Mika Viinikkala, Veli-Pekka Jaakkola, Seppo Kuikka
Institute of Automation and Control, Tampere University of Technology, P.O.Box 692, 33101 Tampere, Finland
Keywords: Web service, operation and maintenance, O&M, service oriented architecture, SOA
Abstract: Efficient information manag
ement is needed at industrial manufacturing plants that compete in the present
demanding business environment. Requirements to enhance operation and maintenance (O&M) information
management emerge from problems within internal information flows of a plant, supporting the networked
organization of O&M, and accomplishing the new demand-driven business model. O&M information
management of an industrial process plant is here proposed to be enhanced by value added web services. A
service framework will work as a supporting architectural context for the value added services. Information
from existing systems, such as automation, maintenance, production control, and condition monitoring
systems, is analyzed, refined and used in control activities by the services.
From the phases of the life cycle of an industrial
manufacturing plant, the operational phase is the
longest and during it the plant is expected to pay
back all costs of its life time and produce profit for
its owner. Changing operating environment,
squeezing global markets, and decreasing profit
margins, among others, place challenges in fulfilling
these expectations. To remain competitive,
manufacturers need to adopt advanced and
integrated operation and maintenance (O&M) of a
O&M means activities that are performed during
perational phase of a plant. The primary goal of
O&M is to secure the undisturbed operation of a
production plant and to make manufacturing as
profitable as possible. Critical aspect in reaching this
goal is information management and application of
modern information technology. Manufacturers
should be able to fully utilize the information in
O&M information systems. Achieving this requires
the ability to easily integrate these systems.
An effort called OpenO&M For Manufacturing
Joint Work
Group has been recently formed by three
non-profit organizations, MIMOSA, the OPC
Foundation, and the Instrumentation and
Automation Society (ISA) SP95 committee to
harmonize the various existing standards from these
organizations (OpenO&M, 2004). The goal of the
joint effort is to provide the standards and
technology that form an interoperable framework for
exchange of O&M information.
In addition to standards, technologies which
able integration functionalities are also needed.
Web services provide standard means for
interoperating between heterogeneous information
systems utilizing different platforms, programming
languages, and component models (Booth et al.,
In this chapter, current state and problems of O&M
information management are discussed. From these,
requirements for integration are derived. Discussion
is based both on literature and a set of production
plant interviews (Viinikkala, Jaakkola, 2004) that
were conducted in order to examine O&M
information management at Finnish industrial plants.
2.1 Information systems
Figure 1 depicts the information systems used on
different levels of manufacturing organization,
which are defined in (ISA SP95, 2004). The
information systems function as repositories of
various kind of information about the past and
present state of the production plant. Information
Viinikkala M., Jaakkola V. and Kuikka S. (2005).
In Proceedings of the Seventh International Conference on Enterprise Information Systems, pages 361-365
DOI: 10.5220/0002528203610365
concerning the process, its state, control signals and
measurements, is collected and stored by the
distributed control system (Paunonen, 1997).
Maintenance activities are supported by
Computerized Maintenance Management Systems
(CMMS), which are basically database systems that
usually contain information about maintenance work
orders, equipment maintenance history, and
preventive maintenance schedules among other
things related to maintenance.
Charasteristic for the information systems at
industrial plants, though alike in their database
centricity, is that they are built with different
techniques on different software platforms and are
not thus inherently interoperable.
2.2 Integration of operations
Production and maintenance are not separate, but
related and parallel activities. The production
process requires energy, materials, and work force as
its primary inputs, but also production capability as
its secondary input. This is produced by
maintenance. On the other hand, the secondary
output of the production process is the need for
maintenance which is the input for maintenance
(Jonsson, 1999). Ideally, related operations should
be supported by information systems.
Many information systems exist at a plant, but
they are rarely considered as a whole. A given
system serves the needs of a particular activity only
– not the O&M of the whole plant. The starting point
for integration is poor in this sense, because legacy
systems have not been designed to support
Also, because of the information fragmentation
and due to lack of standard interfaces, finding the
relevant information to support O&M decision
making from various information systems is hard.
