1. Congratulations on the success of Practical Reliability Engineering, now in its fifth edition which was published in 2012. How did the book come about in the first place?

Patrick O’Connor wrote the first editions of the book and I came in later after he moved into a retirement phase. Reliability engineering is a fairly young discipline, so at the time of the first edition it was in the early stages of development. The books on reliability at the time were mostly written by college professors and were mainly statistical books. Statistics is great; however, if you are a practicing reliability engineer and for example writing an environmental test plan, statistics will only be of limited use to you.

So Patrick O’Connor wrote a book for engineers as a practical guide, involving statistics, which is extremely important to reliability engineering, but also including testing methodology, reliability modelling, accelerated testing, case studies, reliability management, and several other topics. A book like this did not exist before, which made it an instant success.

2. What were the primary objectives that you and your co-author Patrick O’Connor had in mind when originally writing the book?

Reliability engineering is a fairly young field and it was launched by the military around 1950s during the Cold War. The military wanted reliable products, they wanted to be sure that a weapon wouldn’t jam or a missile wouldn’t misfire, hence they contributed a lot to the development of reliability engineering.

Reliability engineering is a mix of different disciplines – statistics, material science and physics of failure (i.e. how things fail, was it due to fatigue, overstress, corrosion, etc.) There was a big gap in reliability education, which needed to be filled and the first edition of this book was intended to fill that gap, as it turned out, quite successfully.

3. Who should read the book and why?

For starters, this book would appeal to graduate students and senior undergraduate students as it is also a textbook, which in fact, has been accepted by more than 30 universities worldwide. I am an adjunct professor at Purdue University and we ourselves use this book for a systems engineering course, since reliability as an important characteristics of any system.

Reliability engineers just starting their careers will probably benefit most from reading this book. We have a joke that reliability engineers are not trained, they are rather grown. What we mean by that is paradoxically, reliability engineering education is very much lacking in the college curriculum. Handfuls of schools in the entire USA have programs on reliability engineering and the situation is not any better in Europe or Asia. So basically what is happening is that a lot of young reliability engineers will receive their training on the job, attending professional seminars, and collaborating with their colleagues. This is where this book will be very helpful in augmenting this education.

I believe this book will also help seasoned reliability professionals. I have received positive feedback from people practicing in the field for 20 or more years and they have commented that they found new and useful information in the book. Also design engineers (EE, ME, SYS) will find it useful, since reliability is usually part of most product specifications, which needs to be addressed at the product level.

4. The fifth edition was released in 2012. For those who have not yet been introduced to the book who will read this, what can the reader expect in the latest version?

Like any other discipline, reliability engineering continues developing: there are new technologies, new products, new environments, new materials, etc. Therefore naturally, every new edition adds a new material or revises old one in order to keep up with the evolving discipline and all the new developments.

It has been over 30 years between the first edition and the fifth. Computers are becoming more popular and more powerful, electronics is becoming smaller, quality and reliability expectations getting higher, warranties are getting longer, and so on. Although the core reliability material, which is timeless, remains the same, we keep introducing state of the art and state of practice changes to the book to keep it current and most importantly relevant.

5. Why is the book still of particular interest?

First of all, as I mentioned before, we keep the core reliability material in, but continue reinforcing the strength of this book by editing old material or adding new one. For example, I work in the electronics industry, which is very dynamic. For example, in the early 2000-s the RoHS legislation was passed in Europe intended to remove all the toxic materials from electronics, most noticeably transitioning to the use of lead-free solder. This created tremendous impact on the reliability of electronics, since now we are dealing with a new soldering process and a new material with totally different properties. All this needs to be reflected in the book to keep it current. Even small changes like augmenting or replacing some of the statistical tables with Excel formulas and solutions help to keep this book current.

Couple of years ago, I have developed a training program in my company to train advanced reliability experts for our global sites and we use this book a lot. One of the chapters that I added specifically to the fifth edition was the Monte Carlo simulation using the Microsoft Excel. Years ago, you needed sophisticated software and powerful computers to run a Monte Carlo simulation. Now I show my students that all you need is Excel spreadsheet and a little bit of knowledge.

Despite the fact that everyone wants to see how things actually work, dealing with failures and helping a team to design products which will not fail, makes reliability engineering a hugely interesting and satisfying field and I encourage everyone to explore it.

6. What is the role of a reliability engineer and how are they different from other engineers?

To answer that, let me give you a humorous quote: ‘Reliability engineers are a sad and embittered race; a lonely group despised by both the design team and management. Their sole function being to generate failures and generate failures, they will. For it is their very life, their ambrosia, their reason for being. Many good designers have quietly disappeared after receiving one too many failures. Management have lost their promotions because the reliability growth curve didn’t grow. No wonder the poor reliability engineer eats alone and has no friends.’ I don’t know who wrote that but when I read it to my students, it always gets a good chuckle.

But in reality, it’s not that bad! Basically the role of reliability engineers is multi-faceted. They contribute to the programs many different ways: for example by working with design teams in an advisory role, e.g. by suggesting to place in a heavy component closer to the mounting point on the circuit board to reduce the effect of vibration. Also they often contribute by testing the product on the resistance to the environments such as rain, heat, cold, humidity, etc. Sometimes reliability engineers have to assess the probability of a product failure in the field, conduct a risk assessment and other similar duties.
These days products are expected to be reliable and requirements keep getting more and more stringent – for example, in the automotive industry, we design electronics for your car. At one time, it was expected that the lifetime of a car radio would be 5 years, then gradually it changed to 10 years, then to 15 years – it is the job of the reliability engineer to adjust and help the team to design the product to last that long and also update the testing practices to accommodate that longer mission life e.g. exposing the product to a harsher environment.

