Semantic Web Enabled Network of Maintenance Services for Smart Devices

1
Semantic Web Enabled Network of Maintenance
Services for Smart Devices
"OntoServ.Net"

• Industrial Ontologies Group
• Tekes Project Proposal
• http//www.cs.jyu.fi/ai/Metso_Maintenance.ppt

Agora Center, University of Jyväskylä, March 2003
2
Our Team Industrial Ontologies Group

• Head
• Vagan Terziyan
• Researchers
• Oleksandr Kononenko
• Andriy Zharko
• Oleksiy Khriyenko
• Supervisor and Consultant from Metso
• Jouni Pyotsia

vagan_at_it.jyu.fi Agora Center, University of
Jyväskylä
Industrial Ontologies Group http//www.cs.jyu.f
i/ai/OntoGroup/index.html
3
Emerging Semantic Web

• Knowledge is an important productivity factor
• However to make your knowledge to be really such
  you should consider managing it based on emerging
  Semantic Web Technology
• Then it would be possible to take better care of
  your businesses, products, services, processes,
  etc. using automatically collected and integrated
  experience from different heterogeneous
  distributed sources worldwide
• This makes possible also to make your own
  knowledge and experience reusable, shared and
  permanently beneficial

4
Enterprise Integration Technologies

• Web Service Technology (SOAP, WSDL and UDDI)
• Enterprise Integration (Enterprise Application
  Integration and E-Commerce in form of
  Business-to-Business Integration as well as
  Business-to-Consumer)
• Semantic Web Technology (ontology languages).

The promise is that Web Service Technology in
conjunction with Semantic Web Technology
(Semantic Web Services) will make Enterprise
Integration dynamically possible for all types
and sizes of enterprises compared to the
traditional technologies
5
Semantic Web

• The Semantic Web is a vision the idea of having
  data on the Web defined and linked in a way that
  it can be used by machines not just for display
  purposes, but for automation, integration and
  reuse of data across various applications.

• http//www.w3c.org/2001/SW

6
Semantic Web basics

• RDF
• is a W3C standard, which provides tool to
  describe Web resources
• provides interoperability between applications
  that exchange machine-understandable information
• RDF Schema
• is a W3C standard which defines vocabulary for
  RDF
• organizes this vocabulary in a typed hierarchy
• capable to explicitly declare semantic relations
  between vocabulary terms

7
Ontological Vision of Semantic Web

• Semantic Web needs ontologies
• An ontology is
• document or file that formally and in a
  standardized way defines the hierarchy of classes
  within the domain, semantic relations among terms
  and inference rules
• Use of ontologies
• Sharing semantics of your data across distributed
  applications

8
Knowledge Management based on Semantic Web
concepts

• A commitment to a common ontology is a guarantee
  of a
• consistency and thus possibility of data (and
  knowledge) sharing
• It seems feasible to use standards of the
  Semantic Web research community for the
  development of next-generation information
  systems based on ontology-driven knowledge
  management, e.g.
• Intelligent process automation systems
• Intelligent condition monitoring systems
• Decision support systems (embedded AI)
• Intelligent maintenance systems and services

9
Project Primer Goal

• The primer goal is to study and implement the
  benefits of the
• Semantic Web (interoperability based on
  ontological support and semantic annotations),
• Intelligent Web Services (modelling, automated
  discovery and integration), and
• (Multi)Agent technologies (agents communication,
  coordination and mobility)
• to improve the performance of the Field Device
  Management Process by launching a network of
  distributed intelligent maintenance services.

10
Pilot Implementation Goal

•  .
• More specifically the goal is to develop
• a prototype of a global intelligent diagnostics
  and maintenance support system,
• an appropriate multiagent support for it,
• ontological support for it,
• pilot prototype implementation,
• case study.

11
Global vision agents in action
Agents acting as service components in the
Maintenance Service Network have ability to learn
during work improving services performance.
New vision assumes a Maintenance Services Network
of smart-devices and Maintenance Service Centers,
in which maintenance experience is accumulated
independently by agents of each Maintenance
Center with a possibility to be integrated
together when needed. Smart-devices are becoming
users of provided maintenance services.
12
Challenge 1 Service Users are devices

• The class of service requestors is extended with
  new group of service users smart devices.
• We add semantic-enabled descriptions of services
  to facilitate
• automated discovery and use of services by
  smart-devices
• automated integration of services
• communication between heterogeneous services.

13
Types of Maintenance Service
Product based
Location based
Profile based

• We consider 3 types of Maintenance Services
• Product based all types of maintenance
  activities for specific products
• Profile based specific maintenance activities
  for wide class of products
• Location based based on a location where
  products are used
• Actually each node related to maintenance center
  may combine all of these
• three types of maintenance.

