Failure modes and effects analysis (FMEA)

Description

Failure modes and effects analysis (FMEA) is a systematic method for identifying and preventing all possible failures in a design, a manufacturing or assembly process, or a final product or service [2, 6, 7, 18].

 

When to use

  • When processes, products/services are designed or redesigned
  • When existing processes, products/services are applied in new way
  • When goals for an existing process, product/service are improved
  • When failures of existing processes, products or services are analyzed

 

FMEA Types

There are two basic types of FMEA: design FMEA and process FMEA:

  • The design FMEA (DFMEA) deals with components, assemblies, parts, and other aspects of a product’s design. The DFMEA objective  is the elimination of the failures caused by poor design.
  • Process FMEA (PFMEA) deals with the steps of the process. The PFMEA objective is elimination of failures caused by the process. It is also used for identification of process variables that must be controlled.


Procedure

  1. 1. A multifunctional team with diverse knowledge about the process, product/service, and customer needs must be gathered
  2. 2. The scope of the FMEA must be Identified. What is the purpose of FMEA - concept, system, design, process, or service? What are the boundaries of FMEA? How detailed should be the FMEA?
  3. The functions of the scope must be identified. This can be accomplished by asking a questions like “What is the purpose of the system/design/process/service?
  4. The potential failure modes for each function must be identified. This means all the ways that the function could fail must be described.
  5. The potential effects of failure must be identified. This can be done by asking a question “What happens when this failure occurs?” for each failure mode.
  6. The significance of each effect must be determined – this is the severity ranking (S) for each failure mode. Severity is usually in the range from 1 to 10 (Table 4.1):
Table. 4.1 Severity Evaluation Criteria
  1. Severity Evaluation Criteria
    Rank Effect
    1 No effect
    2 Annoyance
    3
    4
    5 Loss or Degradation of Secondary Function
    6
    7 Loss or Degradation of Primary Function
    8
    9 Failure to Meet Safety and/or Regulatory Requirements
    10

  2. All the potential root causes for each failure mode must be determined. All possible causes for each failure mode must be included in the FMEA blank.
  3. The occurrence ranking (O) for each cause must be determined. This ranking represents the probability of occurring a failure. Occurrence is also rated in the range from 1 to 10 (Table 4.2):

Table 4.2 Occurrence Evaluation Criteria
  1. Occurrence Evaluation Criteria
    Rank Likelihood of Failure Defective items
    1 Very Low Failure is eliminated through preventive control
    2 Low 1 in 1000000
    3 1 in 100000
    4 Moderate 1 in 10000
    5 1 in 2000
    6 1 in 500
    7 High 1 in 100
    8 1 in 50
    9 1 in 20
    10 Very high 1 or more in 10

  2. The current process controls for each cause must be identified. In fact these are the tests, procedures, or mechanisms that keep failures from reaching the customer.
  3. The detection ranking (D) for each control must be determined. This rating estimates how well the current process controls can detect the failure modes. Like severity and occurrence, the detection is also in the range from 1 to 10 (Table 4.3):

Table 4.3 Detection Evaluation Criteria
  1. Detection Evaluation Criteria
    Rank Likelihood of detection Opportunity for detection
    1 Almost Certain

    Detection not applicable; Error Prevention
    2 Very High Error Detection and/or Problem Prevention
    3 High Problem detection at source
    4 Moderately high Problem detection prost processing
    5 Moderate Problem detection at source
    6 Low Problem detection prost processing
    7 Very Low Problem detection at source
    8 Remote Problem detection prost processing
    9 Very Remote Not likely to detect at any stage
    10
    		 No detection opportunity Almost Impossible

  2. The Risk Priority Number (RPN) is calculated.

    \( RPN = O \ast S \ast D \) (4.1)

  3. The recommended actions must be identified. They could be design or process changes to lower severity or occurrence. They also could be additional controls for detection improvement. Very important is responsible persons for the actions and target completion dates to be noted.
  4. New values for severity, occurrence and detection are set. The new value of RPN is calculated.