Some defense-related applications require a special type of criticality analysis, called Quantitative Criticality Analysis to supplement FMEA applications. This is the “C” in what is called FMECA: Failure Mode, Effects and Criticality Analysis. I’ll shorten Criticality Analysis to CA in this article.
What is Quantitative CA? When and why it is used? Can Quantitative Criticality Analysis be used in commercial applications?
First some background
The following excerpt is from a 2006 Department of Army manual, describing the history of FMECA.
The FMECA was originally developed by the National Aeronautics and Space Administration (NASA) to improve and verify the reliability of space program hardware. The cancelled MIL-STD-785B, entitled Reliability Program for System and Equipment Development and Production, Task 204, Failure Mode, Effects and Criticality Analysis calls out the procedures for performing a FMECA on equipment or systems. The cancelled MIL-STD-1629A is the military standard that establishes requirements and procedures for performing a FMECA, to evaluate and document, by failure mode analysis, the potential impact of each functional or hardware failure on mission success, personnel and system safety, maintainability and system performance. Each potential failure is ranked by the severity of its effect so that corrective actions may be taken to eliminate or control design risk. High-risk items are those items whose failure would jeopardize the mission or endanger personnel. The techniques presented in this standard may be applied to any electrical or mechanical equipment or system. Although MIL-STD-1629A has been cancelled, its concepts should be applied during the development phases of all critical systems and equipment whether it is military, commercial or industrial systems/products.
Types of FMECA
There are two primary types of FMECA. One is Quantitative FMECA, and the other is Qualitative FMECA. Both types of FMECA use a defined criticality analysis. They are similar in procedure, with the exception that the Quantitative FMECA uses a Quantitative CA and the Qualitative FMECA uses a Qualitative CA. Each type of analysis is described below.
What is Quantitative CA?
Quantitative CA is a series of calculations to rank items and failure modes according to a formula covered below. To use Quantitative CA to evaluate risk and prioritize corrective actions, follow these five steps.
Each step is followed by an example using a bicycle brake pad to illustrate the application. Example assumptions: time frame = 5 years; customer usage for high end user = 3 hours/day or 5,475 hours over the life of the brake pad. Assumed failure rate = 0.0001 failures/hour.
Step 1. Calculate the expected failures for each item.
This is the number of failures estimated to occur based on the reliability/unreliability of the item at a given time. Reliability is the probability that an item will perform a required function without failure under stated conditions for a stated period of time. Unreliability is one minus reliability. The “time” for the calculation is most often the target or useful life of the item. With an exponential distribution, expected failures is calculated by multiplying the failure rate by the time (?t); but it is estimated differently for other distributions.
Care must be taken to ensure calculations for reliability/unreliability and expected failures are based on correct failure distributions. Some practitioners assume an exponential distribution (constant failure rate); however, this assumption is not always valid. It is wise to solicit the support of a reliability engineer or other practitioner who is well experienced in these types of calculations.
Step 1 example: Calculate the expected failures for the brake pad at 5 years. Based on assumptions, the number of failures at 5 years is 0.548 (5,475 hours multiplied by 0.0001 failures/hour)
Step 2. Identify the mode ratio of unreliability for each potential failure mode.
This is the portion of the item’s unreliability (in terms of expected failures) attributable to each potential failure mode. In other words, this represents the percentage of all failures for the item that will be due to the failure mode under consideration. The total percentage assigned to all modes must be equal to 100%.
The failure mode ratio of unreliability can be based on reliability growth testing data for the current design, field data and/or test data from a similar design, engineering judgment (“best guess”), or apportionment libraries such as MIL-HDBK-338B. Exercise care if engineering judgment is used because the intent of a criticality calculation is an objective number, not a “best guess” number. Apportionment libraries are often only rough approximations, and should only be used if one is confident of their validity for the given application.
Step 2 example: In this example, there are two failure modes: excessive wear (85%) and cracking (15%).
Step 3. Rate the probability of loss that will result from each failure mode that will occur.
This is the probability that a failure of the item under analysis will cause a system failure. The following are guidelines from Mil-Std 1629A for establishing the probability of loss for the criticality calculation.
Actual Loss 100%
Probable Loss > 10% to < 100%
Possible Loss > 0% to < 10%
Step 3 example: In this example, the probability of loss of the system due to excessive wear is 75% and due to cracking is 15%.
Step 4. Calculate the mode criticality for each potential failure mode.
This is the product of the three factors:
Mode Criticality = Expected Failures (for the item) × Mode Ratio of Unreliability (for the failure mode) × Probability of Loss (for the failure mode)
Step 4 example: Mode Criticality for excessive wear = 0.548 × 0.85 × 0.75 = 0.349; Mode Criticality for cracking = 0.548 × 0.15 × 0.15 = 0.012
Step 5. Calculate the item criticality for each item.
This is the sum of the mode criticalities for each failure mode identified for the item. Item Criticality = SUM of Mode Criticalities.
Step 5 example: Item Criticality for the brake pad = 0.349 + 0.012 = 0.361
Below is an Example of Quantitative Criticality Analysis on a bicycle brake pad, from chapter 12 of Effective FMEAs.
When should Quantitative CA be used?
The majority of applications of Quantitative CA are driven by defense industry contracts. In addition, an organization may wish to benefit from the more detailed risk-ranking information from the Quantitative CA, provided there is sufficient objective failure data and time available to perform the more rigorous calculations. Some FMECA applications wish to use the failure rate information in Quantitative CA to calculate or predict reliability. However, caution is advised, as the assumptions that go into failure rate assessments can be misleading, such as the assumption of an exponential failure distribution.
What is Qualitative CA?
In FMECA applications, Qualitative CA is a type of risk prioritization that uses Severity and Occurrence in the form of a matrix to identify high, medium and low risk. It is the same as the Risk Prioritization step in FMEA (reference SAE J1739:2021), if the FMEA team is using an SO matrix. See Risk Prioritization in FMEA – A Summary.
Can Quantitative CA be used in commercial applications?
In most FMEA applications, there is a specific reason to do the FMEA, such as new technology or new applications of existing technology. See my article called “Preliminary Risk Assessment to understand to learn one approach to selecting FMEAs. https://accendoreliability.com/preliminary-risk-assessment/
Also, in most FMEA applications, an FMEA is done early in the product development process, when it can most effectively drive the needed changes.
It is my experience that when FMEAs are most needed to be done, there is not the objective (accurate) data available to use Quantitative Criticality Analysis.
What is the difference between FMEA and FMECA? The answer may surprise you. The next article explains the difference and offers a solution to simplify.