http://lnu.diva-portal.org/
This is an author produced version of a paper presented at “First
International Conference on Maintenance Engineering ICME
2006 New Century New Maintenance October 15~18, 2006,
Chengdu, P.R.China”. This paper has been peer-reviewed but may
not include the final publisher proof-corrections or pagination.
Citation for the published paper:
Al-Najjar, Basim and Kans, Mirka
”A Model to Identify Relevant Data for Accurate Problem Tracing
and Localisation, and Cost-effective Decisions: A Case Study”
Proceedings of the First International Conference on Maintenance
Engineering ICME 2006 New Century New Maintenance
ISBN:7-03-018064-X
A Model to Identify Relevant Data for Accurate Problem Tracing and
Localisation, and Cost-effective Decisions: A Case Study
Basim Al-Najjar, basim.al-najjar@vxu.se Mirka Kans, mirka.kans@vxu.se
Department of Terotechnology, School of Technology and Design, Växjö University, SE-351 95 Växjö, Sweden
Abstract
Mapping the technical and financial effectiveness of a production is crucial for making cost-effective maintenance decisions. But, a successful mapping cannot be achieved without a well-structured database including all relevant data required for this task. The problem addressed in this paper is how to identify the relevant variables and performance measures that should be monitored for achieving cost-effective maintenance decisions that enhance company’s business? The main result achieved is a model for identifying relevant data required for accurate problem tracing and localisation within maintenance and production processes using a top down approach. The potential of applying the model is examined in a case study including databases of two commercial software programs developed for maintenance. The main conclusions are integration of IT and data resources within the enterprise is needed for developing a holistic view of the production process and a well-formulated and documented procedure of data identification will ensure that the data can be traced back to root sources and in this way we can support the work of continuous cost-effective improvement by eliminating root causes of problems at an early stage.
Key words
Relevant data, Maintenance cost-effective decisions, Common database, Top down analysis, IT integration
1. Introduction
In the past, the survival of a manufacturing companies was mainly connected to how much merchandise it was able to push into the market. This situation has changed and today's strategies imply cost minimisation and differentiation and the ability to use available resources in a cost-effective way. The focus on customer needs puts great demand on the production system to meet the goals of high product quality, production safety and delivery on time at a competitive price, Al-Najjar (1997). Manufacturing industries realise the importance of monitoring and following up the performance of a production process by using financial and technical indicators. It establishes a bridge between the operational level in terms of e.g. productivity, performance efficiency, quality rate, availability and production cost, and the strategic level expressed by the company’s profit and competitiveness. In general, company’s business can be influenced by many factors such as the maintenance policy, production method and procedures, quality system, organisation, personnel competence and training level. It is not enough to focus on one or two of these factors in order to develop a holistic approach when mapping the production process in a cost-effective way, Al-Najjar (1996). The problem addressed in this paper is; how to
identify the relevant variables and performance measures that should be monitored for achieving cost-effective maintenance decisions that enhance company’s business?
A well recognised data set for each type of decision will ensure that decisions made are based upon relevant and representative data. Further, by considering relevant data from relevant working areas such as maintenance policy, production method and procedures, quality system, organisation, personnel competence and training level, cost-effective maintenance decisions can be achieved effectively. The development of information technology (IT) in the past decades has enabled complex decision-making in rapid and easy manner and storage possibilities of large amount of data.
In the paper, a literature survey is done, see Section 2, to highlight whether the problem addressed is studied before or not and to explore the researches done on the related topics. In Section 3, a theoretical frame work introducing definitions of the most important keywords (data, information and knowledge) being utilised by this study, integration of IT systems and principles for developing an industrial database. The top down analysis model is developed and discussed in Section 4. A case study, which includes databases of two maintenance-used software programs, for verifying the potential of applying the model is presented in Section 5, which is followed by the results and conclusions in Section 6.
2. Literature survey
A literature survey was conducted in order to cover recent researches that have been done within the area of the problem addressed in this paper. Following article databases were be considered in our three step-survey: Elsevier, Emerald, Proquest and IEEE. The keywords that have been used in the first step-survey, which had its focus on the technical part were: problem tracing, problem localisation, database, data identification model, decision-making and maintenance. The output of the survey was ten relevant papers. In order to cover the area of IT and its integration aspect we conducted the second step-survey using the following keyword: data integration, information systems integration and IT systems integration. The third step-survey focused on joining the following keywords: maintenance, production, business, strategy/strategic, productivity and competitiveness with IT, information systems, information systems integration and IT systems integration in different combinations. The output of the second and the third step-survey was about 100 papers.
The main result of the first step-survey was that no articles dealing with the identification of relevant data for accurate problem tracing and localisation and decision-making were found. Dwight (1999) presents a model for determining maintenance performance based on a top down approach; the systems auditing, where the performance of the maintenance system is measured against a standard "ideal" system but where the different activities also are judged by their relative importance for the system. He introduces a new model, which can be applied in maintenance decision-making and promotes the value measurement approach, i.e. translating technical issues into financial terms.
Maintenance has recently been considered as an activity contributing efficiently to the companies' strategic objectives in profitability and competitiveness. The awareness of maintenance as a strategic factor within a company is established in literature, see for instance Riis et al. 1996, Al-Najjar (1996), Murthy (2002), Tsang (2002), Swanson (2003), Al-Najjar & Alsyouf (2004), Alsyouf (2004). The connection between operational and strategic corporate needs is important otherwise the various activities within the company cannot be compared or valued. This has
been addressed by many authors in different contexts e.g. Pintelon & Van Puyvelde (1997), Al-Najjar et al. (2001), Miraglia et al. (2002), Otto & Kotzab (2003) showing the need of measurement systems and performance indicators to estimate the impact of operational activities on the strategic level.
