JD Solomon, PE, CRE, CMRP and a valuable member and volunteer of ASQ Reliability Division has released a book including some of his expert experiences and knowledge.

Communicating Reliability, Risk and Resiliency to Decision Makers:

Effective communication of concepts and solutions related to reliability, risk, and resiliency is frequently cited by technical professionals as the most challenging and overlooked aspect of their work. Texts and guidance documents often reference the importance of better communication; however, there are few practical examples and limited practical guidance for this type of communication. This book fills many of these gaps for practitioners. Communicating Reliability, Risk and Resiliency to Decision Makers provides the hands-on approaches and techniques that will make you more effective in getting decision makers to move from discussion to action.

More info on http://reliabilityriskresiliency.com/

A more extended review of the book will apear later in the blog section

At times, several items are tested simultaneously, and the results are combined to estimate the mean time before failure/mean cycles before failure (MTBF/MCBF). When performing tests of this type, two common test results occur:
type I censoring(also called time-truncated censoring) and type II censoring (also called failure-truncated censoring).

Note that the MTBF/MCBF (θ) is the reciprocal of the failure rate (λ).

Type I Censoring

This is when a total of n items are placed on test. Each test stand operates each test item a given number of hours or cycles. As test items fail, they are replaced. The test time is defined in advance, so the test is said to be truncated after the specified number of hours or cycles.

A time- or cycle-truncated test is called type I censoring.

The formula for type I censoring is:

Type II Censoring

Type II censoring is where n items are placed on test one at a time, and the test is truncated after a total of r failures occurs. As items fail, they are not replaced or repaired.

Example: A total of 20 items are placed on test. The test is to be truncated when the fourth failure occurs. The failures occur at the following number of hours into the test.

The remaining 16 items were in good operating order when the test was truncated at the fourth failure. What is the estimate of the MTBF?

What is the failure rate?

Source:

Practical engineering, process, and reliability statistics / Mark Allen Durivage

ISBN: 978-0-87389-889-8

Picture © B. Poncelet https://bennyponcelet.wordpress.com

Harold (Harry) E. Ascher was known widely in the reliability engineering community for his many contributions, especially in the area of statistical modeling of repairable systems reliability.

The 1984 book co-authored by Harry Ascher and Harry Feingold, Repairable Systems Reliability: Modeling, Inference, Misconceptions and Their Causes [1] was the first book to thoroughly address statistical modeling of systems that are repaired rather than discarded after the first failure.
Harry Ascher earned degrees in Operations Research from City College of New York and New York University and spent most of his career with the Naval Research Laboratory (NRL) in Washington DC. Following his retirement from NRL, he spent more than 20 years as a private consultant.
He received many distinctions and awards including the Allan Chop Award from ASQC (now ASQ) in 1995.
One of Harry Ascher’s greatest visions, as a professional, was to “clean up” notation and terminology from what he considered to be results of inherent and avoidable ambiguities between properties of repairable systems and, in contrast, properties of (non-repairable) parts.
For example, the term “Failure Rate” in the literature is used in multiple conflicting meanings, sometimes even within the same article.
One meaning is the derivative of the expected cumulative number of failures for a repairable system (also known as a “Rate of Occurrence of Failures” of “Failure Intensity Function”).
The other, quite different, meaning is the ratio of the Probability Density Function and the Reliability Function for a part (also known as the “Force of Mortality” or “Hazard Function”).
He introduced us the fact that the most important analysis to perform on statistical data for a repairable system is to test for trend.
This may possibly classify the system as a “Happy System” (times between failures tend to get larger and larger, a “Sad System” (times between failures tend to get shorter and shorter) or a “Non-Committal System” (times between failures show no significant sign of trend).
He would repeatedly state that, when it comes to modeling a repairable system, “a set of numbers is not a data set.” Often, the particular pattern in which failures occur provide more useful information about
a system, than does a set of numbers describing the observed times between failures.
He taught us that a renewal process is generally an inappropriate model to use for a complex repairable system consisting of many parts that have either “infant mortality” or “wear-out” characteristics.
To demonstrate how absurd the renewal process model is, when used to describe a complex system such as a car he often told his “Honest John” story.
This is about a used-car salesman trying to sell a wornout car with multiple mechanical issues by telling the customer: “Two days ago the battery was dead, so we charged it. So the car is two days old!”

