Improving the Quality of PFMEAs
David J. Benedict Senior Manager Advance Supplier Quality Planning Chassis/Electrical/Electronics DaimlerChrysler Corporation
Why am I here today?
1. Introduction
Todays Agenda:
2. Brief Background 3. Common Errors in PFMEAs 4. Supplier PFMEA Audit 5. Questions
Current Paradigm
New Paradigm
Drives
Maintenance Plans
FMEA
Error/Mistake Proofing
FMEA Process Control Plans
Taken lightly
Selective products/processes chosen Based on opinion All magically below the RPN cut-off
Taken seriously
All critical and new products/processes Data driven Roll-up RPN values (Green Y philosophy)
Too late
Error/Mistake proofing window long gone Rarely validated with field results
Early
Completed while design is still fluid Living document
Paper exercise done when pushed
Ensures our products and processes are robust
1. Not started because production equipment not operational.
What are some common errors?
2. Looking only at RPNs when prioritizing Recommended Actions. 3. Not following minimum guidelines for Severity rankings. 4. RPN below target level and no further action needed. 5. Filled out using ONLY references to other documents.
What are some common errors?
6. Effects of Failure not Customer effects. 7. Not distinguishing between types of controls. 8. Inconsistency in Severity or Detection rankings. 9. Detection ranking too low. 10.Insufficient Controls for Severity rankings of 8, 9 or 10.
PFMEA not started because production equipment not operational
Minimum documents required to start a PFMEA: Process Flow Chart Risk Assessment Work Instructions The PFMEA should be updated later with findings from trial/pilot runs.
Looking only at the RPNs when prioritizing Recommended Actions
Process Function Potential Failure Mode Requirements Potential Effect(s) of Failure S e v
OP:20
Fracture of wheel-does not meet FMVSS requirements Gas Porosity too
great (greater than x criteria
C l a s s
O Potential c Cause(s) / c Mechanism(s) of u Failure r Hydrogen Supersaturation due toexcessive weight melt temperature.
Current Process Controls Prevention
Current Process Controls Detection
D e t e c t
R P N
OP:170
Unacceptable Surface Appearance contaminants not completely noticed by most removed customers
10
4
Automated Temperature Controls in melt 2 furnace
Strip charting and SPC 6
120
Contamination in Periodic exchange of DI 100% visual inspection Rinse Water 5 water-once per shift at Op. 180
160
For Severity, OP. 20 is much more critical than OP. 170. The RPN number suggests OP. 170 is a higher priority. What would be the results of a quality spill for OP. 20 vs. OP. 170?
Not following minimum guidelines for Severity rankings
Misidentifying Potential Effects of Failure can make establishing meaningful Severity rankings difficult. Legal concerns sometimes drive lower rankings. Note: If the severity received a ranking of 9 or 10, it is important to include a Customer- or Supplier-specific safety symbol to highlight its significance.
RPN below target level and no further action needed
D e R t P e N c t Action Results Recommended Action(s) Responsibility & Target Completion Date Action Results S O D R e c e P v c t N
PFMEA is a continuous improvement tool and a living document. Reducing occurrence should continue to be a goal long after erroror mistake-proofing has been implemented.
24
N/R
36
N/R
42
N/R
Filled out using ONLY references to other documents
O Potential Cause(s) / c Mechanism(s) of c Failure u r Current Process Controls Prevention Current Process Controls Detection
D e t e c t
This is great for TS/QS documentation purposes. Modification not required when document changes.
Advantage
Too Small Shot Blast Media Used
W.I. 483-11-04 Form F-19.78
Tool Wear
W.I. 483-11-04 Form F-19.78
Impossible to assess appropriateness of Detection rankings without referenced documents. Encourages check-the-box mentality.
Disadvantages
Process Function
Effects of Failure not Customer effects
S Potential Failure Potential e Mode Effect(s) of Failure v Requirements
Consider what would happen if there were NO CONTROLS! Scrap/downtime are true, but are not what the Customer would notice or experience. Think of what the customer would notice/experience if the Bolt was loose.