In contemporary research we can note the trend towards more integration of both administrative and industrial automation of IT systems, see for instance Long et al. (1998), Kang & Golay (2000), Herzog et al. (2002), Gordon (2002), Xu et al. (2003) and Vojdani (2003), where for example Weber & Pliskin (1996) show that IT systems integration has a positive effect on company performance effectiveness. Until recently, these systems have been separated, but the trend is now moving towards integration of technical systems and administrative systems, see for example Dahlfors & Pilling (1995), Gordon (2002), Herzog et al. (2002), Bratthall et al. (2002) and Kjaer (2003). The need to integrate IT with business objectives and corporate strategy is a commonly addressed topic in literature; see for example Holland & Light (1999), Segarra (1999), Szirbik & Jagdev (2001) and Sarshar et al. (2002). Many authors realise that investments in IT have a positive correlation with profitability and competitiveness, e.g. Johnsson (1999), Kini (2002), Dedrick et al. (2003). On the other hand, IT systems fail to meet the demands of their users, Whittaker (1999), Bennet et al. (2000) and Evgeniou (2002). Whittaker (1999) and Bennet et al. (2000) shows the importance of including disciplines such as law, business and economics in software engineering and IT projects in order to achieve the objectives expected. Evgeniou (2002) pinpoints the current problems today's companies face in IT system integration: either suffering from lack of information visibility across the enterprise due to insufficient integration or suffering from information inflexibility due to enforced information centralisation and standardisation such as in an ERP system. In worse case, the ambition to integrate all enterprise IT systems will result in enterprises that are enforced to adapt to the underlying logic of the IT system and not the other way around. IT integration would therefore better be driven by the strategic goals of the business, Evgeniou (2002). To be able to show the impact of maintenance on the production and on strategic level, data from various IT systems serving different working areas relevant to maintenance and possibilities to process the data are needed. But today’s IT systems for maintenance purposes are mainly administrative and restricted to maintenance budget control, Pintelon et al. (1999).
3. IT integration and cost-effective maintenance decisions: Theoretical framework
3.1 Definitions
Cost-effective decisions can be achieved by integrating, analysing and effectively utilising relevant data, information, databases and knowledge. The terms data, information, databases and knowledge have been defined and interpreted differently in literature. In this section we discuss the terms briefly and define them.
Blanchard (1998) relates cost-effectiveness to system effectiveness that could be expressed in terms of e.g. availability, dependability or performance, depending on the specific system mission and to the total life-cycle cost. Li & Brown (2004) considers maintenance cost-effectiveness as the combination of maintenance tasks that can achieve the best reliability within a limited financial budget while Al-Najjar (1997) expresses it as the long-term benefit of a specific maintenance policy. The cost-effectiveness of a maintenance policy can be determined by comparing the economies output before and after the investment done for improving the performance
of the maintenance policy in question, Al-Najjar (1999). In this paper, following definition is used: Maintenance cost-effectiveness is a measure of how much the invested capital in the considered maintenance policy is economically beneficial in the long term.
In literature, there is no clear agreement about the difference between data and information. Some authors see data as holders of information in some way, for instance expressed as "known facts that can be recorded and that have implicit meaning", Elmansri and Navathe (2003) or as "facts that have meaning in the user's environment", Hoffer et al. (2002). Other authors define data as symbols without meaning, Curtice et al. (1981), Flensburg & Friis (1999). Flensburg & Friis (1999) regard information as data put in syntax, i.e. in a structure, while Hoffer et al. (2002) defines information as "data that has been processed in such a way that it can increase the knowledge of the person who uses it".
In this paper, data are defined as: The abstract description of a set (or sets) of measurable variables from relevant working areas in an enterprise. This is because we recognise that the symbols (data) need an origin of some art, i.e. variables. While information is defined in this paper as: The output that is drawn from data based upon their contents with respect to a particular environment, i.e. with respect to the processing aspect and perspective.
The root source for data is different sets of measurable variables defined and measured in activities during the business processes. The values of these variables vary in time periodically or randomly, see Figure 1. If we combine two or more units of data, we soon may get unlimited ways of interpreting them into information. Further, one unit of data provides the basis for many units of information. For example, the symbol 46 could for instance denote the temperature warning-limit of a process, the number of defective items produced in one week or the minimum time length of a product to be assembled. To be able to determine the meaning of a symbol it must always appear with a definition that indicates the type or unit, Curtice (1981). One difference between data and information as the terms are defined in this paper is the inherent degree of usefulness of the two terms. Data holds low extent of usefulness, as it must be interpret and put into syntax before it becomes information. Information holds higher extend of usefulness, but it is not the same as knowledge. Nakkiran et al. (2003) regard knowledge as the human expertise stored in a person’s mind, gained through experience, and interaction with the person’s environment. The authors' definition of knowledge is: Relevant information able to be utilised for making relevant decisions through applying logical inference for updating human understanding.
One can see that the possible complexity and subjectivity influences the level of usefulness the term holds. Variables and data are seen as objective facts while knowledge is connected to human mind and becomes more subjective. According to Hoffer et al. (2002) and Elmansri and Navathe (2003) a database is a collection of related data. Data integration is a term closely connected to databases, where Lenzerini (2002) defines it as "the problem of combining data residing at different sources and providing the user with a unified view of these data". This is expressed in a slightly different way by Boucelma et al. (2002): "a goal to construct a global description of the data coming from a multitude of heterogeneous sources". In this paper the term database is defined as: An organised collection of related sets of measurements (data). While data integration is defined as: The integration of heterogeneous relevant data resources on a descriptive homogenous level with respect to a particular perspective serving special aspects.
X1.1 at time t 1.1 X1.2 at time t 1.2 ... X2.1 at time t 2.1 X2.2 at time t 2.2 ... X3.1 at time t 3.1 X3.2 at time t 3.2 ... Y1 Y2 Y3 Z1 Z2 Z3 Z4 Information (Processed data with particular aspect and perspective)
Increasing level of usefulness, subjectivity and complexity Knowledge (Utilised information in a particular human being’s environment) Measurable variable Data (Sets of measure ments) H1 H2
Figure 1. Relation between measurable variables, data, information and knowledge Data integration means that data can be stored, processed, analysed, interpreted, combined and accessed at demand in an easy way. Cost-effective maintenance decisions are dependent of analysing and effectively utilising relevant information reached by using relevant and high quality data. Data quality is dependent upon factors such as accuracy, timeliness, consistency, completeness, relevancy and fitness for use, Fisher & Kingma (2001). Information sharing, i.e. supplying relevant, easily accessible, timely and high quality data to right persons for supporting decision making is one of the driving forces behind IT systems integration today, Toussaint et al. (2001), Evgeniou (2002), Boucelma et al. (2002). Although commonly used in literature, the terms information technology integration and information systems integration are not defined by the various authors. In this paper information technology systems integration is defined as: Structural integration of IT systems, mainly computerised systems, including hardware, software and communication media used for information handling and ways of interactions between or union of these systems.
3.2 IT systems integration development in a strategic perspective
Integration of IT systems as well as the connection between IT systems and business needs has been discussed ever since the use of IT systems emerged in companies some 40 years ago, Bennet et al. (2000). Three or four decades ago, computers were used to automate already existing routines and to take advantage of the enormous information storing and retrieving capacity of computers. This is in order to reduce the administrative costs mainly in form of reduced work force, Persson et al. (1981). The automation was specific for one function or department and resulted in different IT systems for different tasks that could not interact, see for instance Mullin (1989).
