WeibullR.learnr: An Interactive Introduction to Life Data Analysis
10 July 2023
Statement of Need
Life data analysis is the study of how things, from machines to
people, function over time. Life data shows how long things last or how
long before they need to be repaired or replaced. Although numerous
learning resources exist, most reference proprietary software that is
often unavailable to those hoping to get started with life data. This
proprietary software is usually meant for commercial applications and
may not be affordable for students or young professionals.
WeibullR.learnr (Govan 2023a)
is an open source set of interactive learning modules, exercises, and
functions for learning introductory life data analysis. The primary
motivation for the development of WeibullR.learnr was to
simultaneously introduce the basic concepts and also open source
software for analyzing life data. The target audience for this project
are beginning practitioners and university students.
Description
WeibullR.learnr is written in R (R Core Team 2023) and is built using
WeibullR (Silkworth and Symynck
2022), a R package for Life Data Analysis in the tradition of
Waloddi Weibull (Abernethy 1993), and
learnr (Aden-Buie et al.
2023), a framework for building interactive learning modules in
R.
Currently, three primary learning modules exist. These modules can be taken in either order and can be taken separately or together. The learning modules are designed to be plug-and-play, but changes can be made by forking the software repository and modifying the fork.
WeibullR.learnr() provides an interactive introduction
to Life Data Analysis. The learning objectives include basic Weibull
analysis, data censoring including right censored data and interval
censored data, different types of Weibull models including the 2P
Weibull, 3P Weibull, and Weibayes model, parameter estimation methods
Median Rank Regression (MRR) and Maximum Likelihood Estimation (MLE),
and different plotting methods such as Probability Plots and Contour
Plots. The estimation duration for this module is approximately 2
hours.
RAMR.learnr() is a quick reference for common
Reliability, Availability, and Maintainability (RAM) concepts. The
learning objectives include the basic concepts and application of
Reliability, Availability, Mean Time to Repair (MTTR), Mean Time to
Failure (MTTF), Mean Time Between Failures (MTBF), Failure Rate,
Probability of Failure, and \(B_n\) or
\(L_n\) life. The esimated duration of
this module is about 1 hour.
TestR.learnr() provides an interactive introduction to
Reliability Testing. The learning objectives include defining key
reliability growth concepts, including Crow-AMSAA and Duane models;
fitting a reliability growth model to real-world data using R;
interpreting reliability growth plots and identifying trends; applying
the Crow-AMSAA model to assess reliability growth; explaining
fundamental concepts of accelerated life testing, including the use of
Arrhenius and Power Law Models; conducting an accelerated life test with
real-world datasets, utilizing R for analysis; nnalyzing plots that
illustrate the relationships in accelerated life testing, identifying
key patterns and data trends; utilizing Arrhenius and Power Law models
to evaluate the impact of stress factors on product reliability. The
estimated duration of this module is about 2 hours.
The modules can also be accessed in a browser at WeibullR.learnr, RAMR.learnr, and TestR.learnr.
Several helper functions for common RAM calculations are also included. These functions make it easy to apply the concepts covered in this module.
rel()- reliability functionavail()- availability functionmttf()- mean time to failuremtbf()- mean time between failureserv()- serviceability factorfr()- failure rate
The project documentation includes installation instructions for
WeibullR.learnr and the required dependencies, examples of
running the programs, and references to previous work used to build the
modules. The documentation also references more resources for learners
looking for expanded applications. These resources include
WeibullR.plotly (Govan
2023b), a R package for interactive Weibull probability plots,
and WeibullR.shiny (Govan
2023c), a shiny (Chang et al. 2022)
web application for Life Data Analysis.
This project was inspired by the Reliability Program at a major technology company, which has been well established and received at the company. While the program has been popular, it relies on proprietary software that is not available to all researchers and practitioners. The type of proprietary software has also changed over time, making some of the original learning material obsolete. The main goal of this project is to reach a wider audience through open source resources. These learning modules are not only open to all, but will also evolve with the underlining software. The potential benefits of making these resources available to a broader audience include fostering a community of proficiency, collaboration, and innovation.
Learners are encouraged to try the modules and contribute to the project. The repository includes a Contributor Code of Conduct for making contributions. Issues and requests may be filed by raising an Issue or Pull Request.
Acknowledgements
The author would like to acknowledge the authors of the Reliability Program that originally inspired this project.