14
Distributed knowledge
Im competent in domain 1..
I have a problem from domain 23.. Who can help?
Im competent in domain 2..
Im competent in domain N..
As a result of independent maintenance experience
accumulation by service components (agents) every
Maintenance Service Center in the net provides
specific set of service components. When a
problem arises maintenance service components
with the most relevant knowledge for that case
might be found in the net.
15
Using agents
Field agents are already considered to be used in
condition monitoring. Agents are also key
resource in a new web-services framework. Our
goal is to apply agents in a maintenance system,
enabling devices and maintenance centers to
communicate and cooperate with each other
16
Internal and External Agent Platforms
Maintenance Platform Environment to run
Maintenance Services, contains a set of
expert-agents both in maintenance and
diagnostics. Agents are service components

• Service Platform
• Environment where service components perform
• Condition monitoring
• Maintenance activities

Based on the online diagnostics, a service
component-agent, selected for the specific faulty
or emergency situation, can be moved to the
service platform to help the host agent to manage
it and to carry out the predictive maintenance
activities.
17
Challenge 2 Two Types of Service Platforms

• Service Platform is an environment for running
  services and hosting service components (agents).
• Services can be provided either locally, i.e. by
  embedding them to smart-device internal platform,
  or remotely by querying them from a Web-based
  external platform.
• External service can be queried either from
  Web-based external platform or from another
  internal platform.
• External Web service platforms provide more rich
  services since they are used by many clients and
  quality of services can be permanently improved
  according to growing experience.
• Various interactions between service platforms
  (internal-internal, internal-external,
  external-external) can be organized as a P2P-like
  network.

18
Internal Platform
Recovery components
Diagnostic components
Field Agent
Management component
Wrapper component
Field Agent device-dependent embedded condition
monitoring component (e.g. FieldBrowser) Wrapper
component for integration with device-dependent
(software and hardware) resources, acts as a
semantic adaptor, mediator between
semantic-enabled and traditional parts of service
infrastructure Management components for
management of maintenance activities and
distributed resource allocation Diagnostic
components for online discovery of problems
within a device based on its state parameters and
ontology-based classification of these problems
(component is mobile agent) Recovery components
for automatic planning and performing
appropriate maintenance activities for a
discovered diagnosis (component is mobile agent).
19
External Platform
Recovery components
Diagnostic components
Management component
Management component, Diagnostic components,
Recovery components service components of
Maintenance Service Center. There is similar
service components set as in the Internal System
structure, but these components have more rich
experience and abilities to solve problems.
20
Agents in Semantic Web
3. Wait a bit, I will give you some pills
1. I feel bad, pressure more than 200, headache,
Who can advise what to do ?
Agents in Semantic Web supposed to understand
each other because they will share common
standard, platform, ontology and language
4. Never had such experience. No idea what to do
2. I think you should stop drink beer for a
while
21
The Challenge Global Understanding eNvironment
(GUN)
How to make entities from our physical world to
understand each other when necessary ?
22
GUN Concept
2. I have some pills for you
1. I feel bad, temperature 40, pain in stomach,
Who can advise what to do ?
Entities will interoperate through OntoAdapters,
which are supplements of these entities up to
Semantic Web enabled agents
23
Semantic Web Before GUN
Semantic Web Applications
Semantic Web applications understand, (re)use,
share, integrate, etc. Semantic Web resources
Semantic Web Resources
24
GUN Concept All GUN resources understand each
other
Real World Objects OntoAdapters GUN
Resources
Real World objects
GUN
OntoAdapters
25
Maintenance Services

• Organizing the maintenance
• Service 1 Remote diagnostic
• Service 2 Recovery and predictive maintenance
• Service 3 Preventive inspection
• Service 4 Emergency service
• Service 5 Human resource execution

26
Remote diagnostics scenario
parameters
parameters
diagnosis
diagnosis

• Alarm situation is locally detected however
  Internal Maintenance Platform (IMP) is not able
  to classify it as certain diagnosis. Thus IMP
  sends request with parameters to an External
  Maintenance Platform (EMP).
• As a result, EMP sends discovered diagnosis back
  to the IMP.
• If similar request for diagnosis is sent often
  enough, then it is considered to send appropriate
  diagnostic service component (mobile agent) from
  EMP, to operate locally at the IMP.