As the companies became more computerised in the 70s and 80s, the demands for co-ordination of these computerised routines arose in order to reach competitive
advantages through reduced lead times and higher productivity by better planning and follow up. The driving force of the co-ordination phase was to enable communication between sub-systems and modules, realised by the use of middleware, Tuunainen (1998), Chi & Wolfe (1999). The concepts of relational database management and distributed database management that emerged in the 70s and 80s separated the data from the information technology and was the first big step towards more integrated data resources, Persson et al. (1981).
The importance of the corporate knowledge grew, as the economy changed from production centred to customer centred in the 80-90s, Kelly et al. (1997), Dahlbom (1997). To gather and distribute important information to the right persons at the right time became necessary, as well as networking with external parts, such as customers and suppliers. These challenges, together with the globalisation of businesses, made IT system integration necessary, Holland & Light (1999), Chattopadhay (2001). A wish to integrate not only corporate data, but also information resources, provided the basis of today’s business wide IT systems, for instance the enterprise resource planning (ERP) systems, manufacturing planning systems (MPS) or enterprise asset management (EAM) systems, Nikolopoulos (2003). In parallel with the development of integrated business administrative IT systems, development of computer technology in production resulted in for example robotics, programmable logic controllers (PLC) and supervisory control and data acquisition (SCADA) systems, Gordon (2002). Today, integration of administrative and automation IT has become an important issue on the corporate agenda, in order to reach cost effective decisions on all levels of the organisation, see for example Dahlfors & Pilling (1995), Bratthall et al. (2002).
Easy accessible high quality data make the selection of the best combination of the factors required for achieving the highest profit easier. But, criteria and procedures for this selection are necessary, as well as for the data collection, according to the following reasons:
• It is not enough to base maintenance decisions on technical or organisational feasibility or cost minimisation alone, but it should be based on cost-effectiveness. The most cost-effective maintenance decision is the one that gives highest benefit per amount of resources invested.
• Maintenance decisions should be based upon facts and without high quality relevant data it would be impossible to achieve cost-effective decisions. The quality of the output is directly affected by the quality of the input in the process and the quality of the process itself. The way to assure cost-effective maintenance decisions is to assure the quality of input data and of the decision making process itself.
• Optimisation of the production process with respect to only one or two of the process essential elements (subsystems) such as maintenance or quality can in many cases be misleading. Every decision made is therefore necessary to be related to the corporate goals and strategies.
3.3 Principles for developing an industrial database
This section presents five principles recognised by the authors for developing an industrial database, which enables effective tracing of the basic causes behind deviations in the monitored performance measures and cost-effective maintenance decisions.
I. Principle one; Identification of the variables that will be measured, monitored and saved should be achieved by using backward analysis.
II. Principle two; The data gathered should be able to provide a possibility for
developing a reasonable holistic view for rapid technical and economic mapping of the situation at need.
III. Principle three; The database should not be too big to enable easy
management of the data and make it cost-effective. It should primary focus on the most essential measurable variables that can provide possibility for covering relatively wide spectrum of the situation contents.
IV. Principle four; It should be build up based on providing possibilities for
applying a cost-effective continuous improvement. This is why it is rather important to design it with respect to flexibility and changeability.
V. Principle five; It should acquire a dynamic interface with the user enabling
the company to enhance its content and structure at need.
In most cases, when developing an industrial database the connection to the full decision making process is disregarded. Measuring, monitoring, mapping and following up maintenance performance are, in many cases, not defined as important maintenance activities that should be considered when developing the IT applications. Principle one where a top down approach is promoted ensures that the measurable variables needed for measuring, monitoring, mapping, following up performance and preparing decision material are available in the database and further, directly connected to one or more of the process’s or company’s objectives that should be achieved. Data must be gathered in such a way that they reflect as much as possible the true situation in hand representing technical, organisational and economic perspectives in a holistic manner. Principle two promotes the use of data from different sources for creating the holistic view.
While using different data sources, the data gathering must be filtered in a way that the decision making process is not based upon too much data as principle three declares. Large amount of irrelevant data slows down the decision making process and makes the process of tracing root causes hard and maybe confused. Also, lack of or inconsistent data lower the quality of decisions. Instead, the database should be based upon a minimum relevant data needed for solving the most important problems, see principal three. Principles number two and three emphasis the application of a reasonable way to utilise IT integration.
IT systems are, in general, developed based on the specification provided by the customer. The inquiries of customers, even in the different company’s departments, are different and usually concern which functionality that should be considered in the IT application, which in turn directly affects the set of measurable variables to include for achieving the desired functionality resulting in specialised databases. When the software is developed in-house, it often leads to great heterogeneity regarding data, information and technological solutions. Off-the-shelf software such as computerised maintenance management systems (CMMS) and ERP systems uses an approach where standardised data structures are enforced onto the business, and to be able to suit every customer, the data content will be either huge or too restricted. The information standardisation may also lead to a terminology that the users do not recognise and to an IT system that will not be utilised in full extent. Both situations are unsatisfactory; therefore a solution on data integration level rather than on IT integration level would be the most suitable for an industrial database. Different heterogeneous data sources can in this way be used to create a homogenous, common database for specific needs without changing the structure of the root source, keeping
the ability to trace back to the root sources. This common database connected to a decision support system (DSS) or to an integrated DSS module in the current CMMS provides the basis for cost-effective decision making.
In general, the company's internal and external environment undergoes changes during the lifetime of the company’s database. To ensure that the database meets the requirements and has the ability to adjust itself to support improvements continuously, it is important to design the database and its applications with respect to flexibility and changeability as promoted in principle four. Otherwise the database could be out of date already when it’s taken into use and further improvement work within the application area will be disabled. IT alone cannot ensure a cost-effective maintenance decision, as long as it is humans that will make the final decision. It is therefore essential that the human interaction with the IT tools is enabled according to principle five. This would for example allow for reducing the uncertainty in a situation where some essential information is lacking when simulating different situations. In accordance with principles four and five, possibilities to enhance the common database without changing the application software, or the other way around, are needed. An object-oriented module based system architecture using web-based applications and distributed database management would give us the opportunity to achieve this objective.
