Reliability Calculations

1
Reliability Calculations

• What, Why, When
• How do we benefit from them?

2
Who am I?

• Harvey Altstadter
• 34 years experience in Component Engineering and
  Reliability
• Military
• Commercial
• Commercial Space
• HR Electronic Components Consulting Services Inc
• Consultant to Industry
• 631 928-2847

3
What are Reliability Calculations?

• Methodology for analyzing the expected or actual
  reliability of a product, process or service, and
  identifying actions to reduce failures or
  mitigate their effect.
• Stress Analysis
• Reliability Predictions
• FMEA (Failure Mode and Effects Analysis) or
• FMECA (Failure Mode Effects and Criticality
  Analysis)
• Yardstick for comparison of design approaches
• Cost-Benefit Trade

4
Why Do Reliability Calculations?

• Make the product more reliable
• Selling feature
• Reduce returns
• Lower costs
• Enhance or maintain company reputation
• Comparisons with competition
• Customer request
• Design goal
• Hard Requirement

5
Stress Analysis

• Establishes the presence of a safety margin
• Good engineering practice
• Enhances system life
• Provides input data for Reliability Prediction
• Describes operating condition as a percentage of
  rating
• Customer requirement
• Validates compliance with Derating Criteria

6
Reliability Predictions (MTBF)
THE MYTH
If you dont like the numbers... ...give me five
minutes, I will make up a better one
7
Reliability Predictions (MTBF)

• Form the basis of Reliability Analyses
• Compute predicted system failure rate or
  Mean Time Between Failures
• Failure Rate is usually expressed in Failures per
  106 or 109 hours
• MTBF is usually expressed in terms of hours
• Example for a system with a predicted MTBF of
  1000 hours, on average the system experiences one
  failure in 1000 hours of operation or a Failure
  Rate of 1000 per 106 hours
• Methodology
• Use accepted standards
• Model failure rates of components
• Analyze system
• Calculate the system predicted failure rate or
  MTBF
• Evaluate prediction vs target or required MTBF
• Evaluate stress or temperature reduction design
  changes
• Evaluate practicality of design change especially
  when MTBF is self imposed

8
Reliability Predictions (Continued)

• Common Standards
• MIL-HDBK-217
• Generally associated with military systems
• Models are very detailed
• Provides for many environments
• Provides multiple quality levels
• Bellcore (Telcordia)
• Telecommunications Industry standard
• Seems to have supplanted French CNET and British
  Telcom standards
• Models patterned after MIL-HDBK-217, but
  simplified
• Provides multiple quality levels
• Can incorporate current laboratory test data
• Can incorporate current field performance data
• Other Standards
• Auto Industry
• Resources
• Software packages cover both MIL-HDBK-217 and
  Bellcore models
• RELEX is widely available

9
Reliability Predictions (Continued)

• General approach to Prediction
• Model for a single part consists of a number of
  factors multiplied together
• ?ss ?G ?Q ?S ?T
• ?ss Steady State Failure rate
• ?G Generic or Base Failure Rate
• ?Q Quality Factor
• ?S Stress Factor
• ?T Temperature Factor
• Other factors First Year Multiplier or
  Experience Factor
• Model for a unit
• Consists of the sum of all of the individual part
  failure rates multiplied by an Environmental
  Factor ?E
• Source of Factor Information- varies with method
  used
• Lookup table
• Calculation based upon complexity

10
Reliability Predictions
The first cut is made with little analysis to get
a rough idea where the design is relative to the
desired outcome Better numbers come from better
insight into the design Factors to be considered
include Duty Cycle and refined Thermal and Stress
Analyses
THE TRUTH
11
FMEA or FMECA

• Design FMEA
• FMEA is a bottoms up method of analyzing and
  improving a design
• Heavily used by US automotive industry
• Chrysler, Ford, GM require this type of analysis
• Many different company and industry standards
• Most widely used is the AIAG (Automotive Industry
  Action Group) standard
• Analytic Process
• Consider each component or functional block and
  how it can fail (Failure Modes)
• Determine the Effect of each failure mode, and
  the severity on system function
• Determine the likelihood of occurrence and of
  detecting the failure.
• Calculate the Risk Priority Number, or RPN,
  using the formula as follows
  
  RPN Severity x Occurrence x Detection
• Consider corrective actions (may reduce severity
  of occurrence, or increase probably of detection)
  
• Start with the higher RPN values (most severe
  problems) and work down
• Recalculate RPN after the corrective actions have
  been determined, the aim is to minimize RPN

12
FMEA or FMECA (Continued)

• Process FMEA
• Similar to a Design FMEA but is applied to a
  manufacturing process or service. The object is
  to use this methodology to optimize processes.
• FMECA
• A FMECA is similar to a FMEA,
• Criticality is computed in place of RPN .
• FMECAs are used extensively in military,
  aerospace and medical equipment fields, for both
  design and process reliability analysis.
• MIL-HDBK-1629 is a widely accepted standard for
  FMECAs.

13
When

• Stress Analysis
• Prior to release of design to production
• Prior to implementation of design changes
• Reliability predictions should be done at all
  stages of design
• Early design stage- Reliability Prediction may a
  rough estimate
• Late design stage- Reliability Prediction is
  refined
• Fielded system- revised prediction can
  incorporate field data for future use
• Design FMEA or FMECA
• As design matures, impact of failure needs to be
  addressed
• Process FMEA
• During process design
• Prior to implementing new or updated processes

14
How do we benefit from them?

• No system benefits from a calculation
• Calculation without action is window dressing
• Contributes to good or bad feelings about system
• Could make customer happy ...or not
• Calculation after design is complete is a waste
  of time
• Feedback of results into design yields the
  benefits
• Longer predicted life
• Fewer field failures
• Lower warranty costs
• Better customer relations

15
Famous Flubs

• BART (Bay Area Rapid Transit)
• FMEA not performed or inadequate
• Oscillator Crystal Failure- Open. Oscillator
  Frequency went up
• Train speed increased rapidly
• Train overshot last stop
• Train rammed barrier at high speed
• Very serious accident
• Browns Ferry Nuclear Power Plant Accident
• Results of FMEA waived
• Main Redundant Instrumentation Control Wiring
• Required to be in separate cable trays in case of
  fire
• Requirement waived as a cost saving
• Fire in containment wall insulation during leak
  check
• Fire destroyed main and redundant I C wiring
• Nuclear Plant on fire and out of control for
  several days

16
Reliability Calculations

• Summary
• What Analysis Toolkit
• Why Product Improvement Reduced Cost
• When Early for Design Feedback Prior
  to Completion to Validate Goals
• Benefit Reduced Field Failures
  Reduced Warranty Costs Better Customer
  Relations