Failure mode and effect analysis, or FMEA, is an attempt to delineate all possible failures, their effects on the system, the likelihood of occurrence and the probability that the failure will go undetected. FMEA provides an excellent basis for classification of characteristics such as identifying CTQs and other critical variables. As with Pareto analysis, one objective of FMEA is to direct the available resources toward the most promising opportunities. An extremely unlikely failure, even one with serious consequences, may not be the best place to concentrate preventative efforts. FMEA can be combined with decision analysis methods such as Analytical Hierarchy Process (AHP) and quality function deployment (QFD) to help guide preventive action planning.

FMEA came into existence on the space program in the 1960s. It was later incorporated into military standards, in particular Mil-Std-1629A. There are two primary approaches for accomplishing an FMEA:

• The hardware approach—Lists individual hardware items and analyzes their possible failure modes. This FMEA approach is sometimes used in product DFSS projects.
• The functional approach—Recognizes that every item is designed to perform a number of functions that can be classified as outputs. The outputs are listed and their failure modes analyzed. This approach to FMEA is most common on both DMAIC and DMADV projects involving improvement of processes or complex systems.

The FMEA process
FMEA is an integral part of early design process and it should take place during the improve phase of DMAIC or the design phase of DMADV. FMEAs are living documents that must be updated to reflect design changes, making them also useful in the control or verification phases. The analysis is used to assess high-risk items and the activities underway to provide corrective actions. The FMEA process is also used to define special test considerations, quality inspection points, preventive maintenance actions, operational constraints, useful life, and other pertinent information and activities necessary to minimize failure risk. All recommended actions that result from the FMEA must be evaluated and formally dispositioned by appropriate implementation or documented rationale for no action. The following steps should be used when FMEA is performed:

1. Define the system to be analyzed. Complete system definition includes identification of internal and interface functions, expected performance at all system levels, system restraints and failure definitions. Functional narratives of the system should include descriptions of each goal in terms of functions identifying tasks to be performed for each goal and operational mode. Narratives should describe the environmental profiles, expected cycle times and equipment utilization, and the functions and outputs of each item.
2. Construct process maps. These should illustrate the operation, interrelationships, and interdependencies of functional entities.
3. Conduct SIPOC (suppliers, inputs, process, outputs, customers, requirements) analysis for each subprocess in the system. All process and system interfaces should be indicated.
4. List the intended function of each step in the process or subprocess.
5. For each process step, identify all potential items and interface failure modes. Also define the effect on the immediate function or item on the system and on the mission to be performed for the customer.
6. Evaluate each failure mode in terms of the worst potential consequences that may result and assign a severity classification category (SEV). (See Figure 1.)
7. Determine the likelihood of occurrence of each failure mode and assign an occurrence risk category (OCC). (See Figure 1.)
8. Identify failure detection methods and assign a detectability risk category (DET). (See Figure 1.)
9. Calculate the risk priority number (RPN) for the current system. RPN = SEV * OCC * DET.
10. Determine compensating provisions for each failure mode.
11. Identify corrective design or other actions required to eliminate failure or control the risk. Assign responsibility and due dates for corrective actions.
12. Identify effects of corrective actions on other system attributes.
13. Identify severity, occurrence and detectability risks after the corrective action and calculate the subsequent RPN.
14. Document the analysis and summarize the problems that couldn’t be corrected and identify the special controls necessary to reduce failure risk.

RPNs are useful in setting priorities. Larger RPNs should receive greater attention than smaller RPNs. Some organizations have guidelines requiring action based on the absolute value of the RPN. For example, Boeing recommends that action be required if RPN > 120.

A worksheet similar to the one shown in Figure 2 can be used to document and guide the team in conducting an FMEA. FMEA is incorporated into software packages, including some that perform QFD. There are numerous resources available on the Web to assist you with FMEA, including spreadsheets, real-world examples of FMEA, and much more. (Visit www.fmeainfocentre.com).