Skip to main content

FMEA Types and steps involved

FMEA means Failure Mode and Effects Analysis. There are 3 types of FMEAs. 


  1. System FMEA (SFMEA)
  2. Design FMEA  (DFMEA)
  3. Process FMEA  (PFMEA)

SFMEA is for System and used to analyze the failures of the end product in a customer's system. It focuses on the potential failure modes associated with the functions of the end product. 

DFMEA begins with the development of information to understand the system, subsystem or component being analyzed and defines their functional requirements and characteristics. It focuses on potential failure modes associated with the function of each system, subsystem or component being analyzed. 

PFMEA is used to analyze manufacturing and assembly processes. It focuses on process inputs. 

FMEA Processes - Steps Involved 

There are 13 steps in creating an FMEA. Please find the attached diagram.


Step 2: Define the Function 
What is the system, the sub system or process steps supposed to do in view of customer requirements?

Step 3: Determine Potential failure modes
Possible categories in which a function can go wrong:- No function, Partial function, Intermittent function and Unintended function. 

Step 4: List Potential failure effects 
Examples: Noise, Loss of fluid, Loss of function, Loss of system, Rough surface, poor appearance, Customer dissatisfied, Intermittent operations, Unpleasant odor, Potential safety hazard, No/Low output. 

Step 5: Assign Severity Ratings
Assess the seriousness of the effect.

Step 6: List all potential causes of the failure mode.


Step 7: Assign Occurrence Ratings
Rate the occurrence based on the likelihood that a specific root cause will occur.

Step 8: Define Current Design/Process Controls
Detecting the cause and subsequent failure mode prior to production and prevent the cause from occurring. Example- Lab test, Load test, Field test, Fleet test, Design review, Final test, Gauging, Operator training, Method sheets, Work instructions, Maintenance, Environmental test, Fail/Safe features, Warnings, Error proof by design, setup verification etc.

Step 9:  Assign Detection Settings.
Rate the detection based on each detection control mentioned.


Step 10: Calculate RPN
RPN = Severity of the effects * Probability of occurrence * Ineffectiveness of design/process verification.

Step 11: Develop an Action Plan to address high RPNs
Order of priorities :
  1.  High Severity
  2.  High Multiple
  3.  High RPN
Note: RPN threshold values should not be a target.

Step 12 - Take Action
  •  Define the necessary actions in order to reduce occurrence and/or detection of the failure mode.
  • Redesign the product to reduce severity.
  • Determine the list of recommended actions.
  • Assign a person responsible with a due date foe each task.
Examples of Recommended actions: Revise design, revise test plan, revise material specification, reliability testing, finite element analysis, training, process instructions etc.

Step 13: Reevaluate the RPN
Reevaluate all the three ratings - severity, occurrence and detection and calculate the new RPN.

Reason for the Failure of FMEAs

  • Only one person is assigned to complete the FMEA.
  • Not customizing the rating scales with business units specific data, so that they are meaningful.
  • The design or process expert is not included in the FMEA or is allowed to dominate the FMEA team.
  • Members of the FMEA team are not trained in the use of FMEA, and become frustrated with the process.
  • The FMEA team becomes bogged down with minute details of design or process, losing sight of the overall objective.
  • Rushing through identifying the failure modes to move on to tht next step of the FMEA.
  • Listing the same potential effect for every failure.
  • Stopping the FMEA process when the RPNs are calculated and not continuing with the recommended actions.
  • Not reevaluating the high RPNs after the corrective actions have been completed.

Thank you for reading this article. 






 






Comments

Post a Comment

Popular posts from this blog

Introduction to IoT ecosystem

Basics of IoT IoT (Internet of Things) refers to a network where devices communicate to devices and devices communicate to software applications over the internet . This web-enabled devices/internet enabled devices will usually collect data, send and act upon data from their surrounding environment. •       The IoT network consists of Edge side (sensors, actuators and controllers connected to the internet) and Cloud side. •       The devices on the Edge side can either connect to the internet directly or using an Edge (IoT) gateway . The Edge gateway is used for: •        Providing a connection to the internet for devices that cannot connect to the internet on their own. These devices don’t have feature to connect to the Internet. •        Local functionality/control that must work even without internet availability •       ...
Post a Comment
Read more

Introduction to Modelling and Simulation

What is about Today?    Modeling & simulation is use of the physical, mathematical or logical representation of a system , entity or process and these models  reduces the development time and cost of the system and it optimizes the design.  Simulation is the imitation of the real-world process or system. To simulate something, it requires a model to be developed, and these model represents the key characteristics, behaviors and functions of the selected system or process.Modelling an act of building a model and Simulation is act of using a model. Simulation software is based on the process of modeling a real phenomenon with a set of mathematical formulas.  Advantages: Allows to understand how the system operates in real-time system in virtual environment. Easy to make changes in to the system and their effect on the output. Can determine the system requirements by applying different configurations. Performing bottleneck analysi...
Post a Comment
Read more

Introduction to Deep Learning

Deep Learning is a type of machine learning in which a model learns to perform classification tasks directly from images, text or sound. Its usually implemented using a neural network architecture. Here deep refers to the "Number of layers in the network" i.e the more layers, the deeper the network. Traditional neural network contain only 2-3 layers, but in deep network contains hundreds of layers. Deep learning example is  A self-driving vehicle slows down as it approaches a pedestrian network . Deep learning is especially well-suited to identification applications such as face recognition, text translation, voice recognition and advanced driver assistance systems and traffic sign recognition etc. Deep learning is growing faster because it really works.  What is the Difference Between Deep Learning  and Machine Learning? Deep learning is a sub-type of machine learning. For example with machine  learning, you manually extract the relevant features of...
Post a Comment
Read more