Module 3 Assessing & Selecting Models for Time-to-Failure Data

Lesson M03-01

Text: Section 5  pages 1 – 4

Duration: 12 minutes

In this lesson, we introduce the approach for selecting models that adequately describe the time to failure data.  Models selected in this step will be subsequently used to estimate reliability metrics of interest.  A graphical tool (Probability Plot) for assessing model fit is introduced.  While some details of constructing probability plots are introduced, the emphasis will be on applying this tool to determine models that adequately describe the data.

Module 3 Assessing & Selecting Models for Time-to-Failure Data

Lesson M03-00

Text: Section 5

Duration: 2 minutes

Introduction to Assessing & Selecting Models for Time-to-Failure Data Module

Module 3 focuses on how we determine appropriate models for describing the time-to-failure data.  We introduce Probability plots and goodness-of-fit statistics that assist in determining which distributions may provide a good fit to the data.  We also learn how to handle censoring and multiple failure modes when attempting to find reasonable models for the data.

[show_to accesslevel=”cre-jan-2020″]

Enjoy the Course

— select available sessions via the menu to the right (or scroll down on smaller screens – Note: sessions to be added as lesson material becomes available.)

I, Fred, am here to support you along the way. At any time, if you have a question, would like to chat about a concept, or check-in, let me know. Email or call (+1.408.710.8248), or arrange a Skype session (fmsrel) anytime, seriously.

[/show_to]

 Click the Enroll Now button
to join the course starting January 14th.

Course Information

The ASQ CRE certification exam preparation course allows you to brush up or review (and for some learn for the first time) every element of the current ASQ CRE Body of Knowledge.

The course provides a mix of live lectures, sample problems, and reading, to guide your mastery of the content.

The live sessions start on January 14th, 2020 at 3pm US Pacific time. The sessions occur Tuesdays and Thursdays at the same time. Each of the 16 sessions are two hours. We are using Adobe Connect (like Webex or GotoWebinar) for the live sessions and once enrolled you will receive information on setup and login details.

The course provides support and encouragement so you stay on track and have the best possible preparation for the certification exam.

The lectures also include advice on how to best use the concepts, tools, and techniques at work. The real value of the certification is found by improving the value you provide by being a better reliability engineer.

[hide_from visible_to=”public”]If you have already signed up for the course, login and enjoy.

[login-form redirect=”https://fred-schenkelberg-project.prev01.rmkr.net/cre-preparation-live-course”]
[password-recovery-link text=’Lost Password? Click here to have it emailed to you.’]

[/hide_from][hide_from accesslevel=”cre-jan-2020″]

Your Online Live Course with Instructor Support

32 hours of lecture/discussion over 8 weeks on Tuesday and Thursdays from 3 – 5 pm US Pacific time. (subject to change with the consensus of students to different dates/times.) We will also schedule makeup sessions in case of technical difficulty conducting the live online sessions.

Plus, I’m here to support you upon request.

Plus each session will include the ability to download the slides, view the video recording, or download the audio for review offline.

The homework will be assigned from the CRE Primer and additional handouts or sample exams and discussed at the start of each session.

Pilot students requested a set of glossaries one for terms and one for formulas. The terms glossaries are now available and included in the course. These will be available in the downloads section.

How Much Time Will the Course Take?

This will depend on how much you already know and/or desire to study. The lectures are 16 two-hour sessions for 32 hours, plus the expectation for working through the homework and studying/reading your reference material.

A general guideline as you prepare for the ASQ CRE exam is to spend at least 100 hours preparing. This will improve your working knowledge of the BoK topics which may be handy at work.

What are the course perquisites?

An interest in reliability engineering.

For those specifically preparing to sit for the ASQ CRE exam, you should meet the certification eligibility requirements (a mix of education and work experience, including time in leadership positions)

Give yourself  at least 2 month before exam day for the course material, reading, and study time[/hide_from]

Course and Exam References

For this course and for use during the exam, you are highly recommended to secure a copy of the QCI CRE Primer as your primary reference. QCI also maintains an errata for the various editions.

I highly encourage you to get a copy, read, study, markup, and get to know this reference. It is essentially the best and goto reference during the exam, and as a goto reference going forward.

You may want, if you do not have them already, a few other references, including:

Practical Reliability Engineering by O’Connor and Kleyner*

A good statistics book you know well. I use An Introduction to Statistical Methods and Data Analysis by Ott and Longnecker*

ASQ Exam Registration

If you plan to take the CRE exam, sign up separately for the course directly with ASQ. Learn more about exam dates, certification requirements and more at the ASQ CRE Certification page.

What is your return policy?

If you are not satisfied with the course, send me an email within 30 days after the course start date for a full refund.

Click the Enroll Now button to join
the course starting January 14th.

Thank you for the CRE course, it was really helpful to brush up my skills in ASQ CRE BOK under your guidance. — Douzi[/hide_from]

Module 2 Probability & Reliability Statistics

Lesson M02-07

Text: Section 3 pages 30 – 40

Participant Guide: Module 2 page 54

Duration: 35 minutes

When dealing with pass/fail outcomes or count data, discrete distributions are necessary.  In this lesson, we introduce the binomial distribution which models the probability of observing a specified number of failures from a sample size, given an assumed failure probability.  In addition to directly supporting probability calculations involving inspection data, the concepts will be very useful in understanding system probability calculations later in the course.  The Poisson distribution is also covered as it’s often useful in modeling the number of defects (where a particular unit may have more than 1 defect).