For example, maintenance history knowledge may
be fragmented to several information systems (Mäki,
Some custom links that are typically not based on
any standard interface have been developed in order
to integrate a system with another system. By using
this kind of point-to-point approach, information can
be exchanged between two information systems.
However, each system has its own data model,
which has to be made compatible with each other.
Thus, the point-to-point approach based on non-
standard interfaces is laborious and inefficient and
thus expensive.
Information that is read from a system or a set of
systems could be analyzed by some intelligent
entity. Based on the result of the analysis,
information could be written into some other system,
thus accomplishing automatic transactions between
participating systems. This kind of intelligent co-
operation, orchestrated by intelligent entities, does
not exist at industrial plants yet.
Majority of information flows between systems
are currently accomplished manually by means of a
human mediator. This situation causes latencies in
information flows between systems. Human effort is
required even in routine information exchanges
between information systems.
In order to overcome these problems, integration
of the existing systems is required. Going further,
information is required to be accessed by using
standard interfaces, which the information systems
are required to provide. Further, information is
required to be processed by services that use the
standard interfaces to access the information. Such
services are required to contain business logic or
intelligence to accomplish the information
integration or to implement the required co-
operation among systems.
Figure 1: O&M information systems. Modified from
(ISA SP95, 2004) and (Viinikkala, Jaakkola, 2004)
2.3 Networked O&M
The responsibility for maintaining the production
equipment is transferring from proprietary
organizations of the production plant to dedicated
service providers (Viitamo, 2000). The traditional
arrangement, especially in the heavy process
industry, has been organizing maintenance with
plant maintenance personnel.
Responsibility for equipment effectiveness
requires the awareness of the equipment condition
and production state. Thus constant information
exchange between the external service provider and
the production plant is needed. The current
information systems and the integration techniques
that are used in operation and maintenance do not
adequately respond to this need. The collaborative
efforts of O&M rely on the collaboration of human
experts and the plant operators.
Ideally in case of a problem, the required
knowledge and information to solve it is available
immediately, and the necessary notifications to all
relevant directions are sent automatically. Means to
also external interest groups to access O&M
information in a controlled way are needed. The way
of accessing information should be the same for both
internal usage and external interest groups.
Currently, the friction in operations that require the
attention of several people leads to latencies in
various stages of the situation, which causes
unnecessary loss of equipment availability.
To reach the ideal performance in collaborative
O&M, the operations must make use of new kinds of
information processing solutions that support them
in a holistic manner. These solutions must not
override the existing information systems
infrastructure but include them as an essential
sources and storages of data.
2.4 Visibility into operations
The business model that the manufacturers prefer
has changed from a model, in which products are
made to stock, to a model, in which production is
entirely based on orders and customer needs.
Currently, supply chains are becoming faster and
more integrated with support of e-business.
Demand-driven business model cannot be
accomplished without information from lower
automation levels at plants. In enterprise resource
planning-level, which can be seen from figure 1,
plans, that satisfy customer needs, are designed.
These plans can not be accurately made without
knowledge about O&M operations and status.
Forrester recently published results from a survey
of the top 50 global manufacturing executives, in
which they were asked to answer the question:
“What are your biggest problems with global
manufacturing?” The biggest problem the survey
indicated was “poor visibility into plant operations”
(38%). (Forrester, 2000) Thus, the biggest concern
at the enterprise resource planning level is that
managers and other personnel are not seeing the
right information to make the best possible decisions
on how to run the plant. As supply chains are
becoming more efficient, the plant and its operations
are becoming the bottleneck of the whole value
chain. Current systems, as they are, are inadequate in
providing the required visibility into operations.
Also this viewpoint emphasizes the integration
2.5 Architectural requirements
The general integration requirement discussed in
previous sections implies further requirements for
the integration architecture.
The integration architecture should enable
extending the existing information systems, which
will continue to serve the plant level functions. The
extension functionality must be generic so that it can
be reasonably provided by parties other than
information system vendors.
