KEY DATA QUALITY ISSUES FOR ENTERPRISE ASSET
MANAGEMENT IN ENGINEERING ORGANISATIONS
Shien Lin
*, Jing Gao and Andy Koronios
School of Computer and Information Science
University of South Australia
Mawson Lakes (5095), Australia
ABSTRACT
Data Quality (DQ) is a critical issue for effective asset management. DQ problems can result in severe negative consequences for an organisation. Several research studies have indicated that most organizations have DQ problems. In response to these problems, organisations have developed various policies (e.g. fair use of data, information privacy, etc). However, it is often difficult to implement these policies at the tacit and operational level, especially identifying the areas or problems to focus on. Through a case study with two state-wide Australian utility companies, this paper tries to present the DQ issues which emerged from a number of asset management processes. The findings are thought to be valuable for practitioners, especially policy implementers, who want to align policy with real-world business.
Keywords: Data Quality, Enterprise Asset Management, Engineering Asset Management
1. INTRODUCTION
*
Australia’s economic and social viability is critically dependent on the capacity of the major physical assets which underpin the industrial, social and economic environments, to function soundly and to meet quality performance management standards. These major physical assets include aircraft, shipping, railways (both in terms of rolling stock and track), defense, roads, bridges, buildings (both commercial and domestic), industrial processing equipment, manufacturing plant, oil and gas extraction and refining plant and equipment, power generation plant, power distribution systems, water storage and distribution facilities. Security of the services provided by these industries is fundamental to the lifestyle and economy as well as the health of the social and environmental fabric.
To address the concerns within various types of asset management, the Australian government has decided to establish a Cooperative Research Centre for researching into Integrated Engineering Asset Management (CIEAM) in order to facilitate the implementation of innovative technologies, processes and programs in a strategic, integrated framework that optimises asset management systems and ensures that Australian industry is globally competitive [6]. Systems Integration and IT, one of the five CIEAM research programs, aims at developing a suite of tools
*
Corresponding author: [email protected]
and standards for data exchange between the asset-based technical systems and management information systems, and supporting strategic decision-making asset management tools at the corporate level. This research was conducted within the program of Systems Integration and IT through CIEAM.
As Asset Management (AM) is a broad area, this paper places a special focus on data quality (DQ) issues. In order to address the various aspects of DQ issues in enterprise asset management (EAM), the generic data quality concepts need to be summarized (see section 2). Similarly, an overview of generic asset management processes is provided (see section 3). Within these understandings, a close examination is conducted to search for asset management related DQ issues (see section 4). A research framework is then developed for conducting empirical research (by case studies). Based on the preliminary findings, an insightful understanding of various AM related DQ issues will be extracted. These findings will be useful to determine the problem areas that policy implementers need to focus on.
2. DATA QUALITY
Numerous researchers have attempted to define data quality and to identify its dimensions [15,20,26,27,29,32,45,51,53,55,56]. Dimensions of data quality typically include accuracy, reliability, importance, consistency, precision, timeliness, fineness, understandability, conciseness, and
usefulness [2,11,37,40,53]. Wand and Wang [53] used ontological concepts to define data quality dimensions: completeness, unambiguousness, meaningfulness, and correctness. Wang and Strong [53] categorize data quality into four dimensions: intrinsic, contextual, representation, and accessibility. Shanks and Darke [45] used semiotic theory to divide data quality into four levels: syntactic, semantic, pragmatic, and social. Recently, Kahn et al. [27] used product and service quality theory to categorize information quality into four categories: sound, useful, dependable, and usable.
From the literature, it appears that there is no universal definition of data quality; neither is there an agreed framework of DQ dimensions. The four most obvious data quality dimensions are accuracy, relevance, fineness, and timeliness [2,23,57]. These dimensions are inter-related and may even be in conflict with each other. For example, the improvement of accuracy is often associated with a cost of timeliness [2]. Moreover, different stakeholders in an organisation may have different data quality requirements and concerns [17]. For example, the data quality requirements often vary across the different levels of decision-making.
Nevertheless, there is a need for selecting a widely-accepted data quality dimension as the starting point of this research. It is believed that Wang and Strong [56]’s definition of data quality “quality data are data that are fit for use by the data consumer” is suitable.
Maintaining the quality of asset data is often acknowledged as problematic [9,18,59], but critical to effective asset management. Examples of the many factors that can impede data quality are identified within various elements of the data quality literature. These include: inadequate management structures for ensuring complete, timely and accurate reporting of data; inadequate rules, training, and procedural guidelines for those involved in data collection; fragmentation and inconsistencies among the services associated with data collection; and the requirement for new management methods which utilize accurate and relevant data to support the dynamic management environment.
Clearly, personnel management and organizational factors, as well as effective technological mechanisms, affect the ability to maintain data quality. Wang et al. [57] clarify this relationship by drawing an analogy between manufacturing and the production of data. In this way they derive a hierarchy, adapted from international ISO 8402 product quality standards, of responsibilities for data quality, ranging from management down to the individual procedures and mechanisms. Their framework specifies a top management role for data quality policy, i.e. overall intention and direction relating to data quality, and a
data quality management function to determine how that policy is to be implemented. This, in turn, should result in a data quality system for implementing data quality management, within which data quality control is enforced through operational techniques and activities. Data quality assurance then comprises all of the planned and systematic actions required to provide confidence that data meet the quality requirements.
3. ASSET MANAGEMENT
Asset management means the management of the plant and equipment during its whole life; that is, from specification through manufacturing, commissioning, useful life, maintenance, and then managing the consequences from the decision to refurbish or replace the item before final decommissioning and recycling any components. Asset management is a structured program to optimize the life-cycle value of the physical assets by reducing the cost of ownership, while providing the required level of service. The objective of an asset management system is to minimize the long-term cost of owning, operating, maintaining, and replacing the asset, while ensuring reliable and uninterrupted delivery of quality service [10,47]. At its core, asset management seeks to manage the facility’s asset from before it is operationally activated until long after it has been deactivated. This is because, in addition to managing the present and active asset, asset management also addresses planning and historical requirements [48].
