UWA Homepage  /  Home  /  Maintenance, Maintainers and Data

Navigation


Maintenance, Maintainers and Data



Aim


Assist industry to move maintenance practices and decision-making into the 21st Century so that Maintenance delivers value through minimising of risk and ensuring asset systems operate in a predictable and controlled environment.




Maintainer of the Future White Paper

In the last two decades we have seen significant changes in how assets are operated due to advances in process control, remote operations systems, and the development of autonomous assets. In contrast, the nature of maintainers’ day-to-day work has changed very little. Engines still need lubricated, motors wired up, pipes welded and sensing systems repaired. However we may be at a tipping point as a combination of business drivers (cost, productivity, safety) and technical developments (automation, augmented reality, real-time diagnostics) combine to change the nature of maintenance work. This paper examines these drivers and looks at how these might influence the work and training of the maintainer of the future.

Collaborators: W.Jacobs (co-author), numerous contributors from industry, government and academia

Outputs:

Funding: CRC Mining 2013-14




Maintainers of the Future – Adapting to Change

This project started in 2014 as a follow up to the Maintainer of the Future White Paper. In the course of the White Paper it became apparent that the voice of the maintainers is rarely heard in studies about their work practices and training requirements. This new project aims to address this with the help of colleagues in Anthropology here at UWA. Modern anthropologists work with the likes of Google to understand consumers and employees. Here we are looking at the following questions: “How does modern technician’s work shape their learning, identity and community?

How is this changing with changes in technology, training practices, and organisational work practices? If you change the nature of the work, what does that do to people’s identity and commitment to the work? What are the impacts of these changes on maintainers? What are the impacts of these changes on organisational productivity and safety?”

Collaborators: A/P Martin Forsey and Marc Schmidlin (both at the UWA School of Anthropology and Sociology)




A tool for assessing Maintenance Data Quality

This paper presents an approach to decide a priori what types of decision support models an organisation’s operational and maintenance data will support. It is based on the idea that the model selected should be fit for purpose for the decision that needs to be made and the data available to support the model. The approach is tested using a set of decisions from component life prediction through to more complex condition based maintenance decisions. Only a limited number of cases involving critical equipment had the data to support routine condition based maintenance decisions. The approach supports the organisation to improve by identifying which data would be most useful for specific decision support models they might wish to use. It also provides a scoring mechanism so that improvements in data over time can be tracked.

Collaborator: Dr. Neil Montgomery (Centre for Maintenance Optimisation and Reliability Engineering, University of Toronto)

Outputs:




Improving the quality of data collected by maintainers

This paper argues that there are benefits to more fully understanding the psychological factors that lay behind data collection. Thus, rather than assuming what the data collectors want, a goal hierarchy approach determines that empirically. Practically, this supports the development of customized interventions that will be much more effective and sustainable than previous efforts.

Findings include that managers should be careful of the degree to which “push” factors, such as managerial pressure and technological input control, are relied upon. While they may be helpful for motivating those data collectors who are not intrinsically motivated, they are either not helpful or may discourage those data collectors who are intrinsically motivated. Instead, self-concordance may act as a longer-term, more stable approach to increasing the motivation of data collectors and thus increasing the quality of data that enter reliability systems.

Collaborator: Professor Kerrie Unsworth (UWA Business School and Lead Author)

Outputs:

Funding: CRC Mining and Water Corporation




Glossary of Reliability Terms and Definitions for the Mining Industry

Following a comprehensive review of national, international and industry standards and guidelines a set of preferred Terms and Definitions was developed for use by mining industry practitioners to promote a more unified approach within the industry and support benchmarking.

The glossary can be viewed here.

Collaborators: Mark Ho (UWA Engineering) and a number of mining industry focus groups

Outputs: A Glossary of Reliability Terms and Definitions. This is available for free download.

Funding: CRC Mining




Mobile Mining Equipment Reliability Database

Mobile mining assets are often part of fleets with small populations, there are site-specific operational and maintenance contexts and the maintenance work order data is often regarded as not fit-for-purpose for reliability analysis. This research project has produced statistically significant failure data sets for mobile mining assets and methods to extract, clean, present and use data to identify reliability improvements.

Collaborators: Mark Ho (UWA Engineering), CRC Mining Industry members

Outputs:

Funding: CRC Mining




Novel approaches to data cleaning of maintenance data in CMMS

To tackle the challenges, natural language processing techniques have been explored to improve the structure of the textual data. Textual data also contains typographical errors which can be detected using spelling suggester. In collaboration with Computer Scientists we are exploring the use of n-gram analysis, a probability language model. Using a sample of 1000 maintenance logs, this cleaning tool detected 82% of the words and typographical errors. The tool also identifies 88% of the word's structure correctly, which can be used to provide an insight about the quality of the data. This work will continue to focus on textual cleaning for applications in areas that store free text fields such as maintenance records and the measure that can be used to quantify its quality.

Collaborators: Professor Amitava Datta (UWA Comp Sci) and Hui Li Leow

Outputs: Honours thesis by Hui Le Leow 2014




Models to assist in Maintenance Decision making

Much of the work described above is about the collection and cleaning of data about maintenance activities. The purpose of this data is to support maintenance decisions. Much work has been done on the development of models. Some example publications are given below.

Publications:


Links


CRC Mining http://www.crcmining.com.au/

AM Council http://www.amcouncil.com.au/

C More http://cmore.mie.utoronto.ca/



Back to top