Chapter 8: Human Error in Maintenance

1
Chapter 8 Human Error in Maintenance

• Overview
• Maintenance Error
• Techniques to Predict the Occurrence of Human
  Error in Maintenance

2
Maintenance Error

• Maintenance Error in System Life Cycle
• Breakdown percentage of a maintenance persons
  time

3
Maintenance Error

• Top human failure problems

4
Maintenance Error

• Reasons for Maintenance Error
• Poor work layout
• Poorly written maintenance procedures
• Complex maintenance tasks
• Improper work tools
• Poor Environment, etc.
• Guidelines for Reducing Human Error in
  Maintenance
• Areas procedures, human error risk management,
  tools and equipment, training, design,
  supervision, communication, shift handover,
  maintenance incident feedback, etc.

5
Maintenance Error Prediction Techniques

• Markov Model 1
• Problem the system can either fail due to
  maintenance error or other failures. Determine
  the probability of the occurrence of human error
  in maintenance
• Assumptions
• The failed system is repaired and PM is performed
  periodically
• Maintenance error and failure, other than human
  error, rates are constant
• Failed system repair rates are constant
• The repaired system is as good as new

6
Markov Model 1
System failed Due to other failures 2
System down due to maintenance error 1
µm
µn
System working Normally 0
?m
?n

• Model and solution

7
Markov Model 2

• Problem the system can only fail due to
  non-maintenance related failures, but its
  performance is degraded by the occurrence of
  maintenance error
• Assumptions
• The totally or partially failed system is
  repaired and PM is performed regularly
• The maintenance error can only lead to system
  degradation, but not failure
• The system can fail from its degradation mode due
  to failures other than maintenance errors
• The system repair rates from both failed and
  degradation states are constant
• Maintenance error and non-maintenance error
  failure rates are constant
• The repaired system is as good as new

8
Markov Model 2
µS
System working Normally 0
µd
µme
Failed
Degrade
?d
?me
?S

• Model and Solution

9
Examples

• Example 1 Assume that a system can fail due to
  PM-related human error and non-human error. The
  value of human and non-human error rates are
  0.0002 errors per hour and 0.005 failures/hour,
  respectively. Calculate the probability of the
  system failure due to maintenance error for a
  100-h mission.
• Example 2Assume that the following data are
  specified in model 2
• ?S 0.004 failures/h ?me0.0001 errors/h
  ?d0.001 failures/h µS0.02 repairs/h µme
  0.008 repairs/h µd 0.06 repairs/h. Calculate
  the steady state probability of the system
  degradation due to maintenance error

10
Fault Three Analysis Method

• Example Assume that an engineering system can
  fail due to a maintenance error caused by factors
  such as poor equipment design, inadequate
  training, poor work environment, use of deficient
  maintenance manuals, or inadequate work tools.
  Two major factors for poor equipment design are
  oversight or no formal consideration of the
  occurrence of maintenance error. The factor of
  maintenace error can be caused by either no
  requirement in design specifications or
  insufficient allocated funds. Two important
  factors for poor work environment are inadequate
  lighting and inaccessibility. Similarity, two
  main causes for the use of deficient maintenance
  manuals are unavailability of compatible
  maintenance manuals or porly written maintenance
  procedures. Develop a fault tree for the top
  event engineering system failed due to
  maintenance error. Assume the the probability of
  occurrence of events E1, E2, , is 0.02. For
  independent events, calculate the probability of
  the event of the engineering system failed due
  to maintenance error?

11
Fault Three Analysis Method