Exercise Question

A supplier of integrated circuits claims that at most 5% should fail a 20-hour stress test (high relative humidity). A customer performs the same test on 30 ICs and observes 6 failures. What is the probability of observing at least 6 failures if the supplier claim is true?

Module 2 Probability & Reliability Statistics

Lesson M02-06

Participant Guide: Module 2 pages 42 & 43

Duration: 13 minutes

Successful completion of these exercises indicates you have a good grasp of the various reliability metrics and how they are related to the probability density function.  The exercises also cover the basic probability distributions and how they play a part in estimating reliability statistics.

Exercise Questions

1. How is the Cumulative Distribution Function (CDF) related to the Probability Density Function (PDF)?

2. How is the Cumulative Distribution Function (CDF) related to the Reliability Function?

3. What does the B10 life indicate?

4. The reliability of units supplied by two different suppliers could be compared by comparing their reported MTTFs (True/False).

5. The MTTF estimates the time at which about 50% of the population is expected to fail (True/False).

6. For an exponential distribution (constant failure rate), it is true that R(100) = R(200) (True/False).

7. Why would the times to failure for bearing wear likely not follow an exponential distribution?

8. An electronic device is modeled using a Weibull distribution. The parameters are estimated to be b = 0.73 and h = 1000 hours. What is the reliability at 500 hours?

9. For the previous problem, how does the reliability at 500 hours change if we add 100 hours of burn-in?

Module 2 Probability & Reliability Statistics

Lesson M02-05

Text: Section 3 pages 41 – 46

Duration: 11 minutes

Here, we revisit conditional probability as applied to reliability estimation.  We present an example when burn-in is used to weed out defective units.  As illustrated in the example, burn-in is only effective when we are trying to strengthen a population that is subject to infant mortality failures.

Module 2 Probability & Reliability Statistics

Lesson M02-04

Text: Section 3 pages 13 – 29

Duration: xx minutes

When analyzing non-repairable systems or components, we typically find probability distributions that best describe/fit the available time-to-failure data.  This lesson introduces some of the common models (distributions) used in practice.  Although the Weibull model is emphasized somewhat due to its popularity and convenient properties, we should also consider alternate models when estimating reliability statistics.

Module 2 Probability & Reliability Statistics

Lesson M02-03

Text: Section 3 pages 1 – 12

Duration: 32 minutes

We use reliability metrics to set quantitative targets and to estimate performance against those targets.  In this lesson, we define the most common reliability metrics/statistics.  We also relate them to the probability density function which describes the time-to-failure data.

Module 2 Probability &Reliability Statistics

Lesson M02-02

Text: Section 2 pages 6 – 13

Participant Guide: Module 2 page 11

Duration: 28 minutes

While conditional probability is not necessarily that intuitive, it plays a role in various reliability applications such as warranty forecasting, reliability prediction with burn-in, and repairable systems analysis.  This lesson introduces conditional probability as well as independent events.

Exercise Questions

Module 2 Probability & Reliability Statistics

Lesson M02-00

Text: Sections 2 & 3

Duration: 1 minutes

Introduction to Probability & Reliability Statistics Module

In preparation for the reliability methods and tools to follow, this module covers many of the fundamental building blocks.  Since Reliability is a probability, we review some essential probability ideas and rules.  We introduce the concept of conditional probability since it plays a role in various methods such as warranty forecasting and burn-in.  We discuss probability models that may be considered in describing time-to-failure data with some emphasis on the Weibull distribution.  The Weibull distribution is very popular due to its flexibility and meaningful parameters although we should not limit ourselves solely to this model.  Finally, we discuss some useful discrete distribution models and their applications to reliability.

Module 2 Probability & Reliability Statistics

Lesson M02-01

Text: Section 2  pages 1 – 5

Duration: 8 minutes

Since Reliability is a probability, we should understand some basic probability ideas.  Here we review basic probability concepts which will be important for applications later in the course.

Module 1 Reliability Concepts & Data

Lesson M01-05

Text: Section 1 pages 5 – 7

Duration: 18 minutes

The bathtub curve is a useful model for understanding the types of product failures that may occur.  Here, we define the three phases of the bathtub curve and how they are characterized by the failure rate.  We also introduce the reliability function as an extension of the reliability definition covered earlier.

Exercises for Lesson 5

Module 1 Reliability Concepts & Data

Lesson M01-04

Text: Section 1 pages 8 – 11

Duration: 22 minutes

In addition to failure data, Reliability analysis must consider information about un-failed units as well.  Even for the failures, some uncertainty may exist regarding when the failure occurred.  In this lesson, we cover the types of censoring situations that typically exist in addition to the known failure times.

Module 1 Reliability Concepts & Data

Lesson M01-03

Text: Section 1 page 4

Duration: 21 minutes

The type of data we collect affects how much data we need, and the methods used to analyze it.  In this lesson, we explore different types and sources of reliability data, classifications of analyses, and challenging/unique aspects of reliability analyses.

Module 1 Reliability Concepts & Data

Lesson M01-00

Text: Section 1

Duration: 2 minutes

Introduction to Reliability Concepts & Data Module

In this introductory module, we define Reliability and cover many important reliability concepts.  We discuss why reliability performance is so important and the types of questions that may be answered by analyzing failure data.  Different types of reliability data are reviewed.  We introduce the concept of censored data which allows us to correctly utilize partial or incomplete data in the analysis of test or field data.  We cover the Bathtub Curve which illustrates the types of failures that may occur over a product’s lifetime.  Finally, we introduce the reliability function.