27
Challenge 3 Service Components are
Autonomous Intelligent Agents

• Service components are mobile
• Service components are able to learn
• Service components are Semantic Web enabled

28
Maintenance Service Network
Service components are certified. Certification
system is a basis for guaranteed quality of
maintenance services.
All interactions in the Maintenance Network are
performed between Management Service components

• Requirements for Management Service component
• Check of request correspondence to available
  local services, based on profile of MC.
• Request to other components of the network, in
  case if request cant be satisfied.
• Enabling peer-to-peer semantic search in the
  Maintenance Service Network

29
Service management

• Profile is a file, that contains information
  about
• what type of maintenance activities MSC
  provides
• what level of quality its gained during
  certification
• economical aspects (cost).

Since we have independent services in
distributed environment, the peep-to-peer
concept must be implied on base of Semantic Web
(profile web).
High-level functions are performed on the base of
profile processing. Each Maintenance Service
Center has a corresponding profile which
describes its services. Profile is created in
machine understandable form on a basis of common
ontology.
30
Challenge 4 Semantic P2P Concept for
Service network Management

• The concept assumes decentralized management
  architectures with.
• centralized ontologies for e.g.
• Service certification management
• Service discovery management
• Service responsibility management
• Quality of Service management
• Trust management
• Privacy and security management .
• Also transaction management issues related to
  transportation of mobile.
• components between platforms should be addressed
  in this project.
• Two levels of management are considered for
  interactions between.
• local service platforms of smart-devices (P2P
  network) and for.
• interactions between service centres on
  enterprise level.

31
Where are ontologies?
We understand
Ontology
each other!
All necessary information from Maintenance Domain
is annotated using common Maintenance Ontology,
that provides common vocabulary for all involved
agents and services. Knowledge represented in
the system is restricted by terms given in the
ontology.
32
Subdomain ontologies
The following set of subdomain ontologies can be
defined
Product
describes device structure, its components and
states (for maintenance/control processes)
ontology
Faults' Detection
describes breaks and faults classifications,
maintenance cases bindings to certain products or
components, specification of detection methods,
rules, etc.
ontology
Maintenance
describes maintenance activity classification and
prerequisites of use rules/inference tools to
use, etc.
ontology
33
Upper Maintenance Ontology
Class of Restrictions
Class of Maintenance Activities
Class of Diagnosis
Subclass-of
Restriction
Subclass-of
Subclass-of
Restricted-by
Applied-to
Maintenance Activity
Diagnosis
Procedure
Applied-to
Requires
State of product
Specification
Product
State
Resource
Standardized-by
Subclass-of
Class of Resources
Subclass-of
Standard
Class of Standards
34
Creating ontologies
Class details
Class properties (slots)
Classes hierarchy
35
RDF in XML
RDFS vocabulary
RDF description
36
Ontology of Control Valves with Protégé
37
Our project implementation goals
Design Maintenance Centers Infrastructure
Implement minimal set of maintenance service
components (agents)
Provide minimal set of necessary information
structures and ontologies
Design Pilot Service Platforms
38
Necessary data for pilot implementation
To select some product as a case for
implementation and consider different diagnostic
cases.
What is the relation maintenance activity
equipment state
What maintenance activities exist?
How equipment state is described?
How does breakage class associate with
maintenance activity
What is the relation equipment state
breakage class
What breakage classes exist?
39
Project Main Objectives

• Development of upper-ontologies for the
  maintenance domain
• Development of samples (a) an embedded
  agent-enabled platform and (b) Semantic Web
  maintenance service for smart-devices
• Development of P2P semantic search techniques in
  semantic-enabled network of maintenance services
• Pilot implementation of embedded platform and set
  of maintenance services Development of ontology
  for Metso smart-device case
• Testing of pilot system on the Metso smart-device
  case

40
Project Deliverables

• Requirements to a Maintenance Service Network for
  Smart-Devices
• Requirements to possible service components
  (agents)
• Requirements to an embedded service platform
• Requirements to ontology management in a semantic
  P2P network
• Requirements to maintenance service ontology
• Scenarios for certification, security, privacy
  and trust management
• Service platform specifications and
  implementation plan
• Upper-ontologies for smart-devices maintenance
  domain
• Devices ontology
• Diagnostics ontology
• Maintenance activities ontology
• Maintenance service ontology
• Pilot implementation of the Service Platform

41
Conclusions

• Traditional Enterprise Integration technologies
  are able to address some of maintenance
  management problems today. However, new
  technologies like Web Services Technology in
  combination with Semantic Web and Agent
  Technologies have the potential to address
  maintenance needs much better
• We have experience and human resources to develop
  the concept of Distributed Maintenance Network
  and provide implementation starting from a pilot
  system and pilot ontologies
• Results can be used by co-operating companies
  e.g. Metso for management of their field devices
  based on embedded agent platforms and Web
  services Sonera for providing communication
  infrastructure for embedded agents and launching
  appropriate Web services for this and also for
  other cases