The process of asset management is sophisticated because it is an engineering and planning process that requires substantial information to be collected from many different parts of the organization [25]. This information must be maintained for many years in order to identify long-term trends. The asset management engineering and planning process uses this information to plan and schedule asset maintenance, rehabilitation, and replacement activities. The information management system that captures, maintains, and provides the needed asset information is critical in providing effective asset management [46].
There are a number of AM processes involved in AM. For example, Figure 1 shows that asset management has a number of unique processes, which have not been previously discussed. Further work can be found from Woodhouse [60,61].
Sandberg [41] and Paiva, Roth and Fensterseifer [38] argue that contemporary asset management demands an elevated ability and knowledge to continually support the asset management process, in terms of data acquisition, real-time monitoring, and computer supported categorization and recording of
divergences from standard operations. Figure 2 illustrates an overview of the process of asset management and the information flow associated with the process.
Figure 3 illustrates an overview of the integrated engineering asset management framework and indicates the need to establish enterprise-wide asset management information systems.
4. DATA QUALITY AND ASSET
MANAGEMENT
Asset data quality issues have become increasingly prevalent in practice, and the evidence indicates that most organizations, particularly engineering enterprises, have experienced some level
of DQ problems [9,10,13,22,59]. Data quality problems can have significant business impacts. Data quality is an overlooked issue within most large enterprises [13,34]. Expensive and time-consuming data integration initiatives often fail as a result of a lack of awareness or lack of focus on flaws in the source data [21].
Poor data quality in individual operational systems may have a large, negative effect on the enterprise, including:
• Lost productivity and higher error rates for routine tactical tasks.
• The inability to deliver results from initiatives such as straight-through processing or business activity monitoring.
Figure 1: Asset management processes [60]
Figure 3: Integrated asset management framework [7] Data quality management can be regarded as a
business issue. This requires business leaders to have a high degree of awareness and understanding of the effects of poor data quality on business results and key strategic initiatives [24,28]. Most large enterprises have examples of project failures due to poor data quality. Making these visible to the management team allows the rest of the organization to learn from such experiences, in order to improve. Building on awareness, the enterprise must follow a structured approach (see Figure 1) to ensure that business and IT resources are focused on achieving and sustaining improvements in data quality.
There is a growing need to address the DQ issues in engineering asset management systems, by analyzing existing practices and developing models to assist engineering enterprises to capture, process and deliver quality information [49]. It is essential to ensure the quality of data captured from the monitoring systems, used for control systems, maintenance systems, procurement systems, logistics systems, and range of mission support applications for facilitating asset availability, readiness, reliability, effectiveness, and management [35].
However, asset management is not considered as a core business activity by many businesses and therefore they depend on the traditional organizational information sources to manage assets.
These traditional sources represent the tacit, implicit knowledge of engineers, operators, and information contained in information systems, which have been primarily designed to increase productivity rather than improving the efficiency of the processes involved in production. At the same time, a variety of operational and administrative systems exist in asset management, which not only manage the operation of asset equipment but also provide maintenance support throughout the entire asset lifecycle. In practice, data is collected both electronically and manually, in a variety of formats, processed in isolation, stored in a variety of legacy systems, shared among various operational and administrative systems, and communicated through a range of sources to assorted business stakeholders. Data captured and processed by these systems is not comprehensive; it is process dependent, making it difficult to be reused for any other processes [19].
Due to the multiplicity of systems and stakeholders, as well as the level of unpredictability in asset operation within asset management, it is impossible to tap user requirements, which consequently contributes to the ‘dirtiness’ of asset data. In managing physical assets through the entire asset life cycle, large amounts of data are needed for long term performance and reliability prediction, as well as informing the decision making process on
when to retire an asset. Although very large amounts of data are being generated from asset-condition monitoring-systems, little thought has been given to the quality of such generated data. Thus the quality of data from such systems may suffer from severe limitations [43].
It appears that there has been little discussion in the literature on managing DQ for asset management. Therefore, it is necessary to develop a model to enhance the capability of DQ in asset management. To address this problem, an asset management specific DQ framework will first need to be developed by conducting a pilot case study to identify unique DQ issues in asset management. This paper aims to present the preliminary findings from the pilot case study. The findings will be further used, together with the available literature, to build the proposed DQ model.
5. DATA QUALITY RESEARCH
FRAMEWORK FOR ASSET
MANAGEMENT
Mitroff and Linstone [36] argue that any phenomenon, subsystem or system needs to be analyzed from what they call a Multiple Perspective method – employing different ways of seeing, to seek perspectives on the problem. These different ways of seeing are demonstrated in Linstone [33], Mitroff and Linstone [36] and Werhane [58] through their TOP model. This model allows analysts to look at the problem context from either Technical or Organizational or Personal points of view:
• The technical perspective (T) sees organizations as hierarchical structures or networks of interrelationships between individuals, groups, organizations and systems
• The organisational perspective (O) considers an organization’s performance in terms of effectiveness and efficiencies. For example, leadership is one concern.
• The personal perspective (P) focuses on the individual concerns. For example, the issues of job description and job security are some of the main concerns in this perspective.
Therefore, based on the processes of asset management (as illustrated above – Figure 1, Figure 2 and Figure 3) and the TOP perspectives, a research model is needed in order to guide the process of exploring the data quality issues emerging from the modern approach (EAM Information Systems facilitated) to asset management.
Based on the discussion on various data quality issues [16,57,62,63,64] and the unique requirements of asset management, the data quality research
framework for asset management was developed as shown in Figure 4, using the TOP perspectives theory.
