switch its research focus from the incremental local improvement within the separated research disciplines, like e.g. distributed systems or communication technology towards a holistic systemic re-engineering. Instead of having isolated closed boxes we should focus on open systems and on smart integration of computation with their physical representation!
© Dr. Jochen Schlick, DFKI-IFS 2012-42
Questions or comments
THANK YOU
© Dr. Jochen Schlick, DFKI-IFS 2012-43
21