OP. 210 Valve Hole Deburr/ Burr Free
OP. 220 Punch Hole/Hole Diameter 10.0 mm +/-0.2
Burrs
Excessive scrap/ 7 downtime
Hole Diameter >10.2 mm.
Bolt loose
Not distinguishing between types of controls
Prevention affects ONLY Occurrence rankings. Detection affects ONLY Detection rankings. Not knowing the difference often leads to incorrect Detection rankings. When using 2nd edition FMEA Manual form, Controls must be labeled (P) for Prevention and (D) for Detection. 3rd edition form has two separate columns for Prevention and Detection Controls.
Inconsistency in Severity or Detection rankings
Current Process Controls Prevention Standardized work instructions Standardized work instructions Current Process Controls Detection
D e t e c t i o n
Final Functional Testing is called out as both a 3 and 2. Why? Clarify/be more specific if Control is truly different.
Final Functional Testing Machine 3 Final Functional Testing Machine 2
Detection ranking too low
Process Function S Potential Failure Potential e Mode Effect(s) of Failure v Requirements C l a s s O Potential Cause(s) / c Mechanism(s) of c u Failure r Current Process Controls Prevention Current Process Controls Detection D e t e c t
OP. 210 Valve Flat tire-Value Hole deburr/all Not all burrs Stem cut during burrs removed removed. tire mounting.
Operator did not remove all burrsinsufficient 4 lighting.
Standardized Work Instructions
100%Visual Inspection at Op. 220.
In this example, the Detection Controls are 100% visual inspection. The Detection ranking should be an 8, not 3.
Insufficient Controls for Severity rankings of 8, 9 or 10
Process Function Potential Failure Mode Requirements Potential Effect(s) of Failure S e v C O l Potential Cause(s) / c a Mechanism(s) of c s u Failure s r Hydrogen Supersaturation due toexcessive weight melt temperature. Current Process Controls Prevention Current Process Controls Detection D e t e c t R P N
Fracture of wheel-does OP:20 Aluminum Gas Porosity too meet FMVSS 10 not Melt/Gas Porosity not great (greater than x to exceed x criteria requirements criteria
Automated Temperature Controls in melt 2 furnace
Strip charting 6 and SPC
120
A non-conformance for OP. 20 can result in a Field Campaign/Recall. While SPC is good at detecting trends, it cannot detect with much certainty a single, random non-conformance. Error/Mistake-proofing must be implemented (Detection ranking of 3 or less) for Severity rankings of 9 or 10. Limit Detection ranking to no greater than 4 for Severity ranking of 8.
Supplier PFMEA Audit
What is the Audit?
An objective tool to evaluate the quality of the PFMEA
Why was the Audit Form Created?
Previous process involved individuals applying their own criteria for determining what made a good vs. bad PFMEA (often inconsistent and confusing). PFMEA best practices were not being shared. Needed to find a way to accelerate PFMEA improvement.
Who will use the Audit?
Suppliers (self-assessment) Customers (evaluate Supplier PFMEA submissions)
What can the audit do?
The form is really 2-in-1 and has the flexibility to be used:
by those creating or developing the PFMEA (supplier manufacturing engineer, design engineer, quality engineer, etc.), or by those conducting a 3rd party audit of the PFMEA (supplier or customer corporate quality function, etc.)
What are some of the things the audit form can help detect?
It can help detect whether:
a cross-functional team provided input during the creation of the PFMEA the probability of occurrence was created using real world data previous internal or external rejects were included the DFMEA was referenced/used in the creation of the PFMEA
How do you use the Audit form?
The use of the form is straight forward:
If you are creating or developing (or wish to do a "deep dive" into the data behind the PFMEA), answer ALL questions on the form. If you are conducting a 3rd party audit of a PFMEA, answer only those questions not shaded/italicized.
Where can I find the Audit form?
Go to Covisint and log in. Go to DaimlerChrysler page. Go to bulletin 76738.
Questions?