There is strong evidence that most organisations have far more data than they possibly use; yet, at the same time, they do not have the data they really need [31]. Therefore, identification of business needs is essential. The TDQM (Total Data Quality Management) methodology [27,55] emphasized the importance of continuous improvement and delivery of high-quality information products. It demonstrated that defining requirements, measuring, analysing, and monitoring improvement are important processes within the information manufacturing system. The defining component of the TDQM cycle identifies important DQ dimensions and corresponding DQ requirements. The measurement component produces DQ metrics. The analysis component identifies the root causes for DQ problems and calculates the impact of poor-quality information. Finally, the improvement component provides techniques for improving DQ.
As data is not static, it is a dynamic, fluid resource. It flows in a data collection and usage process. The DQ problems that may arise at each stage are different, and require different metrics as well as solutions [8]. Because of the continuum nature of data, DQ is not a one-time, fix-it-and-forget-it practice [14]. Building and keeping good quality corporate data takes constant vigilance and feedbacks in the context of the entire data life cycle. Feedback loop is an important DQ monitoring tool. The application of strategic feedback loop can serve to ensure business needs are up-to-date.
This framework is useful to guide the research into data quality issues in asset management, because it highlights the three root perspectives (TOP) on data quality problems, illustrates how they emerge during the process of asset management; and in particular, outlines the four basic data quality measurement criteria.
Although the key factors for high data quality in asset management have not been addressed, there have been many studies of key factors in quality management such as Total Quality Management (TQM) and Just-in-Time (JIT) [1,5,39,42,66]. Some of the data quality literature has addressed the key points and steps for DQ [12,22,23,30,44,52]. Regarding asset management processes, the DQ factors, as shown in Figure 5, are summarised from the literature review of DQ related research efforts in order to develop a semi-structured research protocol to be used in interviews aiming to understand the DQ issues emerging from individual processes.
Figure 4: Data quality research framework for asset management
6. RESEARCH METHOD
For this research, two large Australian organisations were selected as case sites. Case study research is used to study the contemporary phenomenon in its real-life context [65] and it can be used where the research and theory are at their early, formative stages [3]. As data quality for asset management has received little attention in the research community, there is a need to examine whether data quality is a critical issue to any asset management programme and what the key factors are to ensure data quality in managing assets. Therefore, case study research appeared to be appropriate for this study.
In data quality studies, four types of stakeholders have been identified: data producers, data custodians, data consumers and data managers [50,54].
To apply this stakeholder’s concept in an asset management environment:
1. Data producers are those who create or collect data for asset management systems.
2. Data custodians are those who design, develop and operate the asset management system. 3. Data consumers are those who use the asset
information in their work activities.
4. Data managers are those who are responsible for managing data quality for asset management systems.
Data was collected using semi-structured interviews with key stakeholders of asset management systems. In total, 31 interviews (15 and 16 respectively at each site) were conducted with contractors, employees and managers at all levels from the participating organisation within the field of asset management. Each interview included questions about the background of the organization, as well as its asset management practice, the participants’ roles, and their views about data quality issues in managing assets. Additional information was obtained from secondary data including reports and internal and external documents. The purpose of the case study was to investigate what was actually happening in the real-world organisation in relation to data quality issues associated with the implementation of EAM systems within engineering enterprises.
Through a number of interviews, new issues were raised. These issues were recorded and were rephrased as questions during the follow-up interviews. It is believed that, after a number of interviews, the majority of all the important issues have been covered. Moreover, in order to ensure validity, the issues have been raised with different people (at different levels at the same organisation, and same level but different company) for cross-checking.
Technology
Organization
People
-DQ control & improvement approaches and techniques -Data acquisition architectures, tools, systems
-DQ interoperability standards -Application and process integration -DQ techniques for data integration -Data storage architecture
-Data cleansing techniques
-Top management's commitment to DQ
-Appropriate DQ policies, standards and implementation -Role of DQ managers
-Organisational structure -Organisational culture
-Information supplier's quality management -Customer focus
-Audit and reviews
-Evaluation of cost/benefit tradeoffs -Teamwork communication -Change management -Internal control system -Input control
-DQ feedback
-Personnel competency
-Performance evaluation and rewards -Employee relations
-Management's responsibility
Data Quality
Figure 5: Summary of factors influencing data quality
7. PRELIMINARY RESEARCH
FINDINGS
Due to the limitation on this paper’s length, only the key findings are presented below.
7.1 Technology Perspective
7.1.1 Integration of AM-related Technical Systems as well as Integration between Business Systems and Technical Systems
A variety of specialized AM-related technical systems exist and include reliability assessment systems, asset capacity forecasting systems, enterprise asset maintenance systems, electrical motor testing systems, turbo-machinery safety systems, rotating machine vibration condition monitoring systems, operational data historians, root cause analysis systems, and physical asset data warehouse systems. Such specialized systems are acquired from multiple vendors and as they are quite disparate, they often lead significant integration problems.
There appears to be little cognizance when adopting business systems such as financial, human resource and inventory information systems of the need to ensure compatibility with technical systems such asset register systems, work order management systems and condition monitoring systems. Most
users are unable to translate the vast amounts of available asset data into meaningful management information to optimize their operation and control the total asset base. This has led to the notion of ‘islands of information’.
Such disconnects make it extremely difficult to bring real-time information from the plant into business systems. There are disconnects between the transaction-driven, product-centric business data environment and the continuous data, process-centric open control system and manufacturing data environments. The lack of process-to-product data transformation capabilities in linking business systems and plant floor EAM applications have significant data quality consequences and thus negatively affect data-driven decision-making.
7.1.2 Asset Hierarchy
The objective of developing an asset hierarchy is to provide a suitable framework for assets, which segments an asset base into appropriate classifications. The asset hierarchy can be based on asset function, asset type or a combination.
The intent of the asset hierarchy is to provide the business with the framework in which data are collected, information is reported, and decisions are made. In most cases (as found in one organization) organizations work with an informal asset hierarchy.
This often leads to data being collected to inappropriate levels, either creating situations where costs escalate with minimal increases in benefits, or insufficient information is available to make informed decisions.
