Course Aim

The 24-hour Weibull Analysis Course teaches attendees how to extract information from reliability failure data to make better design, operational, maintenance and fleet management decisions. These decisions can include optimizing preventive maintenance intervals, identifying failures caused by manufacturing defects, selecting the right component, and plenty more. Attendees will be able to make decisions like these confidently, improving reliability and minimizing unavailability. Weibull analysis is one of the most widely used and powerful statistical tools to investigate reliability data – especially when there is not a lot of data available (which describes most commercial and operational scenarios).

The course focuses on practical, real-world applications of Weibull analysis. This includes predicting reliability performance (such as warranty reliability), determining key reliability milestones (such as warranty period, service-life or ‘BX’ lives), the nature of failure (wear-in, wear-out and everything in between), if your system has multiple failure modes … and lots of other key inputs to product development decisions.

Attendees also receive the Acuitas Weibull Plotting Tool which is based on Microsoft Excel ®. Commercial software the provides the same outputs usually involve licenses that cost tens of thousands of license fees each year!

By the end of this course, attendees will be able to look at a Weibull Plot like the one above and be able to determine key failure and reliability characteristics without having to rely on equations and theory only. Attendees will practice in-depth and increasingly challenging worked exercises to ensure they feel comfortable solving even the most challenging problems.

This course is suited to engineers who need to be able to interrogate reliability and failure data to make better design, maintenance, and operational decisions by both understanding how their systems are failing, and quantify reliability performance to see if it meets requirements.

Course Conduct

This course is offered virtually. Classes will be delivered virtually via Zoom or teams from 8:00 am to 3:00 pm Australian Eastern Time zone (with a 1-hour break for lunch) over four days. Attendees will be provided with access to course material via a course webpage for 12 months, allowing them to review material as required. Each lessons incorporate videos with corresponding guidebooks featuring matching visuals, and revision questions and answers, providing seamless learning across resources.

Case studies are used to reinforce key concepts, and teach attendees how to use the Acuitas Excel-based Weibull Plotting tool. Asynchronous support is provided by the presenter after the course is complete. Attendees will need to complete online revision questions to receive a certificate.

On registration, you will be sent an e-mail with instructions advising how to log on to the course to access course materials and other details

Course Outline

1. Introduction – Outline of the course, administrative details, learning objectives and course content.

2. 10 Reasons to do reliability – Description of the benefits of Weibull analysis (as part of reliability engineering more broadly) along with what attendees can expect to be able to do once they complete the course.

3. Reliability and Randomness – Introduction to how failure processes are random, and how all random processes have characteristics that help us understand system reliability characteristics.

4. How we describe what random looks like – Key concepts like probability density functions (PDFs), cumulative distribution functions (CDFs) and hazard rates are explained, along with how they can be used to describe reliability characteristics.

5. The exponential distribution – Attendees are taught about the ‘simplest’ probability distribution, which describes how systems that never age or experience infant mortality. The exponential distribution is mathematically simple, and overused as a result. However, it can model some basic failure mechanisms.

6. The normal distribution – Also known as the bell curve, the ‘normal distribution’ models random variables that are the sum of other random processes, much like how the total tread of a tire is the sum of all tread lost per unit distance travelled. Normal distributions are useful for modelling failure based on typical wear failure mechanisms.

7. The lognormal distribution – The ‘lognormal distribution’ models processes that are the product of other random processes. Several wear-out failure mechanisms (like fatigue and corrosion) accumulate damage in this multiplicative way, meaning the lognormal distribution is able to model them well.

8. The Weibull distribution – The ‘Weibull distribution’ is able to mimic all of the previous probability distributions and other failure mechanisms including those based on manufacturing defects that cause infant mortality. Attendees are taught how this distribution, and its inherent flexibility, can model a huge range of failure processes.

9. Probability plotting – Probability plots are specially scaled charts that will show if failure data comes from a process that can be modelled by any of the distributions above. These charts ‘arrange’ data points in a way that will create a visibly straight line, corresponding with a nominal probability distribution.

10. Probability plotting exercise (part #1) – Attendees will practice organizing and plotting example failure data on a Weibull probability plot, focusing on identifying key quantitative reliability characteristics.

11. Probability plotting exercise (part #2) – Attendees will expand on the previous practice, focusing on identifying the ‘statistical signatures’ of key failure mechanisms, and generate recommendations to improve reliability performance.

12. Acuitas Weibull Plotting Tool (worked exercise #1) – Attendees practice implementing the techniques from the previous exercise in the Acuitas, Microsoft Excel ® based Weibull Plotting tool.

13. Detailed Weibull Analysis – Failure mechanisms are the physical, chemical, thermal or electrical process that ultimately causes failure. These are the mechanisms that are triggered by root causes, and they are the processes that are modelled by probability distributions. Attendees are taught how to identify likely failure mechanisms from data, using Weibull plots.

14. Acuitas Weibull Plotting Tool (worked exercise #2) – Attendees using the Acuitas Weibull Plotting tool to identify likely failure mechanisms from failure data, and make specific recommendations based on them.

15. Getting the Most Out of Your Data – Attendees are taught how to look for key reliability performance characteristics in failure data using Weibull plots. These characteristics can help inform key maintenance, operational and fleet management decisions.

16. Acuitas Weibull Plotting Tool (worked exercise #3) – Attendees use the Acuitas Weibull Plotting tool to identify key trends and signatures in data to inform key maintenance, operational and fleet management decisions.

17. Confidence in Our Conclusions – All data analysis that involves regression from random variable values involve uncertainty. The extent of this uncertainty is measured in terms of ‘confidence.’ Attendees will be taught what ‘confidence’ is, and how it applies to Weibull analysis.

18. Acuitas Weibull Plotting Tool (worked exercise #4) – Attendees learn how to quantify and illustrate the uncertainty inherent in Weibull analysis, and hot to illustrate it using the Acuitas Weibull Plotting tool.

19. 3-Parameter Weibull Analysis – Attendees are taught about the ‘3-parameter Weibull distribution’ which incorporates a failure free period that can help model special failure mechanisms in certain scenarios, and reliability failure data that is incomplete or corrupt.

20. Acuitas Weibull Plotting Tool (worked exercise #5) – Attendees learn how to analyze failure data with the 3-parameter Weibull distribution using the Acuitas Weibull Plotting tool to identify key failure characteristics and operating parameters, including aspects of incomplete or corrupt data.

21. Weibayes Analysis – Weibayes analysis is a special type of analysis that allows the ‘statistical signature’ of a failure mechanism that a system or product is known to have. This allows reliability data analysis to be more informative with less data (or testing).

22. Acuitas Plotting Tool (worked exercise #6) – Attendees will learn how to conduct Weibayes analysis on a small data set when the failure mechanism (and its statistical characteristic) are known.

23. Weibull Analysis and Accelerated Life Testing (ALT) – ALT involves increasing stresses (such as temperature and stress) in a controlled and quantifiable way to ‘speed up’ testing. Attendees are taught how Weibull analysis applies to ALT to ensure the validity of the test parameters.

24. Conclusion and Wrap-Up

Course Material

The following resources will be provided to attendees of this course:

1. An editable PDF course workbook

2. Acuitas Weibull Plotting Tool (Microsoft Excel) – licensed for attendee use only

3. Revision questions and answers for each lesson.

4. A completion certificate on passing the course assessments.