7. When and how does statistics play a role in reliability engineering?

Statistics is one of the integral parts of reliability engineering. A lot of products are produced in large quantities (cell phones, DVD players, cars, etc.) and we cannot test every single one of them. Therefore, we only deal with a sample and somehow we need to ensure that all those millions of units in the field actually perform well, even though we can only test a limited number of them. That is where statistics comes to our rescue, because whatever we observe during the product testing (most often accelerated testing) we need to translate to the product performance in the field and a risk assessment. And there are a number of statistical methods to help us with that.

Another big application of statistics is warranty analysis, both past and future. Often before we launch a product, management will request warranty assessment for this product in order to make appropriate warranty funds accruals to pay the future claims. In fact many countries have legislations requiring these kinds of allocation of funds and we extensively use statistics for these kinds of estimations.

8. You currently work at the Delphi Corporation as an Engineering Manager. Please could you tell us more about these roles?

My official title is Global Reliability Engineering Leader, which is more of a technical management role. I have a team of reliability engineers distributed throughout the world. These engineers report to me functionally and I provide guidance in terms of the latest reliability engineering practices including testing, analysis, statistical methods, roadmaps, and future product developments.

I also support a number of engineering programmes helping product teams to understand failure modes and failure mechanisms, develop test plans, analyse the results and often to develop corrective actions. I also perform statistical data analysis of test results and field returns and assess the risk of product failures in the field.

9. What is your current research focussing on? What are your main objectives and what do you hope to achieve through the results?

As an engineering scientist working in the industry I pursue a number of different projects with specific practical applications. One topic, which might be of interest to your readers, is reliability demonstration by test. Most of the product specifications have reliability requirements, such as 99% reliability with 90% confidence or something similar. However, since we can only test a limited number of samples some of the requirements become unrealistic from statistical standpoint. With my sample size I can, for example, demonstrate 97% reliability with 50% confidence, which is obviously not very impressive. By the way, any statistician will laugh you out of the room if you mention 50% confidence.

One of the focuses of my research is to improve or resolve that situation. For example, we can apply Bayesian statistics and utilise prior knowledge about the product, such as test results on the previous models or field returns for a similar product. That would help us to improve our reliability metrics while testing the same sample size. Another approach is to assess the testing conditions, which are often extreme and reassess the population reliability. For example, our vehicle vibration test often corresponds to the 99.8th percentile driver, a person who lives in a rural area and drives a lot on unpaved roads. However, most of the drivers will see more benign conditions, since the population of the vehicle has a wide distribution of users ranging from an old lady who only uses her car to drive to a church on Sundays to a pizza delivery driver in a rural area. Therefore the reliability of the whole product population will be much better than what we demonstrate for an ‘extreme’ driver. The methods I developed help us to provide more realistic reliability demonstration estimates.

10. You have also co-authored another for Wiley, Applied Reliability Engineering and Risk Analysis: Probabilistic Models and Statistical Inference. What will be your next book undertaking?

This was my third book, but it actually spun out from my activity as the editor of the Wiley Series in Quality in Reliability Engineering. A group of authors wanted to commemorate the centennial of the birth of a renowned Russian statistician and reliability theorist, so they approached me with this book idea and invited me to contribute as well. This book contains a number of different chapters on various reliability and statistical analysis topics including risk analysis, accelerated testing, Markov analysis, multistage solutions and many others.

As far as my book projects, at the moment I am trying to focus on my responsibilities as a Series editor for Wiley and there are several projects that I am currently pursuing with other authors, such as green reliability, design for reliability, obsolescence and a few others. I would also personally like to write a book on warranty forecasting and management but don’t have a specific timing for that project.

I would very much welcome statisticians who are interested in writing books on statistical data analysis, warranty data analysis, big data or any other topics with applications to quality or reliability engineering.

11. As you mentioned, you are also the Series Editor for the Wiley Series in Quality and Reliability Engineering, for which there are ten books currently published. Are you able to give us a sneak preview of those in the pipeline?

We have four books coming out and we also are negotiating a few others. One of the books in process is on green reliability, which covers various reliability aspects of renewable energy sources, like wind turbines and solar cells. Another book is on accelerated and highly accelerated testing methodology, which is becoming increasingly popular these days, since it helps to achieve better reliability and shorten a product development cycle.

Another book is focusing on design for reliability. Many product development teams still use the old approach we call ‘test, analyse and fix’. You build the product, test it, find problems, fix them, test again and basically repeat the cycle until you get it right. However this approach is not very efficient and takes too long. Design for reliability process intends to get the design right the first time and minimize the amount of testing and consequently get the product to market faster.

12. What are you looking for in terms of contributions to the series? Are there any hot areas that you’d like to see the series publish in?

Absolutely – I would very much welcome statisticians who are interested in writing books on statistical data analysis, warranty data analysis, big data or any other topics with applications to quality or reliability engineering.

Another topic I am interested in commissioning a books is functional safety. It is less related to statistics, but covers new standards coming out of Europe, regulating how you design products related to human safety, e.g. vehicles, trains, heavy duty equipment, where a human can potentially be seriously hurt. These standards cover a large number of design rules and regulations but this is the reality that we need to deal with and I think this would also make an excellent addition to the series.
In conclusion, I’d like to say that reliability engineering is a very exciting field, despite the fact that I jokingly call it ‘a dark side of engineering’, where you study when and how things fail. Despite the fact that everyone wants to see how things actually work, dealing with failures and helping a team to design products which will not fail, makes it a hugely interesting and satisfying field and I encourage everyone to explore it.