Infrastructure assets generally have a clear hierarchical relationship that breaks down from the asset type as a whole to large units (facilities), then to assets and their components. The information needs of the organization vary throughout the management structure. At the workface the key elements are operations, maintenance, and resource management, at a component level. At higher management levels this information needs to be aggregated to provide details on assets, facilities and (infrastructure) systems as a whole in terms of finance, strategic and policy.
7.1.3 Data Access
Designers and asset manufacturers represent the external source of asset data. As part of the asset acquisition process, all asset information required to own and operate the asset should be handed over to the user organisation at the commissioning of the asset, in a form that can be assimilated readily into the user organization’s asset information systems. These asset data may include a fully-fledged technical information database with Geographical Information System (GIS) maps, technical specifications, and even video clips of the equipment and its operation.
The research has found that a data gap may exist between the maker and the user of asset equipment. The user organization needs to populate the EAM with data from the manufacturer — particularly the component structure and spare parts. These capabilities exist in manufacturers’ product data management (PDM) and product lifecycle management (PLM) systems. Unless arrangements or contract conditions are made, in many cases, the data is not passed on to the buyer in a usable electronic format. In some cases, the asset data handed over to the user organisation does not conform to the physical assets delivered. In other cases, updated asset data, particularly the component structure, may not always be passed on to the user organisation.
Information such as job instructions, maintenance cycles and advisory notices is also available. However, without standards and interfaces to share this information across systems, it is often held offline either as paper documents or poorly linked electronic copies of instructions.
7.1.4 Database Synchronization
The capability of EAM systems can be enhanced through a link with GIS to provide the ability to access, use, display, and manage spatial data. The ability to effectively use spatial asset data is
important for utilities with geographically dispersed utility networks. However, it was found that one of the most critical activities is to establish synchronization between the two database environments. One asset manager indicated that there has been an issue existed for overcoming the synchronization of asset register in a very common work management system with GIS in the company. Both automated and manual processes needed to be defined and implemented to maintain synchronization between the GIS and EAM databases. Database triggers and stored procedures need to be defined to automate the attribute update process maintaining synchronization between the GIS and EAM databases. Workflows and business rules must be developed for GIS and EAM data editing, to ensure synchronization from both applications.
7.1.5 Data Exchange
Both of the organizations indicated that there is a need for data exchange between asset management applications for seamless access to information across heterogeneous systems and different departments within an engineering enterprise. The life cycle performance data of the various assets are kept in individual uncoordinated databases, which make inter-process, inter-functional analysis very difficult.
Moreover, the various computer software programs designed for condition monitoring and diagnostics of machines, that are currently in use within the two organizations, cannot easily exchange data or operate in plug-and-play fashion without an extensive integration effort. For example, the maintenance work event forms, which contain the hours that a contractor has worked, cannot be directly input to the payroll system as an alternative to the traditional timesheets. This makes it difficult to integrate systems and provide a unified view of the condition of assets to users.
7.1.6 Data Standards for Condition Monitoring Systems
Although it appears that condition monitoring equipment and systems are proliferating, an apparent lack of dialogue among vendors (as found in one organization) has led to incompatibilities among hardware, software and instrumentation. Data collected by current outdated equipment could become obsolete and inaccessible to new, upgraded systems. To fully realize the integration of systems over the various levels of asset maintenance and management, new standards and protocols are needed. A focus on the standardization of condition monitoring data modeling and exchange tools and methodologies, such as Standard for the Exchange of Product model data (STEP) is critical.
Moreover, facility managers in one organization indicate that there is a need to be able to accurately
assess asset condition with the capacity to comprehensively record key asset condition data, as a fundamental objective of the inspection and maintenance policy. Whilst comprehensive asset databases generally exist, not all of them have been designed to accurately record asset condition data. For example, one plant records system holds data on all items of operational plant in terms of age, location, and date last inspected/maintained, but it does not provide the means to hold data on the condition of the plant. Where a comprehensive asset database exists containing both asset details and recorded defects, the database however does not permit the recording of successive inspection data (any new record overwriting the current record) and so the important function of developing generic aging curves is not supported.
7.1.7 Sensor Calibration and Integrity Check for Condition Monitoring Process
Interviews with asset maintenance field workers indicate that data captured by intelligent sensors may not always be accurate. Data capturing devices typically used in condition monitoring are electronic sensors or transducers, which convert numerous types of mechanical behavior into proportional electronic signals, usually voltage-sensitive signals, producing analog signals which in turn are processed in a number of ways using various electronic instruments. As signals are generally very weak, a charge amplifier is connected to the sensor or transducer to minimize noise interference and prevent signal loss. The amplified analogue signal can then be sent via coaxial cables to filtering devices to remove or reduce noise, before being routed to a signal conditioner and/or Analogue-to-Digital converters for digital storage and analysis. To ensure the data received by the SCADA system conforms to the original signal data captured by sensors, integrity checks for signal transmission process and sensor calibration need to be performed and maintained. However, as the sensor calibration and integrity checks are often neglected in asset maintenance in most industries, the extent to which acquired data is correct and reliable was shown to be of concern with respondents.
7.1.8 Reliability Data Systems 7.1.8.1 Data Collection Process
The data collection process is often time consuming and many organisations have concluded that the information they get is not worth the money they put into it. One additional problem here is often that the operator of the reliability database has to get raw data from a computerised maintenance management system. The transfer of data between the two databases must often be done manually because they were developed independently. An example that
shows one possible way ahead is that the maintenance systems and the reliability data systems are built from the start to co-exist in order to minimise the manual effort to collect information on failure.
Other problems are the transfer of operational data from the main plant computer and the evaluation of correct estimates from the plant production records. The quality of a specific reliability data bank may depend on the quality of these operational data, the transfer software or the estimation model. Particular attention must be drawn to those aspects. And, as the data recorder is not the data user, the data collection process could be significantly improved by motivating the data recorder with a feedback on what has been done with the data.