4. Top down analysis for identifying maintenance relevant data 4.1 Maintenance data analysis process
In this paper, identifying relevant data is considered a structured process, where prerequisites such as business goals, strategies and objectives, working procedure for decision-making and the maintenance tasks to be performed all affect the final choice of the data and data sources required. If decisions and activities carried out on a work task level are not based upon the overall objectives of the company business stated on the strategic level, the risk of doing things with less efficiency or effectiveness is not small. The maintenance activity that is done somewhere in the production process might be appropriate to solve an immediate problem, but not necessary being the most cost-effective one. Therefore, the starting point of top down analysis approach of maintenance data is the corporate strategic goals.
The major strategic goals of vast majority of companies would be to gain more market shares and higher profitability, but the way of reaching these goals could differ. Therefore, the corporate goals are decomposed into objectives and strategies for the different processes such as production and maintenance processes. The objectives of the production process in a paper mill could be defined as producing high quality paper in the most cost-efficient way while a nuclear power plant stresses more the safety aspect. Promoting safety does not necessarily mean neglecting the dimension of cost-effectiveness. The same level of safety could be achieved using different approaches and each may result into different impacts on the company's business cost-effectiveness. In this paper, the role of maintenance is considered to be; to maintain the quality of all the essential elements that contribute in the production process to achieve high quality products that are delivered on time at a competitive price. Achievement of maintenance objectives can be recognised using different measures such as availability, productivity, safety, quality and savings in the production cost.
To ensure a correct mapping and assessment of the situation, relevant data must then be determined from the total sets of data available. The relevant data can be
collected from the production process and other relevant processes in the company. Technical data can be collected directly from the machines and work orders, financial data can be found in corporate IT systems while other data might not be found in any computerised systems and should be picked directly from the production process. The relevant data are analysed at need, such as when one or more variables exceed predetermined warning levels. Based upon the analysis, the actions will be taken in order to reinstate the production into the desired state. The results of the maintenance analysis must always be compared with the maintenance objectives to assure that correct decisions were made, i.e. feedback. Feedback in the maintenance analysis process can be utilised for comparing the data that have been used with the requirements stated in the maintenance strategies, policies and procedures to ensure data quality and that the correct sets of data were used. Also, it can be used for justifying maintenance objectives achieved with the company’s goals.
4.2 Top down analysis model for identifying maintenance relevant measurable variables
In Figure 2, a top down analysis model for identifying relevant measurable variables required for monitoring and following up the development of the state of a production process with respect to strategic objectives is presented. The model consists of four phases divided into eight distinct steps, where a continuous arrow shows the direction of procedure and a dashed arrow the flow of data.
A decision defined in phase 1, i.e. whether stop the production or not, is based upon strategic objectives within safety, production, life cycle cost, quality or maintenance or any combination of the objectives. This decision is dependent of the real time assessment of the condition of the elements constituting the production process and therefore problem diagnosis and evaluation of the situation at the component, equipment or process level is necessary (Phase 2). But, it cannot be done effectively without monitoring and collecting relevant technical, financial and organisational data, which are included in phase 3. In order to make an accurate cost-effective maintenance decision, accurate warning and action limits should be established properly based upon past data and statistical tools, Al-Najjar (1997). When proper performance measures (indicators) are identified, suitable data for estimating these measures will be easily identified (Phase 4). A common database for this purpose should be established and data sources identified to provide the database with actual data. Finally, the specific measurable variables and data gathering policy defining when, how and where these variables should be measured can be defined effectively. By defining the measurable variables, tracing back the root source of each decision can be achieved easily. If suitable measurable variables are not found for the specific data, other sources and/or measurements must be defined.
In Figure 3 the top down identification procedure is put into a context that explains what is needed before starting each step of the identification procedure (prerequisites), the expected results of various step (results) and reasons behind taking any step in the procedure (motivations). For each step in the procedure, there are prerequisites in form of control documents on corporate or business activity level, tools and knowledge, which are needed for performing the step. The quality and amount of relevant prerequisites at hand will directly affect the result gained of each step. The results are formulated in the shape of lists containing the output from each step. The output from one step lays the ground for continuing into the next step.
Step 4: Identify relevant key measures for
following up the development easily
Step 5: Define the measures’ acceptable and
rejectable limits
Step 3: Monitor specified technical and
financial key measures
Step 1: Define cost effective maintenance
decisions with respect to the enterprise objectives
Phase 1: Decisions at process
level with respect to strategic objectives
Phase 3: Follow up at working
areas’ levels, e.g. production, quality, machine condition and costs, with respect to the strategic objectives
Step 2: Make diagnose, prognosis and prediction
to evaluate the situation when it is necessary, e.g. at deviations or regularly
Phase 2: Diagnosis, prognosis
and prediction at component, equipment and process level
Step 8: Identify relevant measurable variables
suitable variables not found
Phase 4: Relevant data
gathering
Step 6: Identify suitable data sources for the
common database
Step 7: Define measuring policy
Phases Steps
Figure 2. Process phases and steps of identifying relevant measurable variables For example, the outcome from the identification step number two (Make diagnose and evaluate the situation) for instance will be a list of the possible problems that may arise at component, machine or higher level influencing production and demands. A decision to stop a machine can be based upon the condition of a component in machine, but can also be based upon production process level related to productivity and efficiency of the process, product quality or production costs. When the possible problems are identified, we can move to next step for identifying relevant performance measures/indicators required for monitoring the problems in question. For getting a broader understanding of why to conduct the steps defined in the procedure, the motivation of each step is presented in the rightmost column. Step two (Make diagnose and evaluate the situation) is for instance important for identifying all reasons behind the problems on a root cause level. This will ensure that correct actions will be taken that reduces or eliminates the true problem and not only the effects or symptoms of a problem.
8) Identify relevant measurable variables
7) Define data gathering policy 6) Identify suitable data sources for the database 4) Identify relevant performance measures
5) Define the acceptable and rejectable limits
3) Monitor specified technical and financial performance measures
2) Make diagnose and evaluate the situation 1) Take cost effective maintenance decisions with respect to the strategic goals
Measuring and analysis policy, monitoring system Maintenance and production nume-rical objectives
Identification procedure Result Prerequisite Technical, financial, organisational and managerial understanding and considerations Maintenance competence, goals and strategy Corporate, production and maintenance general strategic goals Information strategy: what, where and how to measure and keep?
List of relevant decisions within related working areas
List of possible problems/problem causes on process, equipment or component level that demand decisions List of deviations in the state of the components, equipment or process
List of warning and action limits for each key measure List of relevant key measures
List of data sources: location and type of each key measure List of measuring tools and periodicity for each type of measurement List of measurable variables
Motivation
2) Identify and localise basic reasons behind problems, e.g.
machine/component, operator, production method or maintenance
5) Determine acceptable e.g. vibration, noise and temperature levels, acceptable quality rate, performance efficiency, availability & production costs
6) Identify where data can be found, i.e. either in computerised databases, papers or non-documented
8) Identify measurable variables generated by the production process, from production, LCC, maintenance, quality etc.