Due to the heterogeneous nature of the existing
information systems, the architecture must abstract
the existing information systems and expose their
data content in a unified form. Therefore the
architecture must define a mechanism for describing
the data content that is exchanged. The technology
with which the connectivity is implemented must be
independent of software platforms and operating
Standardization efforts should be used as
guidelines for defining the data content that is used
inside the integration architecture. Great input to the
vertical integration can be extracted from ISA SP95,
but for integration of O&M activities on the factory
floor level, the standardization work is still ongoing.
Results can be anticipated from the OpenO&M
consortium, for example.
Service Oriented Architectural style (SOA) is a
software paradigm developed around the Web
Services technology. In service oriented
architectures, data and functionality are decoupled
from each other, and dependencies between the
service requesters and providers are minimized. The
requesters and providers are said to be “loosely
coupled”. (Booth, D. et al., 2004)
The service oriented approach suits well the
business environment of industrial operation and
maintenance. Services that do not themselves hold
any significant amounts of data, but transform it, can
be implemented by the external service providers.
The functionality of these services represents the
special maintenance knowledge held by the service
providers. This knowledge can be applied at various
process plants, and it should not therefore be tightly
coupled with data, which is always plant specific.
3.1 Value added services for O&M
O&M value added services are pieces of software
that utilize the existing information system
infrastructure in a production plant to create value
addition to their users. The value addition is
obtained of the existing O&M information systems
by integrating them with these services, instead of
using them solely for their original purpose.
In practice, the value addition can be anything
from data access facilitation to complex process
analysis and automatic control actions. A common
denominator of the aforementioned is that the
services help to make the O&M operations more
integrated and effective. The value added services
provide functionality, which is at its best when
deployed outside the existing information systems.
They are sensitive to the state of the production
process and available production and maintenance
resources. Therefore process, production and
maintenance information systems are the key
sources of information for the value added services.
The value added services generate new
information that flows back to the production plant
and partner organizations and information systems.
They can also generate notifications and perform
automatic control actions. The services can be
configured by the human users of both parties of the
partnership model to produce value addition. The
configurability of the services is a very important
issue from the partnership perspective, because it
makes possible for the services to encapsulate
knowledge that might not be available for the other
The information exchange requires consensus on
the data models to be used. OPC XML DA provides
a great basis for conveying single data items such as
process measurements. On more complex data, such
as maintenance work orders or condition monitoring
data, other standardization efforts have a crucial
The operation and maintenance value added
services can be categorized to three levels according
to their complexity and the value added. Services on
the lowest complexity level simply combine the
information to one customizable view. The value
addition provided by these services is the ease of
access to the relevant information. The second level
of service complexity is that the services refine the
collected data and in this way create new
information. The second level services can be used
to produce key figures, such as OEE (Overall
Equipment Efficiency), from the process of interest.
Services at this level must possess and apply
knowledge about how the information is calculated.
They must also be able to maintain state and
preserve process information. The third and the
highest level services interact autonomously with
their environment. Services at this level can act as
high abstraction level controllers that control the
work processes of production and maintenance.
A prototype of an intelligent value added service,
called Reasoning Service, has been implemented by
us. The service refines process measurement data
into new, more abstract information, and into
automatically performed actions that control the
workflows of O&M. The transformation of data was
performed using a fuzzy inference system, which
can be used for both function approximation and
classification tasks. The exemplary service is
presented in detail in (Jaakkola, 2004). The
prototype demonstrates the potential of standard
interfaces such as OPC XML-DA as the enablers of
systems integration. It also shows how modern AI
techniques can be utilized within O&M value added
3.2 Service framework
A web service based framework provides supporting
infrastructure for value added services (Kondelin,
Karhela, and Laakso 2005). Figure 2 depicts layers
of the developed service framework architecture.
The framework is independent of organizations
and software platforms and therefore promotes the
interoperation of information systems, exceeding the
organizational boundaries. The common interface
description and data model specification allow the
value added services to connect to the data contents
of other services.