7.1.8.2 Coding of Information
The coding of information and the terminology can differ from database to database. This makes it difficult to compare data from different sources. The solution is standardisation to the extent possible. The standardisation should cover topics like equipment boundaries, definition of failure modes, definition of failure causes, downtime versus active repair time, the need for spare parts, recording of preventive maintenance, recording of other parameters that are important for equipment reliability and even the definition of “failure”. It may be that standardisation can best be carried out for different types of industries.
7.1.8.3 Information Technology
The information technology has now been developed to a stage where it creates no or only minor problems to the database operators. The main issue in setting up a reliability database is the specification of the content of the database. This specification strongly depends on the objectives of the database: safety, availability or production, cost, maintenance, in service inspection, material, expertise, human factors… These objectives will define the structure and the level of detail of the database, and the precise boundaries of what is followed: plant breakdown, system, functional group, equipment, subcomponent, control, preventive and corrective actions… It is preferable to limit objectives. Too much information kills information (and quality).
7.1.8.4 Extraction of Data
The extraction of data from the database is still a major problem. How to overcome this is an area for further research. Different users of reliability databases have different needs, and it is difficult to set up a database that covers all needs. It may be a solution to tailor the database to a specific user. Examples of users that need information on different
levels are designers who need reliability index allocation to define a plant or a specific system, procedures who want to improve their product, the safety engineer who gives advice on the level of redundancy in safety systems, the reliability engineer who wants to optimise system configuration with respect to production availability, and the maintenance engineer who wants to optimise preventive maintenance.
7.1.8.5 Quality Assurance
Quality assurance of reliability database is often poor. A structured and formalised standard, or a documented guideline should be developed on this subject and on the subject of cross control between coded information and free text analysis. The importance of parameters that influence the reliability of a component is not always well understood. This makes it difficult to record the information that is important for the reliability. This problem is related to the problem of how to extract data from the databases and needs further research.
7.1.8.6 Expert Judgements
Expert judgements are advocated as an important part of data analyses. They often are the only available data or the only possibility to check or to extrapolate past field data polluted by preventive maintenance, aging, equipment modifications or specific environment. Some work is going on in Europe aiming to formalise the process of selection of experts, the interrogation phase, the calibration of experts, the weighting and the aggregation of answers, in such a way that expertise could be in the future considered as good as technical data.
7.2 Organizational Perspective 7.2.1 Organizational Readiness
Many companies that attempt to implement EAM systems run into difficulty because they are not ready for integration and the various departments within it have their own agendas and objectives that conflict with each other. Organizational readiness can be described as having the right people, focused on the right things, at the right time, with the right tools, performing the right work, with the right attitude, creating the right results. It is a reflection of the organization’s culture. EAM implementations involve broad organisational transformation processes, with significant implications to the organisation’s asset management model, organisation structure, management style and culture, and particularly, to its people.
An EAM implementation project within the utility organisation is expected to have a high acceptance of the system in areas that provide just as good or better functionality than the old system. However some functions and processes did not get
the full appreciation the legacy systems once had. Through interviews with technicians and data collectors, it was found that field workers are frustrated with the need to use Maximo [an asset maintenance work order management system] and are losing confidence in it. One staff member said that, “…with Maximo there are so many problems, people are not interested”. Another interviewee said that, “Maximo hasn’t solved the speed problem which you would have thought it would have….. From day 1, workers were starting to acknowledge the good points of old systems compared to Maximo”.
7.2.2 Business Process Reengineering
Organisational fit and adaptation are important to the implementation of modern large-scale enterprise systems. Like enterprise resource planning systems, EAM systems are also built with a pre-determined business process methodology that requires a fairly rigid business structure for it in order to work successfully. They are only as effective as the processes in which they operate. Companies that place faith in EAM systems often do so without reengineering their processes to fit the system requirements. Consequently, this often results in negative impacts on the effectiveness of both the EAM system and asset management practices. It is concluded that the business process for asset management in the organisation was not modified to fit the EAM system.
7.2.3 Management Commitment
Numerous interviews with AM stakeholders in the participating organizations including asset planning managers, GIS manager, field supervisors, senior managers, maintenance technicians, data entry staff, maintenance contractors and sub-contractors indicate that management commitment and supports are critical to the success of DQ improvement programs.
It appears to be that at strategic level managers know DQ problems, but it seems that they don’t treat the problems with high priority. Therefore, only limited resources were allocated to address the DQ problems. It was also found that at tactical level managers were frustrated by the poor quality information stored in the company’s asset management systems. It seems that they can’t do too much about it. They tend to rely on other information sources such as GIS for AM practice. They even suggested DQ should be linked with reward system in order to make staff responsible for the DQ problems they caused. At operational level, staff were aware that the data collected, entered were not right, but it seems that they don’t care. The operational supervisors know DQ problems existed in the AM systems, so they don’t trust its data. Instead, they create their own islands of information (eg in Excel
spreadsheets) for their own AM practice.
7.2.4 Lack of Codified Business Standard
To enforce business rules through monitoring, an organization must have a set of data standards for corporate information. Having a codified set of rules makes data monitoring possible because users know what to enforce when establishing data-quality controls. But these rules must be consistent and agreed upon across lines of business and among business and technical users.
7.2.5 Disconnect between Business and IT while Creating Metrics for Monitoring
While IT departments know how the data is stored and linked, the consumers of data (i.e., line-of-business employees who rely on data to make decisions) know what the data should look like. The two sides must work together to create a meaningful set of control metrics within the existing IT environment. Without collaboration, any data monitoring project will fail to yield measurable results because the metrics won't reflect the needs or "pains" of the business users.