7) Define how and when to measure using which tool 4) Identify those key measures that are required for mapping the situation of the production process 3) Identify and localise technical and financial deviation in an early stage
Knowledge and experience about the company’s working areas
1) Stop production for doing maintenance, allocate and motivate investments
Figure 3. Model for identification of relevant measurable variables with prerequisites, results and motivations for various steps
5. Case study
In this section we introduce a case study, which aims to:
1) Investigate the possibilities of finding technical and financial data required for cost-effective decisions within the available IT applications in maintenance. 2) Analysis of the design and size of the databases with respect to the principles
stated in Section 3.3.
The databases of a computerised maintenance management system and an asset management module in an ERP system were chosen as study objects. The two objects represent different approaches for providing the maintenance organisation with IT support. The CMMS has specialised on maintenance data requirements in a standalone application with small interference with other IT systems, while the asset management module is one of many modules within the ERP system, which in turn has an aim to support all activities within the company. Both objects are based on the relational database model and are implemented in Microsoft (SQL) Server/Microsoft Access and Oracle Application Server respectively, where SQL stand for Standard Query Language. The designs of these two databases are representative for the IT systems that are used for such purposes but the actual structure of the databases are not directly accessible. Therefore reverse of the data structure was conducted in order to recreate the structure of theses databases. Reverse engineering is the process of taking something apart, e.g. a database, and analyzing its work in detail. In this paper, this was made by studying the forms and reports available in the application software to recreate the structure of the underlying logic. The logical database structures are presented in Figure 5 and 6 and a description of the notation used in the models is found in Figure 4. All attributes were named according to the data found in the
applications, while the entity names have been set by the authors to describe the set of attributes in the most proper way.
ENTITY #2 ENTITY #1 Attribute #1 Attribute #2 … Attribute #n Attribute #1 relation 1 M Relational attribute #1 Cardinality denoting the participation of entities within a relation. 1 = single participation and M or N = plural participation.
Figure 4. Notation used in the logical database models
O BJECT N um ber N a me D e signa tion D ow ntim e c ost P la ce me nt CA TEG O RY Cha ra c teristic Ty pe P ERSO N N a me H ourly ra te CO MP A N Y N a me A ddre ss P hone Fa x E-ma il A RTICLE N um ber N a me U nit P urc ha se pric e Inve ntory pla c em e nt Inve ntory ba la nc e Re orde r point JO B Ty pe P riority Sta tus P e riodic ity Sta rt tim e Expe c te d finish tim e Tim e c onsum ption D ow ntim e O the r w ork c ost
1 1 1 1 1 1 1 1 1 1 M M M M M M M M N N M D O CU MEN T D oc ume nt nam e 1 1 1 1 1 1 1 1 1 1 co ns is ts o f
is div ide d into
is div ide d into
is div ide d into
is div ide d into
is d iv id ed in to is c onne c te d to is co nn ect ed to is c onne c te d to is c onne c te d to is co nn ect ed to is m ade on prov ide s c onduc ts is co nn ect ed to M us es
Figure 5. Study object CMMS
5.2 Analysis of data content
The applicability of these two databases for relevant technical and financial decisions within maintenance was tested by using two maintenance performance measures: the direct maintenance costs per produced quality item (DCQ) and losses due to non-utilisation of the fixed cost because of production performance disturbances (NPD). These performance measurements were chosen due to their structure, where both technical and economic data are needed in order to assess them.
DCQ is a measure of the direct maintenance costs measured per quality products produced by the process. DCQ expressed by Eq(1) is a measure for displaying technical and economic impact of a maintenance policy on production process effectiveness. It is measured by the cost paid for maintaining the process capability
and effectiveness may influence the quality rate. Maintenance direct cost is
cS = Spare parts costs
cM = Man hour costs
cO = Overheads (including depreciation of maintenance equipment and consumed
material)
cOUT = Outsourcing costs
DCQ is calculated as:
(
)
(
P R)
c c c c DCQ S M O OUT − + + + = (1) whereP = Total amount of produced items and R = Amount of rejected items.
NPD expressed by Eq(2) displays the losses due to the non-utilisation of the fixed cost because of production performance disturbances due to failures, unplanned but before failure replacements (UPBFR) and short stoppages, Al-Najjar and Alsyouf (2004), which mostly could be regained by eliminating the root causes behind the problems. The losses of UPBFR are in many cases equal to (or a little less than) to the economic losses due to failures. Failures and other stoppages can, in general, be reduced by applying more efficient maintenance policy. NPD is assessed as:
(
)
P F S UPBFR F t t t t c NPD=∑
+∑
+∑
(2) wherecF = Fixed operation costs
tUPBFR = Duration of UPBFR
tS = Duration of failure
tF = Duration of short stoppage
PRODUCT UNIT PARAM ETER ACTION CAUSE ROLE EM PLOYEE ECONOMY W ORK ORDER DEPARTM EN T SITE PART SUPPLIER PERM IT PERM IT A TTRIBU TE ROUTE EVEN T OBJECT FAU LT Default Default Default Amount Name Name Hourly rate Description EmployeeId Pager Name Mobile Phone Signature Fax
Remark Quick Phone Int. Phone E-mail
Cost type Planned cost Planned revenue Planned margin Actual cost Actual revenue Actual margin RoleId Hourly rate Description Name Type Description Name ParameterId Name Type Unit code Type Factor Base unit User defined Description
ObjectId Planned stoppage Status Planned production Type Current position Category Purchase date Criticality Purchase price Warranty Cost per year
ProductId Quality PartNo Inventory part Unit Quantity on hand Quality Price Description SupplierId Attribute code Description Type Description Classification RouteId Name Type Symptom Class Discovery Description W ork orderN o Date Description Status Type Priority Start Completion Execution time Required start Latest completion Directives Quotation Statistical code belongs to has could be included in produces ha s has belongs to is needed to conduct is created due to has uses conducts is ca rr ied o ut b y ha ppens on supp lies is pla ced in co uld n ee d < -tri gg er s col le cts --> w ill r es ult in co uld b e d ue to 1 1 1 1 1 1 1 1 M M M M M M M M N has N N N N N N N M M M M M M M M M M N 1 1 1 1 1 N M is collected in M N Amount Scrapping co ns is ts of M N N M M 1 supplies belongs to TESTPOIN T Name Location M ha s 1 Discount Other costs
Figure 6. Study object ERP system
5.2.1 Analysis of the computerised maintenance management database
• Direct maintenance costs per produced quality item
The CMMS database contains some attributes to register maintenance direct costs (cS,
cM, cO and cOUT). For cS only the purchase prise is available. The additional costs arise
due to transport and inventory-keeping expenses are excluded. This is why it is not
representative. cM is assessed using the attribute “Time consumption” for different
jobs (in the JOB entity) and the hourly rate of the personnel, found in the PERSON
entity. The distinction of cO and cOUT is not available separately, but both are
registered as "Other work costs" in the JOB entity. There are no entities dealing with
production in the CMMS database. Thus, total production P and rejected items R
cannot be found in the CMMS. To be able to calculate DCQ, production related data must be gathered from elsewhere. The maintenance cost data provided by the CMMS is covered in few attributes that makes the traceability very low.
