Failure modes and effects analysis and maintenance strategy development are key contributors to the design of any asset management program. Failure data from the field can be used to determine if the FMEA and strategies are effective. One approach to achieve this is to look at corrective maintenance work orders and determine what event happened, what caused the corrective work, and what remedy was applied. These can be described by PCR (problem cause remedy) codes and need to be supplied by the operators/ maintainers. However achieving consistent data quality can be challenging and a way of checking if appropriate and consistent P, C and R codes are being used is necessary. This information can be fed back to the code providers to get them involved in how to improve the data quality.
This project involves close collaboration with the UWA NLP-TLP team https://nlp-tlp.org/. This team has developed a suite of tools to digitalise unstructured texts in maintenance work orders and extract (at scale) knowledge that can potentially be used to identify the problem and remedy. Once the new PCR system is up and running this digital extraction can be compared to the codes supplied by the operator. In this project specific work is needed to develop an agreed set of failure mode (problem) codes and remedy codes based on historical data for a specific asset class.
The project will look at a specific class of assets and the first task for the successful applicant is to consider different asset classes and make a reasoned selection e.g. chlorination systems. The project will involve the application of NLP tools so strong coding skills in Python and some machine learning and visualisation skills are required.