7.3 People Perspective 7.3.1 Training
From a data quality perspective, training has not been sufficiently addressed. Thus, there was a lack of a general awareness of data quality. For example, staff often made mistakes in entries because they didn’t feel that it was important to ensure a high level of data quality. In particular, they were not aware of the severe consequence caused by these mistakes. In the organisation, when the new state-wide asset management system was first introduced, several ordinary staff members were chosen to take a brief, 3-day training workshop and then were assigned to be trainers for the rest of the organisation. Due to the insufficient knowledge and skills of these trainers, the system implementation experienced tremendous problems. One manager mentioned that, “training has been provided, but a lot of attendants are old and hence can’t be bothered”. However, through the interviews with field workers, one said that “training is the same for everyone” and “most of us have very little training, we’re mostly self-taught”. It was found that the gap between current practices and capabilities, and those required to harness everybody’s best efforts, is wide in the organisation. On the education front alone, simple things like ‘awareness of the cost of downtime’ and ‘how the data being collected is going to be used’ can transform the motivation, performance and creativity of the asset operators/technicians.
Managing assets requires all aspects of training as well as appropriate documentation of the system. It was found that organisations tended to focus more on
the ‘hardware’ part of the systems’ development process, putting less effort on the ‘soft’ part, that is, the training of how to operate and manage the system. People’s skills and abilities to use the system efficiently are very critical to ensure data quality in asset management systems. If people do not have the skills and knowledge to control the system, then even a perfect system would not be able to produce high quality information. Lack of training can cause serious damage and have an adverse impact on information quality. Unfortunately, it is easy for organisations to find reasons/excuses for avoiding adequate training for the staff and management.
7.3.2 Data Recording
In asset management, all of the analytical methods, prediction techniques, models, and so on, have little meaning without the proper input data. The ability to evaluate alternatives and predict in the future depends on the availability of good historical data, and the source of such stems from the type of data information feedback system. The feedback system must not only incorporate the forms for recording the right type of data, but must consider the personnel factors (skill levels, motivation, etc.) involved in the data recording process. The person who must complete the appropriate form(s) must understand the system and the purposes for which the data are being collected. If this person is not properly motivated to do a good, thorough job in recording events, the resulting data will of course be highly suspect.
Research into data collection has found that data quality and validation effectiveness improve, the sooner the collected data is entered, and the nearer the data entry is to the asset and its work. If the data entry point is remote from the asset, then the capability for accurately confirming the data is considerably reduced and the temptation to enter something - anything that the system will accept - is great. One manager said in the interview that, “I feel that most of the (data) errors over time have been because of the lag between the field data and being contained in the computer somewhere….they (field people) might wait a week before they complete their work order (entry)”. It was found that the longer the time lag between using the entered data and the time it was initially created, the less chance of cleaning up the data to make it useful.
7.3.3 Communication and Management Feedback
Competitive asset intensive companies have reported that most of their asset improvements come from their workforce. Despite the fact that “people are our greatest asset”, evidence of the opposite was often found. People’s problems, people’s relationships, people’s aspirations and people’s personal agendas are seldom given the consideration
appropriate to their importance in the successful implementation of an EAM system. In fact, the problem needs to be stated more emphatically. Most system implementations neglect the people factor and, as a result, most systems ultimately fail to achieve the objectives upon which their original funding was justified.
It appears to be that organisations continue to see the operators and technicians as skilled hands, rather than also having brains and being very sophisticated sensors. It was also found that field people within the organisation often generate the view that “year after year they filled out field data without feedback and a lot of them worked out that if they did nothing, nothing happens so why bother?”.
8. IMPLICATIONS AND
CONCLUSIONS
There are some implications for real-world practitioners, which emerged from these two preliminary case studies. The following conclusions and recommendations were drawn from the case study findings.
8.1 Understand Data Quality Issues for Asset Management
Data quality issues are critical to the success of asset management. The framework proposed in this paper provides a useful tool for planning the establishment of an awareness of data quality issues in managing assets. In particular, it indicates the key areas where the policy implementers need to focus and to monitor. Data quality issues need to be widely understood and managed in order to ensure effective asset management. When analysis is required for making decisions, to establish a data quality project regarding the management of engineering assets, issues discussed in this paper can help practitioners to perform a cost/benefit analysis in relation to data quality issues. The identification of data quality issues within the area of asset management will also serve to provide additional research opportunities for the development of tangible solutions to data quality problems in asset management.
8.2 Understand the Key Factors that Impact on Data Quality while Managing Assets
There are certain factors that influence data quality when managing assets. Organisations should focus on those key factors as defined by the framework in this paper, which include training, top management support, employee relations, organisational culture, and legislative requirements. Understanding the key factors should lead to high-level data quality management practices, which is a key to the successful implementation of effective
asset management. The knowledge of specifications of the key factors of data quality management in engineering asset management permits organisations to obtain a better understanding of data quality management practices, and perform better data quality controls in managing engineering assets. A particularly important factor for data quality projects is adequate training. Implementing a data quality project requires an effective project team that works together. Both engineering and IT personnel perform very important roles in the implementation process to ensure that the project is on the right track. Quality communication among engineering, business and IT people will significantly reduce data quality problems.
8.3 Adequate Training is Essential
Adequate training on data quality for all personnel involved in managing engineering assets is important for ensuring and improving data quality. People’s ability to use the system is equally important to ensure a relatively high level of data quality in asset management. Sufficient training should be provided to all employees to obtain a broad understanding of the system as a whole, as well as providing particular personnel with adequate documentation and specific training to deliver the critical mode of knowledge (know-what, know-how, know-why) for their specific data roles (data collector, data custodian, data customer) in their relevant functional areas in relation to the system.
8.4 Future Research
These case studies provided a better understanding of data quality issues for asset management as well as providing useful practitioner findings from real-world practice. Key data quality issues discussed and the use of the identified framework should help organisations obtain a better understanding of data quality issues throughout the process, leading to activities which will help ensure data quality. This research was conducted in two utilities companies, but similar issues could be found across different industries. Although the organisations may not have controls on those factors, organisations can actively manage those changes. Organisations could use external pressures to accelerate the quality management of internal information. Moreover, this paper has provided some recommendations, with implications for practitioners. Thus, the prevention and correction techniques can be applied according to the issues found in order to ensure a high level of data quality. In addition it has provided guidance (at least a starting point) to further explore how to implement Data and Information Quality policies successfully during the next phase of this research. (The questions and problems are listed in this research; the solutions are yet to be proposed).