• Losses due to non-utilisation of the fixed cost
Only one attribute supporting the assessment of NPD is found in the CMMS database: the downtime of an object (machine or equipment) when a maintenance job is performed. This attribute is found in the JOB entity. As these data are connected to a maintenance job, the duration of short stoppages are most likely not included though. Full data coverage to assess the production performance inefficiency is therefore not reached. Fixed operations costs and planned operation time is not to be found in the CMMS database. Assessing NPD is not possible by using the CMMS alone. Further, as long as the data required for categorising different stoppages into failures, UPBFR and short stoppages are not completely available, the possibility of eliminating the root causes and preventing their reoccurrence is low.
5.2.2 Analysis of the asset management module database
• Direct maintenance costs per produced quality item
All direct maintenance costs could be collected from the ECONOMY entity for work orders, where the following cost factors are connected to the work orders: personnel, material, external expenses (outsourcing) and fixed price costs. The latter cost type is
triggered if a job has a fixed internal price. Man-hour costs cM could be also assessed
through the attributes "Execution time" (in WORK ORDER) and the "Hourly rate" of
ROLE and DEPARTMENT. Spare parts costs cS could be assessed more thoroughly
using other attributes in the database, as spare parts costs are registered in the PART entity as price and additional purchase costs are found as a relational attribute to
PART and SUPPLIER. Regarding total production P and rejected items R, we found
them as relational attributes to the relation OBJECT produces PRODUCT. Data needed to assess DCQ is thus available.
• Losses due to non-utilisation of the fixed cost
Fixed operating costs cf and planned operation time tp could be calculated by using
attributes in the OBJECT entity. There are no attributes for capturing the lower
performance efficiency in form of tUPBFR, tS and tF. Only stoppage types and amounts
are recorded in the EVENT entity and the relational attribute connected to EVENT and OBJECT. To be able to calculate NPD, data about stoppage duration must be gathered from elsewhere.
5.3 Analysis of design and size
In chapter 3.3 five principles for developing a relevant database were presented by the authors. In this section, the structure of the databases of the case study will be analysed with respect to these principles.
• Principle I: Use of backwards analysis
The objects of study were commercial software that are designed and developed to suit different kind of users. Data provided in the databases of these software applications are therefore not customised with respect to the strategic objectives of various companies. This means that regardless of what strategy for data identification was used, top down or bottom up, the data set does not fully suit the specific user of the application. The use of backwards analysis enables a thorough tracking of root causes of problems, which is partly lacking in the studied databases. The data found in the databases were in some cases aggregated or confused with other data. One
example of this is man-hour costs cM used for calculating DQC. In the ERP asset
management module the value of man-hour costs could be assessed by using at least
two sources of data; 1) by calculating cM using the attributes "Execution time" in
WORK ORDER and the "Hourly rate" of ROLE and DEPARTMENT or 2) by collecting personnel costs from the ECONOMY entity for work orders. The first enables a deeper traceability to root sources than the second one. In general, the traceability is rather low, especially in the CMMS database. This makes tracing back to root causes harder than if more detailed data is available.
• Principle II: Holistic view
The CMMS is designed for maintenance, but concentrates upon the operational decisions of planning, executing and following up of maintenance activities alone. This gives a narrow operational view of maintenance without exploring its impact on the subsystems and working areas that are affected or affecting maintenance. Parts of
the basic maintenance economic data are provided, but technical data is lacking. The asset management database holds a broader perspective and contains entities and attributes from other areas, such as production and quality. It gives a better coverage of maintenance technical and economic data. The ERP system is not designed specifically for maintenance purposes, which results in confusing attribute names as standardised corporate wide terminology is used.
• Principle III: Complexity
The CMMS database is simple in its design, while the asset management module database is data intensive, making it hard to grasp the full data content. In this study the database complexity is reflected in the entity and attribute amount of the two study objects. The entity and attribute amount of the ERP module is about three times as high as in the CMMS showing that the CMMS is less complex than the asset management module.
• Principle IV: Flexibility and changeability
The CMMS developer offers two database management system choices to their customers, one based upon Microsoft SQL Server and one on Microsoft Access. The latter choice restricts the amount of connected users to about ten, while the larger application handles unlimited amount of users. The ERP database management system is designed for unlimited amount of users. The two applications could thus serve the same amount of users. Both databases are relational, giving the same possibilities for future changes. The ability to make changes in the design could be reflected by the amount of relationships that exists between the entities, implying that the possibility that a change in the design will affect other parts of the database rises with the number of interrelationships that exists between the entities. We noticed that the asset management database consists of about 50% more relationships than the CMMS, suggesting that the asset management database has a less flexible design. • Principle V: Dynamic interface
In the CMMS, the interface is designed with user friendliness in mind making it easy to understand and use. The possibilities to interact with the database for a normal user are low but an experienced IT user can access the underlying database directly for making changes or to reach data sources. The asset management module is hard to navigate without the use of personalised views. It is build with flexibility in mind and provides higher grade of changeability than the CMMS interface. The database is not directly accessible without special software though, giving low possibilities to interact with the data source.
