Does Your Company Know How to Use the Reliability Bathtub Curve?

Companies are often caught by surprise due to unexpected reliability issues that cost a lot of time and money to correct. In many cases, the surprises occur when reliability predictions, and even estimates based on reliability testing, fail to be realistic. Based on these negative experiences, some companies have concluded that the well-known “Reliability Bathtub Curve” does not exist or is not useful to help them make realistic estimates about their product’s behavior. “If the reality differs anyway,” they wonder, “why should one focus resources on reliability testing and predictions?”

As this presentation demonstrates, the “Reliability Bathtub Curve” does exist for many types of products but analysts sometimes make unrealistic reliability predictions because they have an improper understanding of how to use the curve. Their estimates are based only on the center part of the curve (often called the “useful life”) and ignore the portions of the curve in which the failure rate may be decreasing (due to “infant mortality”) or increasing (due to “wearout”). We will discuss the three key factors that determine the intrinsic (best possible) reliability of a product and how reliability analysis techniques can be properly applied to consider the full bathtub curve. The ways in which this correct perspective on reliability affects other management considerations (such as managing suppliers and validating reliability requirements) are also discussed.

Key Words: Bathtub, Reliability, Key items, Reliability Prediction, Reliability Verification, Purchasing