REFERENCES
1. Badri, M. A., Davis, D. and Davis, D., 1995, “A study of measuring the critical factors of quality management,” International Journal of Quality and Reliability Management, Vol. 12, No. 2, pp. 36-53.
2. Ballou, D. P. and Pazer, H. L., 1995, “Designing information systems to optimize the accuracy-timeliness trade-off,” Information Systems Research, Vol. 6, No. 1, pp. 51-72. 3. Benbasat, I., Goldstein, D. K. and Mead, M.,
1987, “The case study research strategy in studies of information systems,” MIS Quarterly, Vol. 11, pp. 369-386.
4. Bever, K., 2000, “Understanding plant asset management systems,” Maintenance Technology, pp. 20-25.
5. Black, S. A. and Porter, L. J., 1996, “Identification of the critical factors of TQM,” Decision Sciences, Vol. 27, pp. 1-21.
6. CIEAM, 2003, Business Plan V1.0.
7. CIEAM, 2005, CIEAM Information Brochure, http://www.cieam.com/aboutus.
8. Dasu, T. and Johnson, T., 2003, Exploratory Data Mining and Data Cleaning, John Wiley, New York.
9. Eckerson, W. W., 2002, “Data quality and bottom line: Achieving business success through a commitment to high quality data,” Total Data Warehouse Institute Report Series.
10. Eerens, E., 2003, Business Driven Asset Management for Industrial and Infrastructure Assets, Le Clochard, Australia.
11. English, L. P., 1999, Improving Data Warehouse and Business Information Quality: Methods for Reducing Costs and Increasing Profits, Willey and Sons.
12. Firth, C., 1996, “Data quality in practice: Experience from the frontline,” Proceedings of Conference of Information Quality.
13. Fisher, C. W. and Kingma, B. R., 2001, “Criticality of data quality as exemplified in two disasters,” Information and Management, Vol. 39, pp. 109-116.
14. Fisher, T., 2005, “Meeting the data-quality challenge,” Computerworld.
15. Fox, C., Levitin, A. V. and Redman, T. C., 1994, “The notion of data and its quality dimensions,” Information Processing and Management, Vol. 30, No. 1, pp. 9-19.
16. Gelle, E. and Karhu, K., 2003, “Information quality for strategic technology planning,” Industrial Management and Data Systems, Vol. 103, No. 8, pp. 633-643.
17. Giannoccaro, A., Shanks, G. and Darke, P., 1999, “Stakeholder perceptions of data quality in a data warehouse environment,” Australian Computer
Journal, Vol. 31, No. 4, pp. 110-117.
18. Group 1, 2004, “Data quality and data integration: The keys for successful data warehousing,” MAS Strategies.
19. Haider, A. and Koronios, A., 2003, “Managing engineering assets: A knowledge based asset management methodology through information quality,” E-Business and Organisations in the 21th Century, pp. 443-452.
20. Hassan, B., 2003, “Examining data accuracy and authenticity with leading digit frequency analysis,” Industrial Management and Data Systems, Vol. 103, No. 2, pp. 121-125.
21. Hipkin, I., 2001, “Knowledge and IS implementation: Case studies in physical asset management,” International Journal of Operations and Production Management, Vol. 21, No. 10, pp. 1358-1380.
22. Huang, K., Lee, Y. W. and Wang, R. Y., 1999, Quality Information and Knowledge, Prentice Hall, Upper Saddle River, N.J.
23. Huh, Y. U., Keller, F. R., Redman, T. C. and Watkins, A. R., 1990, “Data quality,” Information and Software Technology, Vol. 32, No. 8, pp. 559-565.
24. IFS Applications, 2004, Enterprise Asset Management, IFS.
25. IPWEA, 2002, International Infrastructure Management Manual, Australia/New Zealand Edition.
26. Ives, B., Olson, M. and Baroudi, J. J., 1983, “The measurement of user information satisfaction,” Communications of the ACM, Vol. 26, No. 10, pp. 785-793.
27. Kahn, B. K., Strong, D. M. and Wang, R. Y., 2002, “Information quality benchmarks: Product and services performance,” Communications of the ACM, Vol. 45, No. 4, pp. 184-192.
28. Knox, M., 2004, “Asset managers move to centralized data and data management,” Gartner Research.
29. Kriebel, C. H., 1979, “Evaluating the quality of information systems,” Design and Implementation of Computer Based Information Systems.
30. Kumar, A. and Palvia, P., 2001, “Key data management issues in a global executive information system,” Industrial Management and Data Systems, Vol. 101, No. 4, pp. 153-164. 31. Levitan, A. V. and Redman, T. C., 1998, “Data as
a resource: Properties, implications and prescriptions,” Sloan Management Review, Vol. 40, No. 1, pp. 89-101.
32. Lillrank, P., 2003, “The quality of information,” International Journal of Quality and Reliability Management, Vol. 20, No. 6, pp. 691-703. 33. Linstone, H. A., 1999, Decision Making for
Perspectives to Improve Performance, Artech House Publisher.
34. Loshin, D., 2001, Enterprise Knowledge Management: The Data Quality Approach, Morgan Kaufmann, Academic Press, London. 35. McAndrew, S., Anumba, C., Hassan, T. and
Duke, A., 2005, “Potential use of real-time data capture and job-tracking technology in the field,” Facilities, Vol. 23, No. 1/2, pp. 31-46.
36. Mitroff, I. I. and Linstone, H. A., 1993, The Unbounded Mind: Breaking the Chains of Traditional Business Thinking, Oxford University Press, New York.
37. Orr, K., 1998, “Data quality and system theory,” Communications of the ACM, Vol. 41, No. 2, pp. 66-71.
38. Paiva, E. L., Roth, A. V. and Fensterseifer, J. E., 2002, “Focusing information in manufacturing: A knowledge management perspective,” Industrial Management and Data Systems, Vol. 102, No. 7, pp. 381-389.
