Determining mean-time-between-failure (MTBF) estimation for in-service assets is an essential process. Statistical distributions of end-of-life values are used to assess asset reliability performance and the effectiveness of maintenance strategy. However, identifying the end-of-life event for each instance of functional failure is an arduous, manual process dependent on structured and unstructured fields in the maintenance management system and rules used by individual reliability engineers. We emulate the process of end-of-life event detection using a natural language processing pipeline followed by statistical parameter estimation to produce MTBF values for in-service assets from maintenance work order data. Using this pipeline we test how alternate mappings of words in unstructured text and the use, or not, of structured data can impact the identification of end-of-life events. We demonstrate the pipeline on data sets from two industrial users with 14,508 and 89,259 maintenance work orders, respectively. We find that resulting MTBF estimates can vary from, for example, 97 to 226 days for a single asset depending on the mappings and rules used. The main contributions of this paper are a) a demonstration of the impact of undocumented decisions made by reliability engineers in identifying end-of-life events on MTBF estimates, and b) an end to end pipeline for MTBF estimation from raw maintenance work order texts. In addition, we provide open-source code for the pipeline that can be used by asset owning organisations to semi-automate the MTBF estimation process in a manner that is fast and scalable, and ensures the rules used for end of life determination are documented and hence the process is transparent and repeatable.
How to Cite
natural language process, reliability metrics, mttf, mtbf, reliability management, failure event, embedding, word2vec, maintenance, cmms, work order, nlp
This work is licensed under a Creative Commons Attribution 3.0 Unported License.
The Prognostic and Health Management Society advocates open-access to scientific data and uses a Creative Commons license for publishing and distributing any papers. A Creative Commons license does not relinquish the author’s copyright; rather it allows them to share some of their rights with any member of the public under certain conditions whilst enjoying full legal protection. By submitting an article to the International Conference of the Prognostics and Health Management Society, the authors agree to be bound by the associated terms and conditions including the following:
As the author, you retain the copyright to your Work. By submitting your Work, you are granting anybody the right to copy, distribute and transmit your Work and to adapt your Work with proper attribution under the terms of the Creative Commons Attribution 3.0 United States license. You assign rights to the Prognostics and Health Management Society to publish and disseminate your Work through electronic and print media if it is accepted for publication. A license note citing the Creative Commons Attribution 3.0 United States License as shown below needs to be placed in the footnote on the first page of the article.
First Author et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 United States License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.