FMEA潜在失效模式及后果分析(英文版)(共19页).ppt

1、Failure Mode & Effect AnalysisExplanation and Use of FMEAQuality Forward SystemWhat is FMEA?FMEA is a problem solving process adopted by Snap-on as a key component of the Quality Forward System.Generally, FMEA is done by an on-site multifunctional team facilitated locally or by an SEQ Group Adva

2、nced Quality Engineer.Objective is to identify all of the reasons why quality of a process or product does not meet expectations.Uses severity, occurrence and detection ratings to determine corrective action priority for each possible failure reason.Quality Forward SystemWhen should a FMEA study be

3、conducted?As part of a Design/Development phase.When corrective action is needed due to unfavorable performance data (performance based on QFS metrics does not meet goals).When a need to make incremental improvement is identified by facility management or the SEQ Group.Quality Forward SystemFMEA Tea

4、m Start-UpTeam Selection done by local quality coordinator, business unit manager or, on request, SEQ Group Advanced Quality Engineer.Problem Statement A concise and focused description of the quality problem or opportunity for improvement, “what went wrong” or “what needs fixing”.Goal Statement a d

5、escription of what is to be accomplished by the team, in quantifiable terms “reduce by”, “increase by”, “eliminate”, etc. Good metrics to use include scrap cost, rework cost, or Defects Per Million Opportunities (DPMO), etc. For example, a goal statement could be “To reduce DPMO on product family X

6、from 3000 to 2000 by year end”.Complete FMEA Team Start-Up Worksheet (example located in procedure SEQ 80.10, Figure 1).Quality Forward SystemProblem Statement & Goal StatementProblem Statement and Goal Statement should consider:What went wrong?What is the gap between the desired quality level a

7、nd the actual quality level?Which process or processes are involved?Which department or cell is the deficiency occurring in?Is the problem cyclic, can it be attributed to specific events?What are the appropriate metrics to measure?What is the impact on operational fitness and profitability?What is t

8、he impact on the bottom line (reduced, increased, eliminated)?Are there constraints recognized early in the process? If so, have they been explored and clarified with the appropriate management levelQuality Forward SystemFMEA Process FMEAs uncover process problems related to the manufacture of produ

9、ct. Examples: a piece of automated assembly equipment may misfeed parts resulting in products not being assembled correctly, or, in a chemical manufacturing process, temperature or mixing time could be sources of potential failures resulting in unusable product. Think in terms of the five elements o

10、f a process: people, materials, equipment, methods and environment. With these five elements in mind, ask “How can process failure affect the product, processing efficiency or safety?” Nine steps in FMEA process.Quality Forward SystemStep #1 - Review the ProcessEnsure that everyone on the FMEA team

11、has the same understanding of the process being worked on.Review blueprint/engineering drawing of product (if conducting FMEA on product) and/or a detailed flowchart of the process. If these are not available, the team will need to create one.Using the blueprint and/or flowchart, team members should

12、 familiarize themselves with the product or process. They should physically see the product or a prototype of the product, and physically walk through the process exactly as the process flows.It is helpful to have an “expert” on the product or process available to answer any questions the team might

13、 have.Quality Forward SystemStep #2 Brainstorm Potential Failure ModesMethod of generating and documenting ideas of potential failure modes.If product or process is complex, a series of brainstorming sessions, each focused on a different element (i.e. people, methods, equipment, materials, environme

14、nt), may be needed to generate a more thorough list of potential failure modes.Facilitator should write down all ideas generated in brainstorming sessions.Once brainstorming is complete, ideas should be grouped into like categories (as determined by the team). Groups can be type of failure, where on

15、 the product failure occurred, estimated seriousness of the failure, etc. Failure modes may also be combined if they are the same or very similar.Once failure modes have been grouped/combined, they should be transferred to the Potential Failure Mode and Effect Analysis Sheet (example in SEQ 80.10, F

16、igure 2).Quality Forward SystemStep #2 (continued) Brainstorming Guidelines Do not make any judgment on ideas. Team members should not comment on whether an idea is good or bad. Thinking should be unconventional, imaginative, or even outrageous. Aim for a large number of ideas in the shortest possib

17、le time. Team members should “hitchhike” on other ideas, by expanding them, modifying them, or producing new ones by association. All team members should participate in brainstorming process. Quality Forward SystemStep #3 List Potential Effects of Each Failure Mode After the failure modes are listed

18、 on the FMEA worksheet form, the FMEA team needs to review each failure mode and identify the potential effects of the failure should it occur. For some failure modes, there may be only one effect, while other failure modes may have several effects. This step must be thorough this information will d

19、etermine assignment of risk ratings for each of the failures. It is helpful to think of this step as an if-then process: If the failure occurs, then what are the consequences.Quality Forward SystemStep #4 Assign Severity Rating The severity rating is an estimate of how serious the effects would be i