39. Porter, L. J. and Parker, A. J., 1993, “Total quality management: The critical success factors,” Total Quality Management, Vol. 4, pp. 13-22.
40. Salaun, Y. and Flores, K., 2001, “Information quality: Meeting the needs of the consumer,” International Journal of Information Management, Vol. 21, No. 1, pp. 21-37.
41. Sandberg, U., 1994, “The coupling between process and product quality: The interplay between maintenance and quality in manufacturing,” Euromaintenance.
42. Saraph, J. V., Benson, P. G. and Schroeder, R. G., 1989, “An instrument for measuring the critical factors of quality management,” Decision Sciences, Vol. 20, No. 4, pp. 810-829.
43. Saunders, D., 2004, Innovation in Asset Management: Achieving a Nexus of Cost and Capability, UDT Pacific, Hawaii, USA.
44. Segev, A., 1996, “On information quality and the www impact a position paper,” Proceedings of Conference of Information Quality.
45. Shanks, G. and Darke, P., 1998, “Understanding data quality in a data warehouse,” Australian Computer Journal, Vol. 30, No. 4, pp. 122-128. 46. Sokianos, N., Druke, H. and Toutatoui, C., 1998,
Lexikon Produktions Management, Landsberg, Germany.
47. Spires, C., 1996, “Asset and maintenance management – Becoming a boardroom issue,” Managing Service Quality, Vol. 6, No. 3, pp. 13-15.
48. Steed, J. C., 1988, “Aspects of how asset management can be influenced by modern condition monitoring and information management systems,” IEE.
49. Steenstrup, K., 2005, “Enterprise asset
management thrives on data consistency,” Gartner Research.
50. Strong, D. M., Lee, Y. W. and Wang, R. Y., 1997, “Data quality in context,” Communications of the ACM, Vol. 40, No. 5, pp. 103-110.
51. Strong, D. M., 1997, “IT process designs for improving information quality and reducing exception handling: A simulation experiment,” Information and Management, Vol. 31, pp. 251-263.
52. Vosburg, J. and Kumar, A., 2001, “Managing dirty data in organisations using ERP: Lessons from a case study,” Industrial Management and Data Systems, Vol. 101, No. 1, pp. 21-31.
53. Wand, Y. and Wang, R. Y., 1996, “Anchoring data quality dimensions in ontological foundations,” Communications of the ACM, Vol. 39, No. 11, pp. 86-95.
54. Wang, R. Y., 1998, “A product perspective on total data quality management,” Communications of the ACM, Vol. 41, No. 2, pp. 58-65.
55. Wang, R. Y., and Kon, H. B., 1993, “Toward total data quality management (TDQM),” Information Technology in Action: Trends and Perspectives, Prentice Hall, Englewood Cliffs, NJ.
56. Wang, R. Y. and Strong, D. M., 1996, “Beyond accuracy: What data quality means to data consumers,” Journal of Management Information Systems, Vol. 12, No. 4, pp. 5-33. 57. Wang, R. Y., Storey, V. C. and Firth, C. P., 1995,
“A Framework for analysis of quality research,” IEEE Transactions on Knowledge and Data Engineering, Vol. 7, No. 4, pp. 623-640.
58. Werhane, P. H., 2002, “Moral imagination and systems thinking,” Journal of Business Ethics, Vol. 38, pp. 33-42.
59. Wheatley, M., 2004, “Operation clean data,” CIO Magazine, pp. 24-33.
60. Woodhouse, J., 2004, Asset Management, http://www.plant-maintenance.com.
61. Woodhouse, J., 2003, “Asset management: Concepts and practices,” Engineering Asset Management Conference Proceedings, ERA Technology, Surrey, UK.
62. Xu, H., Nord, J. H., Brown, N. and Nord, G. D., 2002, “Data quality issues in implementing an ERP,” Industrial Management and Data Systems, Vol. 102, No. 1, pp.47-58.
63. Xu, H., Nord, J. H., Nord, G. D. and Lin, B., 2003, “Key issues of accounting information quality management: Australian case studies,” Industrial Management and Data Systems, Vol. 103, No. 7, pp. 461-470.
64. Xu, H., 2000, “Managing accounting information quality: An Australian study,” Proceeding of the International Conference on Information Systems (ICIS 2000), Brisbane, Australia.
65. Yin, P. K., 2003, Case Study Research: Design and Methods, 3rd Edition, Sage.
66. Yusof, S. M. and Aspinwall, E., 1999, “Critical success factors for total quality management implementation in small and medium enterprises,” Total Quality Management, pp. 803-805.
ABOUT THE AUTHORS
Shien Lin is a researcher in the Strategic Information Management Laboratory for CIEAM at the University of South Australia. He has extensive experience working on various positions in the Information System and Finance sectors in Taiwan and Australia. He holds master degree in Electronic Commerce, and is currently working towards the completion of a PhD in Data Quality for Engineering Asset Management. His research interests cover many aspects of the data quality domain, particularly in developing data quality model to enhance the quality of data associated with the management of engineering assets.
Jing Gao received his PhD degree from the University of South Australia. He is currently working as a full time research fellow / lecturer for the School
of Computer and Information Science at University of South Australia. He has also been working as a professional trainer / consultant for many Australian leading organisations including the Defence Science and Technology Organisation (DSTO) of Australia, SA Water and etc. His research interest is about how to obtain an alternative view to appreciate complex social and technical problems.
Andy Koronios received his PhD degree from the University of Queensland. He has extensive teaching experience both in the tertiary sector at undergraduate and postgraduate, MBA, DBA and PhD lever as well as in the provision of executive industry seminars. He has numerous research publications in a diverse area such as multimedia and online learning systems, information security and data quality, electronic commerce, and Web requirements engineering. His current research interests focus on data quality and the use of information in strategic decision making. He is currently a Professor and the Head of the School of Computer and Information Science at University of South Australia.
(Received May 2005, revised June 2005, accepted November 2005)
