top of page

ARC Training Centre for Transforming Maintenance through Data Science

Untitled design(4).png
Going Over Data

Transform the maintenance management process

Our vision to enable Australian companies to lead the way maintenance is conducted, increasing both productivity and profitability through the automation of the work management stages.


Purpose of the Centre


To deliver the next generation of data science solutions focused on the problems that industry needs, which is to ensure the efficient and effective maintenance execution.

Our academic-and-industry partnership will deliver the next generation of data science products coupled with training to ensure efficient and effective maintenance execution.

Enable Maintenance Work Management by providing decision support solutions for activity that need a ‘human-in –loop’ for maintenance and automating activities that don’t.

The Centre is a partnership between Curtin University, The University of Western Australia, CSIRO and industry partners Alcoa, BHP and Roy Hill, as well as CORE Innovation Hub and the Minerals Research Institute of Western Australia.

To enable this transformation, the Centre will focus on the three key roles in a new digitally-driven maintenance management system. These are:


  1. The maintainers responsible for executing work on the assets and making crucial observations about as-found condition of the assets,

  2. The engineers responsible for the models and analysis necessary for fault detection, asset health assessment and remaining useful life predictions, and quantifying uncertainty, and

  3. The maintenance managers responsible for ensuring that decisions at the system level balance cost, risk and performance.






Each research themes has a series of three blocks of work:

The development of fit-for-purpose data sets, these will be provided by our Partner organisations. 

The aim here is that all data captured by the maintainers is machine readable, the data is cleaned and stored, and captured data in a way that is repeatable and requires minimal data wrangling by users.

Building Suitable models.

The data sets will be used to build suitable models, taking us to our second block of work. Models are themselves developed, validated, stored and updated in a trackable way. The use of these models by decision makers is also tracked and performance changes measured and assessed.


Computational efficiency. 


The Centre has employed software engineers to ensure research outcomes are more easily translated into existing operations and/or provisioned via a consistent platform


We understand that models by themselves don't deliver value; we also need to change organisational psychology/culture/training/software architecture/business process. What we believe makes this Centre special is that this is built-in alongside our translation themes.


The output of maintenance analytics creates new decision-making challenges for maintainers, engineers, and managers. The most significant risk is that the information is simply not used in Organisations.

This theme will adopt a multilevel perspective to build a comprehensive picture of the factors that support decision making when using complex data. This aspect of the project translates the core research programs into business practice by integrating user issues hierarchically at three levels of analysis:

  1. Cognitive is particularly important for maintainers using information delivered through novel media and in a context that is different to traditional maintenance systems,

  2. Task demands is particularly important for engineers designing models, and for the maintainers using models. Use of data will be inhibited if the output of models is not a meaningful part of the overall task requirements of maintainers or generates tasks that cause fatigue, boredom, or distress

  3. Organisational culture is particularly important for managers implementing decisions based on analytic models. For example, if an underlying cultural belief places primacy on the intuition and implicit skill of senior managers, then analytics models can undermine the role of managers and limit the use of data output


This translation project will work with the industry partners to implement a plan based on research at each level to integrate across the three levels.

A framework around “person, tasks, and culture” will be developed that is specific to the maintenance context of the organisation and involves stakeholders from the maintainer, the engineer, and the manager groups. It is essential that the overall system, not just individual elements, supports the translation of analytic models into business practice.

Project Leads

Mark Griffin (5).jpg

Mark Griffin

Eden Li white background.png

Eden Li

Are we ready to transform?
Building capability to adopt data science solutions
Masterclass session by Dr Keyao Li

bottom of page