20、f a given failure did occur. May be based on past experience, or an estimate based on the knowledge and expertise of the team members. Because each failure may have several different effects, and each effect may have a different level of severity, the effect, not the failure, must be rated. Each eff

21、ect should be given its own severity rating. Ratings based on 10-point scale, with 1 being the lowest rating and 10 being the highest. General example on next page may be customized by team for their specific FMEA project.Quality Forward SystemStep #4 continued Severity Rating Scale ExampleRatingDes

22、criptionDefinition10Dangerously HighFailure could injure the customer or an employee9Extremely HighFailure would create noncompliance with federal regulations8Very HighFailure renders the unit inoperable or unfit for use7HighFailure causes a high degree of customer dissatisfaction6ModerateFailure re

23、sults in a subsystem or partial malfunction of the product5LowFailure creates enough of a performance loss to cause the customer to complain4Very LowFailure can be overcome with modifications to the customers process or product, but there is minor performance loss3MinorFailure would create a minor n

24、uisance to the customer, but the customer can overcome it in the process or product without performance loss.2Very MinorFailure would not be readily apparent to the customer, but would have minor effects on the customers process or product1NoneFailure would not be noticeable to the customer and woul

25、d not affect the customers process or product.Quality Forward SystemStep #5 Assign Occurrence RatingThe occurrence rating is a measure of how often an failure mode may occur.Best method for determine occurrence rating is to use actual data from the process (for example, failure logs or process capab

26、ility data).If actual data is not available, the team needs to estimate how often a failure mode could occur.Team can make a better estimate of how likely a failure mode is to occur and at what frequency by knowing the potential causes of failure.Ratings based on 10-point scale, with 1 being the low

27、est rating and 10 being the highest.General example on next page may be customized by team for their specific FMEA project.Quality Forward SystemStep #5 continued Occurrence Rating Scale ExampleRatingDescriptionPotential Failure Rate10Very High Failure is almost inevitableMore than one occurrence pe

28、r day or a probability of more than three occurrences in 10 events (Cpk 0.33).9One occurrence every three to four days or a probability of three occurrences in 10 events (Cpk approximately equal to 0.33).8High Repeated failuresOne occurrence per week or a probability of 5 occurrences in 100 events (

29、Cpk approximately equal to 0.67).7One occurrence every month or one occurrence in 100 events (Cpk approximately equal to 0.83)6Moderate Occasional failuresOne occurrence every three months or three occurrences in 1,000 events (Cpk approximately equal to 1.00)5One occurrence every six months to one y

30、ear or one occurrence in 10,000 events (Cpk approximately equal to 1.17)4One occurrence per year or six occurrences in 100,000 events (Cpk approximately equal to 1.33)3Low Relatively few failuresOne occurrence every one to three years or six occurrences in 10 million events (Cpk approximately equal

31、to 1.67)2One occurrence every three to five years or two occurrences in one billion events (Cpk approximately equal to 2.00)1Remote Failure is unlikelyOne occurrence in greater than five years or less than two occurrences in one billion events (Cpk greater than 2.00)Quality Forward SystemStep #6 Ass

32、ign Detection Rating The detection rating looks at how likely it is that a failure or the effect of a failure will be detected. Start by identifying current controls that may detect a failure or effect of a failure (example SPC or in-process checks). More controls in place generates a low number (be

33、tter) rating; no current controls means the likelihood of detection is low and therefore the item would receive a high number (worse) rating. Ratings based on 10-point scale, with 1 being the lowest rating and 10 being the highest. General example on next page may be customized by team for their spe

34、cific FMEA project.Quality Forward SystemStep #6 continued Detection Rating Scale ExampleRatingDescriptionDefinition10Absolute UncertaintyThe product is not inspected or the defect caused by failure is not detectable.9Very RemoteProduct is sampled, inspected and released based on Acceptable Quality

35、Level (AQL) sampling plans.8RemoteProduct is accepted based on new defectives in a sample.7Very LowProduct is 100% manually inspected in the process.6LowProduct is 100% manually inspected using go/no go or other mistake-proofing gages5ModerateSome Statistical Process Control (SPC) is used in process

36、 and product is final inspected off-line4Moderately HighSPC is used and there is immediate reaction to out-of-control conditions3HighAn effective SPC program is in place with process capabilities (Cpk) greater than 1.332Very HighAll product is 100% automatically inspected.1Almost CertainThe defect i

37、s obvious or there is 100% automatic inspection with regular calibration and preventive maintenance of the inspection equipment.Quality Forward SystemStep #7 Calculate Risk Priority NumberThe risk priority number (RPN) is calculated by multiplying the severity rating times the occurrence rating times the detection rating for all of the items. RPN = Severity X Occurrence X DetectionTotal RPN is calculated by adding all of the R