(excerpted from article by Christian K. Hansen, President, IEEE Reliability Society)

Previously published in the June 2014 Volume 5, Issue 2 ASQ Reliability Division Newsletter

Picture © B. Poncelet https://bennyponcelet.wordpress.com

Example using “Bayesian” assumptions in proving a new design is X% better.
Suppose my current part is producing failures in the field with a Weibull β=2, η=95 hours.
I have demonstrated that I can reproduce the failure mode in the lab by producing the same β.
I have redesigned the part and placed 8 of the new designed parts on test in the same lab setup.
After a few weeks they have test times of 210, 260, 570, 120, 225, 280 500 and 400 hours with no failures.
How much better is this new part?
One way to demonstrate this is to calculate the MLE η of the 8 times demonstrated on the new designed parts, “assuming” a β=2 and “assuming” 1 failure is imminent:
Therefore, showing you have (conservatively) over 10X better reliability

Previously published in the June 2013 Volume 4, Issue 2 ASQ Reliability Division Newsletter

Picture © B. Poncelet https://bennyponcelet.wordpress.com

ASQ is seeking volunteers to help write example questions to be used for Certified Reliability Engineer (CRE) exam practice questions.
You must be a member of ASQ and a CRE.
You can not have participated in efforts related to the actual CRE Exam content and development within the past two years in order to keep actual and practice CRE efforts separate.
You will be provided RUs for the hours contributed to the project as well as recognition of your contributions.

If you are interested in supporting this effort, please contact Donna Grunewald at dgrunewald@asq.org for more information

Info on ASQ CRE https://asq.org/cert/reliability-engineer

There are some classic methods for determining the unknown parameters in reliability analysis including probability plotting, least square, and maximum likelihood estimation (MLE).
These methods provide a simple value for parameters based on experimental data.
Bayesian approach is the method of choice employed widely in research for estimating and updating parameters values.

The advantages of this method:
1. The Bayesian approach is an updating network which does not ignore the prior information in contrast with the classic methods, but updates the earlier estimations with new obtained knowledge to improve the estimated parameter.
2. The output of Bayesian network is a distribution instead of a simple value for the parameter.
3. In classic methods, a large sample size is required for convergence of the estimate. For analysis of cases restricted by limited data the Bayesian approach is a good approach for parameter estimation.

By considering X as the unknown parameter and E as the new knowledge of crack length, Bayesian theorem modifies a prior probability yielding a posterior probability , via the expression:

where is the likelihood function and is constructed based on new available knowledge and evidence.
The factor f(E|X)/∫f(E|X)π0(X)d(X) is the impact of the evidence on the belief in the PDF of the parameters.
Multiplying the prior PDF of the parameters by this factor provides a theoretical mechanism to update the prior knowledge of the parameters with the new evidence.

A Bayesian network is a complicated method in practice. An analytical solution rarely occurs.
However, it is possible to moderate the Bayesian difficulty by numerical method through Markov chain Monte Carlo (MCMC) solution.
This numerical method is applicable for similar approaches which need to integrate over the posterior distribution to make inference about model parameters or to make predictions.
MCMC is Monte Carlo integration that draws samples from the required distribution by running a properly constructed Markov chain for a long time.
Gibbs sampling is usually used for taking samples.
BUGs is an acronym stand for Bayesian inference using Gibbs sampling with WinBUGS as an open source software package for performing MCMC simulation.

By: Mohammad Pourgol-Mohammad, Ph.D, P.E, CRE, mpourgol@gmail.com

Previously published in the December 2015 Volume 6, Issue 4 ASQ Reliability Division Newsletter

Picture © B. Poncelet https://bennyponcelet.wordpress.com

Joseph Juran (1904-2008) and W. Edwards Deming (1900-93), the two most influential thinkers behind the totalquality movement, both launched their careers a few years apart at Western Electric, which used Statistical qualitycontrol techniques pioneered at Bell Labs to build reliable telephones.
And both gained acclaim while on loan to the government during World War II.
The irony is, Japanese execs heeded the lessons of total quality ahead of American managers.
In 1969, JUSE asked Juran to lend his name to Japan’s top quality award, a sort of super-Deming Prize for companies that maintain the highest quality for five years running.
JUSE deemed Juran’s vision of top-to-bottom quality management even more important than Deming’s manufacturing insights.
Juran demurred-a decision he later regretted.
So what could have been the Juran Medal is instead called the Japan Quality Control Medal.
There is a Joseph Juran Medal, though. It’s awarded by the American Society for Quality.
Juran personally presented the first one in 2001 to Robert W. Galvin, then head of Motorola Inc.’s executive committee.

Previously published in the June 2016 Volume 7, Issue 2 ASQ Reliability Division Newsletter

Picture © B. Poncelet https://bennyponcelet.wordpress.com