Physics-of-Failure Modeling

Physics-of-Failure Modeling


The traditional approach to product reliability has largely been based on the statistical analysis of an item’s failure data. Analysts would predict the probability of a part/system failure from the percentage of times that an item failed either during testing or from historical use in the field. Eventually, statistical distributions were fit to an item’s time-to-failure data, allowing for the estimation of more informative reliability characteristics (e.g., characteristic life, rate of failure, etc.). While statistical methods have been unquestionably valuable, and remain popular to this day, some have criticized the fact that they do not address the actual mechanisms of failure. Though one can determine the probability of an individual failure mode’s occurrence (i.e., modal failure rates) using statistical methods, these estimates still do not consider the underlying root cause of the failure. Over the course of time, reliability practitioners began to develop new physics-based techniques to alleviate these concerns.

*Available as a PDF Download File Only


Product Description

The series name has a double meaning:
(1) the spelling of the name r-e-l-e-a-s-e indicates that the series is intended to “release“ the non-expert down the path to reliable products and
(2) the complementary terms “REL” with “ease”, implying the series goal of “reliability made easy.”

Of course, reliability success is seldom easy, requiring expertise and tailoring with tradeoffs addressing life-cycle costs and other issues, but we hope that the series will help those not familiar with reliability practices understand the basics.

While it is unlikely that the reader will become an instant expert in reliability by reading the RELease guides, it is likely that he/she will gain a better appreciation of the basic tools that lead to designing and building reliability into products and systems. The number of pages in each guide is intentionally limited to address only the basics, with comprehensive authoritative references listed for those wanting to know more. The initial set of guides will be continuously expanded in the future. Please let us know how the series can be improved to meet your needs in introducing reliability to the non-expert, or suggest other topics you would like to see developed.

Additional information

File Type:



Alex MacDiarmid





Publication Date:

July 2013

Table of Contents

1. What is Physics-of-Failure (PoF)       2
2. Physics-of-Failure Terminology       4
3. The Physics-of-Failure Process       5
  3.1. Required Inputs     5
  3.2. PoF Process Flow     6
  3.3. Alternative Prediction Methods     6
4. Performing a PoF Analysis       8
  4.1. Model Search/Selection     8
  4.2. Tailor Existing Model     11
  4.3. Develop New Mathematical Model     12
    4.3.1. Model Validation   14
  4.4. Perform Prediction     15
5. Examples       15
  5.1. Wear Model Reliability Prediction     15
  5.2. TDDB Model Reliability Prediction     17
6. For More Information       18