FMEA is a structured methodology for thoroughly thinking through and correcting potential problems in an engineering design or planned process before in-service implementation. If done well, the procedure can greatly increase reliability by reducing or eliminating failure scenarios. Ideally the procedure is completed very early in the design or planning process when relatively inexpensive corrective decisions can be made and not later after costly in-service failures occur. It is also valuable for existing designs before significant modifications are made.

The procedure begins with the user identifying all potential failure modes. Then a numerical product of essential factors for each identified failure mode is developed. This product is known as the Risk Priority Number (RPN). Relative values of RPN allow a rank ordering of the probability and consequences of different failure modes. This permits the engineer or planner to prioritize corrective measures and effectively allocate limited resources for improvements. The RPN’s factors are as follows:

RPN = S x O x D,    where  S = the relative severity or consequence of the given failure,

O = the probability of occurrence of the given failure, and

D = the probability of not detecting and correcting the failure before it occurs.

Each factor is assigned a value on a 1 to 10 scale and resulting, large RPN values receive first priority for corrective measures. A high severity or consequence of the given failure requires a high value of S, a significant probability that the failure mode will take place (with the given parameters in place) results in a high relative value of O and a high probability of not detecting the failure mode prior to its occurrence dictates a high value of D (again with the given parameters in place).

Obviously the calculation of each Risk Probability Number is simple. The difficult parts of completing a useful FMEA are identifying all the possible failure modes and then defining the appropriate values of S, O and D for each type. To gain the most value the user has to have significant knowledge of all aspects of the application plus the effects of the planned parameters and then assign suitable values to the three input factors of the RPN’s. Assumptions based on the best available information have to be employed. It is always desirable to have input to the analysis from multiple personnel with different specialized knowledge. Nothing is gained if the FMEA involves only adding numbers to a form, spreadsheet or custom software with little or no real analysis of the application.

Implementing the procedure largely depends on the complexity of the given situation. For example, using FMEA in designing the nose landing gear for a large jet aircraft is not simple but using the procedure for all the interacting mechanical, electronic, etc. systems on that aircraft  presents a step change in required effort. A very simple application might be completed using a basic home-made form. More complex applications are normally handled with specifically designed software from specialty vendors. Correctly using the software usually requires training with the developer. In the many applications the procedure is repeated after each modification of design or process parameters. The iterative process continues until the initially large and medium magnitude RPN values are greatly reduced.    

The essential key to successful use of FMEA is complete “buy in” and advocacy of the procedure by management. Without that direction the extra effort necessary for a good, thorough analysis likely will not be expended. This early extra expenditure can provide great future value. Avoiding just one unplanned shutdown or disaster can easily pay for the up-front cost of a good FMEA. Another major benefit of the analysis is that it provides clear evidence of an organization’s due diligence in the safe implementation of a given design or process. Having a record of that effort will be very useful if the organization is sued because of a rare type of failure or other event it could not have reasonably anticipated.                   

Posted in: Industrial/Training Services