Core services transform data from underlying
legacy systems to the common data model of the
framework. This way they bring the data content
from existing information systems infrastructure to
the reach of the value added services. The core
services of the service framework are to some extent
analogous with the vendor specific adapters of
MIMOSA (MIMOSA, 2004). The core services,
however, do not usually interact with each other
directly, but through value added services. Also the
Figure 2: Service framewor
data model of the service framework is not as
explicit as the one of MIMOSA, but rather a generic
The OPC XML-DA interface (OPC Foundation,
2003) has been chosen to be the data access interface
of the service framework. Therefore, an OPC XML-
DA wrapper, which relays Web service messages to
an OPC DA interface, can be considered as a core
service of the simplest kind.
Access control and the discovery of services are
examples of issues that cannot be resolved in the
web service environment only between the client
and the service. The framework must thus include
services that provide functionality for these issues.
In the service framework, these kinds of services are
called Framework Services. Framework services are
common to all, and they do not represent any system
or organization. Access Manager is a framework
service that controls the authorization of users to the
data content of the core services and value added
services. Access control is especially important
when the integration of information systems exceeds
organizational boundaries. The discovery activity
that is essential to the web service architecture is in
the responsibility of the Registry Manager. The
Registry manager provides similar functionality as
UDDI, but typically its services are used at runtime
instead of design time.
Based on requirements derived from current state of
the art, a solution conforming to the ideas of SOA
was proposed. The concept of value added services
for O&M and the supporting service framework
were thus introduced.
It can be concluded that the web service based
architecture enhance the information management
issues discussed. Firstly, internal information flows
can be automated thus leading to enhanced internal
collaboration of integrated operations. Secondly, the
solution supports the organizational changes that
have lead to networked O&M. Finally, better
visibility from top floor into the O&M information
is obtained, alleviating potential bottleneck problems
of the value chains that characterize the current
demand-driven business model.
Web services provide independent means to
create interoperability between heterogeneous
systems of industrial plants. However, in order to
provide more advanced and dynamic reasoning and
services, information must be semantically enriched.
While information can certainly be better managed
by utilizing the proposed solution, the tacit
knowledge, present at plants and within partners still
remains largely unexploited. Currently, we have
begun is to research technologies that are
semantically more advanced and to use these
technologies in the challenging domain of O&M.
Booth, D. et al., 2004. Web Services Architecture, W3C
Working Group Note 11 February 2004, W3C. (20.10.2004)
Forrester, 2000. Forrester Research.
ISA, 2004. The Instrumentation, Systems, and Automation
Society, (20.10.2004)
ISA SP95, 2004. ISA Draft 95.00.03, Enterprise Control
System Integration Part3: Activity Models of
Manufacturing Operation Management, The
Instrumentation, Systems, and Automation Society.
Jaakkola, V-P., 2004. An Intelligent Web Service for
Operation and Maintenance, Master of Science
Thesis, Tampere University of Technology,
Department of Automation. Tampere.
Jonsson, P., 1999. The Impact of Maintenance on the
Production Process – Achieving High Performance,
Dissertation, Division of Production Management.
Lund University.
Kondelin, K., Karhela, T., Laakso P., 2004. Service
Framework Specification for Process Plant Lifecycle,
VTT Research Notes 2277, Espoo 2004.
MIMOSA, 2004. (20.10.2004)
Mäki, K. M., 2000. Kunnossapidon historiatiedon
hallinnan kokonaismalli, Licentiate Thesis, Tampere
University of Technology, Department of Mechanical
Engineering. Tampere.
OPC Foundation, 2003. OPC XML-DA Specification,
version 1.00,
OpenO&M, 2004. Condition Based Operations for
Manufacturing, OpenO&M For Manufacturing Joint
Work Group.
OpenO&M Whitepaper-CBO for Manufacturing-
Final.pdf. (20.10.2004)
Paunonen, H., 1997. Roles of Information Process Control
Systems. Doctoral Thesis. Tampere University of
Technology. Department of Automation.
Viinikkala, M., Jaakkola, V-P., 2004. O&M: State of the
Art, ProServ project documentation.
Viitamo, E., 2000. Services in the Forest cluster –
competitiveness analysis, The Research Institute of the
Finnish Economy. Helsinki.