5.4 Results of the case study
The CMMS covers parts of the basic maintenance related technical and economic data, but lacks production related data. The asset management module contains maintenance specific data as well as data about e.g. production and quality, supporting a more holistic view of maintenance as a concept. None of the databases studied could provide the data needed to assess the two performance measurements chosen, DCQ and NPD. The problem may be much bigger if we consider other performance measures such as overall equipment effectiveness (OEE) and impact of better process effectiveness on maintenance cost, see Al-Najjar et al. (2004). Further, the data found in the databases were in some cases aggregated or confused with other data in the example measures of DCQ and NPD. If a more accurate assessment of e.g.
spare parts costs were made, we would include damage costs due to bad transport and inventory handling expenses. The current structures of the databases are not allowing such a level of detail, which means that full tracing back to root data sources is not possible. Even if data would be available, the applications are not prepared to assess DCQ and NPD or similar maintenance performance measures automatically. The procedure to manually localise and collect data and then assess performance measures would be time consuming for a maintenance engineer to perform. The CMMS has a restricted area of application focusing on operational maintenance management. This combined with restricted data content makes it hard to use the CMMS alone as basis for monitoring production and maintenance and follow up their development in real time. The strength of the CMMS is its relative low complexity and high flexibility allowing experienced users to access and change the data content. The asset management module is data intensive with the positive effect that a more accurate description of maintenance performance can be reached. The size of the ERP database and the standardised, not maintenance specific, terminology makes the database structure inflexible and complex though.
6. Results and conclusions
The most important result that has been achieved in this study is the development of a model for identifying relevant data required for accurate problem tracing and localisation within maintenance and production processes using a top down approach. The top down approach ensures that decisions made on the operative activity level are considering the corporate strategic goals. Real time measurements, or at least proper measuring frequency of the relevant variables are required for effective mapping, monitoring, control and evaluation of the production process. This makes the detection of the deviations from the company's business goals at an early stage without big financial losses easier. In most cases, the data planned to be gathered cannot be used effectively and clearly for linking the technical inputs with financial outputs.
Most industries use standardised software for maintenance management such as the situation in the case study discussed by this paper. This often results in inflexibility due to enforced information standardisation and centralisation and data contents that either will be huge (and not necessarily relevant) or too restricted with respect to the real need of the different working areas. Further, the data provided in
the both cases are not identified with respect to the company’s strategic objectives.
By applying the model of top down analysis for identifying maintenance relevant data, the data identification process itself and not the data contents is necessary to be standardised and structured. The model is data source insensitive. It shifts the focus of the quality aspect from just data level to both data and data collection level. The performance measures will therefore not be chosen depending on what the IT applications can provide in first hand, but upon what is needed for cost-effective assessment, monitoring and following up maintenance performance. If the data collection and analysis process is determined by applying the strategic corporate goals that are decomposed, e.g. into technical, financial and organisational goals, the understanding of what, when and why to measure will grow effectively and continuously. This will also lead to better understanding of the performance measures used and the types of data that are needed as input. If data are lacking, of poor quality or not accessible in the correct level of details, alternative ways of assessing the performance measures must be found. To be able to do this, integration of both technical, financial and administrative IT resources is needed, but to avoid rigid and
centralised information structure standardisation, the integration would be on data level instead of on IT system level. By combining the database content of technical and administrative IT systems an almost full coverage of data could be achieved. This distributed database connected to a flexible maintenance decision support tool would provide support for cost-effective maintenance decisions that fulfils company's strategic goals for higher profitability and better competitiveness.
The main conclusions drown are; integration of IT and data resources within the enterprise is needed for developing a holistic view of the production process. A well-formulated and documented procedure of data identification will furthermore ensure that the data can be traced back to root sources and in this way we can support the work of continuous cost-effective improvement by eliminating root causes of problems at an early stage.
Acknowledgements
We would like to acknowledge Centre of Industrial Competitiveness (CIC) who is the financier of this project. The CIC is an interdisciplinary research platform on Växjö University with the basis in both economics and technology.
References
Al-Najjar, Basim. (1996) Total quality maintenance. An approach for continuous reduction in costs of quality products. Journal of Quality in Maintenance Engineering 2(3), 4-20.
Al-Najjar, Basim. (1997) Condition-based maintenance: Selection and improvement of a cost-effective vibration-based policy for rolling element bearings. Lund University, Department of Industrial Engineering.
Al-Najjar, Basim. (1999) Economic criteria to select a cost-effective maintenance policy. Journal of Quality in Maintenance Engineering 5(3), 236-247.
Al-Najjar, B., Alsyouf, I. (2004) Enhancing a company's profitability and competitiveness using integrated vibration-based maintenance: A case study. European Journal of Operational Research 157(3), 643-658.
Al-Najjar, B., Hansson, M-O. and Sunnegårdh, P. (2004) Benchmarking of Maintenance Performance: A Case Study in two manufacturers of furniture. IMA Journal of Management Mathematics 15, 253-270.
Al-Najjar, B., Alsyouf, I., Salgado, E., Khoshaba, S. and Faaborg, K. (2001) The economic importance of maintenance planning when using vibration-based maintenance. Växjö University Department of Terotechnology.
Alsyouf, Imad. (2004) Cost effective maintenance for competitive advantages. Växjö university, Department of Terotechnology.
Bennet, K., Layzell, P., Breneton, P., Macaulay, L. and Munro, M. (2000) Service-Based Software: The Future for Flexible Software. Proceedings of Seventh Asia-Pacific Software Engineering Conference, 214-221.
Blanchard, B., S. (1998) Logistics engineering and management. Upper Saddle River, Prentice-Hall.
Boucelma, O., Castano, S., Goble, C., Josifovski, V., Lacriox, Z., Ludascher, B. (2002) Report on the EDBT’02 Panel on Scientific Data integration. SIGMOD Record 31(4), 107-112.
Bratthall, L.G., van der Geest, R., Hofmann, H., Jellum, E., Korendo, Z., Martinez, R., Orkisz, M., Zeidler, C., Andersson, J. S. (2002) Integrating hundred's of products through one architecture - the industrial IT architecture, Proceedings of the 24rd International Conference on Software Engineering, 2002, 604-614.
Chattopadhyay, S. P. (2001) Relationship marketing in an enterprise resource planning environment, Marketing Intelligence & Planning 19(2), 136-139.
Chi, Y.-L., Wolfe, P. M. (1999) Bring Collaboration to Enterprise Application Using
a COM/CORBA Bridge, Proceedings of the 25th International Conference on
Computers and Industrial Engineering- New Orleans, LA., 335-338.
Curtice, R. M. (1981) Fundamentals of data element definition. Proceedings of the 1981 ACM SIGMOD international conference on Management of data, pp. 49-55.
Dahlbom, Bo. (1997) The New Informatics, Scandinavian Journal of Information Systems 8(2), 29-48.
Dahlfors, F., Pilling, J. (1995) Integrated information systems in a privatized and deregulated electricity market, Proceedings of 1995 International Conference on Energy Management and Power Delivery 1, 249-254.
