Reading Assignment:

Download and read Chapter 3 in Essentials of Quality With Cases and Experiential Exercises.  Review the Discussion Questions at the end of the chapter to be sure that you understand what you have read.

Earlier in Chapter 3 we discussed the concept of robustness. In this part of Chapter 3 we discuss a related concept -- reliability. Both robustness and reliability are primarily quality of design issues.

Reliability can be defined as a probability in two ways:

The probability that a product will perform on a given trial.  This would be most appropriate for a product such as an automobile ignition switch.

The probability that a product will perform for a specified length of time.  This would be most appropriate for a product such as a light bulb.

Two other dimensions of reliability are the definition of a failure and the prescribed operating conditions.  What is a failure?  When you see a few spots (dropouts) on a video recording, does that constitute a failure, or is it only a failure when you can no longer see the picture?  Can we expect the video cassette to perform well in Antarctic weather conditions, or is it designed to operate in a specified temperature range (prescribed operating conditions)?  These are issues that must be properly addressed.

While at NASA recently, I stopped in and counted all the components of the space shuttle that have to function in order for the ship to make a successful trip.  I counted 1,000,000 components--a failure of any one of these and the mission fails (i.e. it is a serial system).  If the reliability of each component was 0.999, what is the probability of a successful mission?  To calculate the reliability, raise 0.999 to the 1,000,000 power using the y^x function on your calculator.  The answer is 0.999^1,000,000 = 3.077x10^-435--essentially zero.  Would you fly on that shuttle?  Of course not.  Well, how about if the component reliabilities were 0.99999 -- 1 failure per 100,000 opportunities?  The reliability is 0.99999^1,000,000 = 4.54x10^-5--still very low.  Again, you would elect not to fly the mission.  How about components that only fail once per 10,000,000 opportunities--i.e. component reliabilities of 0.9999999?  The reliability is 0.9999999^1,000,000 = .905--9 times out of 10 the mission will be successful.  I still wouldn't fly on that shuttle!

Well, just how reliable can we make each component?  Must we have perfection in each component to have established such a good safety record (only two failed missions) with the space shuttle?  The answer is no.  Reliabilities of a serial system can be increased by the incorporation of redundant components or systems.  A redundant component is often referred to as a backup.  All of the essential components on the space shuttle have at least one backup.  If the primary component fails, the backup takes over and the system continues to operate.  By providing redundant components we can greatly enhance system reliability and ensure that the space shuttle completes its mission and returns safely.

The negative exponential distribution is useful for modeling reliabilities of many systems. In conjunction with mean time between failure (MTBF) or mean time to failure (MTTF) data, this distribution can be used to determine the probabilities of failure or no failure before a specified time in service.

The reliability of other systems is better modeled using the normal distribution. This tends to be true for basic systems whose failures cluster around a failure point.

Failure Mode and Effects Analysis (FMEA) and Fault Tree Analysis (FTA) are tools used to examine the ways in which a system can fail and identify the potential causes of that failure.

Product or component traceability through the use of serial numbers or lot numbers can simplify product failure analysis. Traceability can also facilitate product recalls.

Answer Discussion Questions 5-16 at the end of Chapter 3 of the text.

Do Problems 1-11 at the end of Chapter 3 of the text.

Do Exercise & Activity 2 at the end of Chapter 3 of the text.