Dedrick, J., Gurbaxani, V., Kraemer, K. L. (2003) Information Technology and Economic Performance: A Critical Review of the Empirical Evidence, ACM Computing Surveys 35(1), 1-28.
Dwight, R. (1999) Searching for real maintenance performance measures. Journal of quality in Maintenance Engineering 5(3), 258-275.
Elmasri, R., Navathe, S. B. (2003) Fundamentals of database systems, Pearson/Addison-Wesley, Boston.
Evgeniou, Theodoros. (2002) Information Integration and Information Strategies for Adaptive Enterprises. European Management Journal 20(5), 486-494.
Fisher, C.W. ; Kingma, B.R. (2001) Criticality of data quality as exemplified in two disasters, Information and Management 39(2), 109-116.
Flensburg, P., Friis, S. (1999) Mänskligare datasystem : utveckling, användning och principer, Studentlitteratur, Lund.
Gordon, L. A. (2000) The e-skip-gen effect. The emergence of a cybercentric management model and the F2B market segment for industry. International Journal of Production Economics 80(1), 11-29.
Herzog, U., Hantikainen, R., Buysse, M. (2002) A real world innovative concept for plant information integration, IEEE-IAS/PCA 44th Cement Industry Technical Conference, May 2002, Jacksonville, Florida.
Hoffer, J. A., Prescott, M. B., McFadden, F. R., Modern Database Management 6th ed., Prentice Hall, USA, 2002.
Holland, C.P., Light, B. (1999) Global enterprise resource planning implementation. Proceedings of the 32nd Annual Hawaii International Conference on Systems Sciences, 10 pp.
Jonsson, Patrik. (1999) The Impact of Maintenance on the Production Process-Achieving High Performance. Lund University Institute of Technology Department of Industrial Engineering Division of Production Management.
Kang, C. W., Golay, M. W. (2000) An integrated method for comprehensive sensor network development in complex power plant systems. Reliability Engineering and System Safety 67(1), 17-27.
Kelly, G. J., Aouad, G., Rezgui, Y., Crofts, J. (1997) Information systems developments in the UK construction industry, Automation in Construction 6, 17-22.
Kini, R. B. (2002) IT in manufacturing for performance: the challenge for Tai manufacturers. Information Management & Computer Security 10(1), 41-48.
Kjaer, A. P. (2003) The integration of business and production processes, IEEE Control Systems Magazine 23(6), 50-58.
Lenzerini, M. (2002) Data integration: A Theorethical Perspective. Proceedings of the ACM SIGACT-SIGMOD-SIGART Symposium on Principles of Database Systems, 233-246.
Li, Fangxing, Brown, Richard E. (2004) A Cost-Effective Approach of Prioritizing Distribution Maintenance Based on System Reliability. IEEE Transactions on Power Delivery 19(1), 439 – 441.
Long, E., Misra, A., Sztipanovits, J., (1998) Increasing Productivity at Saturn, Computer 31(8), 35-43.
Miraglia S., Blouin C., Boldman G., Judd S., Richardson T and Yao D. (2002) ABC modeling: advanced features. IEEE/SEMI Conference and Workshop Advanced Semiconductor Manufacturing, 336-339.
Mullin, A., (1989) The application of information technology to asset preservation in Ford of Europe. IEE Colloquium on IT in the Management of Maintenance, 2/1-2/4.
Murthy, D. N. P., Atrens, A., Eccleston, J. A., (2002). Strategic maintenance management. Journal of Quality in Maintenance Engineering 8(4), 287-305.
Nakkiran N. Sunassee, David A. Sewry (2003) An Investigation of Knowledge Management Implementation Strategies, Proceedings of SAICSIT 2003, 24 – 36. Nikolopoulos, K., Metaxiotis, K., Lekatis, N., Assimakopoulos, V. (2003) Integrating
industrial maintenance strategy into ERP, Industrial Management and Data systems 103(3), 184-191.
Otto, A., Kotzab, H. (2003) Does supply chain management really pay? Six perspectives to measure the performance of managing a supply chain. European Journal of Operational Research 144(2), 306-320.
Persson, P.-O., Boberg, K.-E., Broms, I., Docherty, P., Kraulis, G., Kreimer, B. (1981) 80-talet på en bricka : datateknik; utveckling och miljö 1980-1990 : en framtidsstudie om datorer och en ADB-värld i ständig utveckling, Riksdataförbundet, Stockholm.
Pintelon, L., Van Puyvelde, F. (1997) Maintenance performance reporting systems: some experiences. Journal of Quality in Maintenance Engineering 3(1), 4-15.
Pintelon, L., Preez, N. D., Van Puyvelde, F. (1999) Information technology: opportunities for maintenance management, Journal of Quality in Maintenance Engineering 5(1), 9-24.
Riis, J. O., Luxhoy, J. T., Thorsteinsson, U. (1996) A situational maintenance model. International journal of quality and Reliability Management 14(4), 349-366.
Sarshar, M., Tanyer, A. M, Aouad, G., Underwood, J. (2002) A vision for construction IT 2005-2010: two case studies, Engineering Construction & Architectural Management 9(2), 152-160.
Segarra, G. (1999) The advanced information technology innovation roadmap. Computers in Industry 40(2-3), 185-195.
Swanson, Laura, (2003) An information-processing model of maintenance management. International Journal of Production Economics 83(1), 45-64.
Szirbik, N. B., Jagdev, H. S. (2001) The future IT systems for virtual enterprises: product-oriented agent providers?, 8th IEEE International Conference on Emerging Techniologies and Factory Automation 2, 261-270.
Tsang, A. H. C., (2002). Strategic dimensions of maintenance management. Journal of Quality in Maintenance Engineering 8(1), 7-39.
Tuunainen, V. K. (1998) Opportunities of effective integration of EDI for small businesses in the automotive industry, Information and Management 34(6), 361-375.
Vojdani, A. F. (2003) Tools for real-time business integration and collaboration. IEEE Transactions on Power Systems 18(2), 555-562.
Weber, Y., Pliskin, N. (1996) The effects of information system integration and organizational culture on a firm's effectiveness. Information and Management 30(2), 81-90.
Whittaker, Brenda. (1999) What went wrong? Unsuccessful information technology projects. Information Management & Computer Security 7(1), 23-29.
Xu, Hong, Wang, JianHua, Zheng, ShiQuan (2003) The distributed data integration and performance evaluation in power automation system, Proceedings of the Fourth International Conference on Parallel and Distributed Computing, Applications and Technologies, 2003, 468-471.
