REPORT ON MAINTENANCE PROGRAMME

RECOMMENDATIONS AND DISSEMINATION

By
All WP 1.6.3 partners
Abstract:
Scheduled Maintenance Programme shall be safe, cost-efficient and contribute to improve
the aircraft operational reliability.
The very competitive economic environment, the airlines are confronted with, leads the
industry to consider optimising their Maintenance Programmes (MP) to reduce maintenance
costs while maintaining safety and reliability.
Although some airlines perform this exercise on their own involving their local authorities, the
A/C manufacturer is in charge when sufficient in-service experience can be collected, to
organize a Maintenance Programme Evolution exercise. This exercise is directed by the
relevant Industry Steering Committee, as part of the MRB process. Results of the evolution
will be approved by the MRB and introduced in a revision of the MRBR/MPD.
Evolution exercises are based on in-service data reported by airlines in the form of “Nil
findings/details of findings” for each scheduled maintenance task issued from the MRB
process. Up to now, the evaluation of these data by the Maintenance Working Group
(manufacturer and operators as members and Regulatory Authorities as advisors) was
based on the operator experience, manufacturer expertise and “engineering judgment”.
In order to improve and homogenize this in-service data assessment, models have been
developed to help the decision-making on the maintenance task interval adjustment.
These models have been implemented, tested and this report describes the recommendation
and dissemination of theses models.

Dissemination:
PU

Deliverable/Output n°: D1.6.3.4 Issue n°: 1.0

Keywords:
in-service data, scheduled maintenance, unscheduled maintenance, maintenance program,
task interval, evolution exercise

VIVACE 1.6/3/EADS-CRC-F/T/05018-2.0 Page: 1/ 14

© 2005 VIVACE Consortium Members. All rights reserved.
VIVACE Model Specification for interval evolution

TABLE OF CONTENTS

1 GLOSSARY, ACRONYMS AND ABBREVIATIONS 4

1.1 GLOSSARY 4
1.2 ACRONYMS AND ABBREVIATIONS 4

2 EXECUTIVE SUMMARY 5

3 REMINDER OF MAINTENANCE PROGRAMME EVOLUTION 6

3.1 CONTEXT 6
3.1.1 MRB PROCESS 6
3.1.2 MAINTENANCE PROGRAM EVOLUTION PROCESS 7
3.2 VIVACE MP OBJECTIVE 7
3.3 MODELS DEVELOPED 8
3.3.1 MODEL 1X 8
3.3.1.1 principles 8
3.3.1.2 Statistical model 9
3.3.2 MODEL 2X 11
3.3.2.1 principles 11
3.3.2.2 Statistical model 11

4 RECOMMENDATIONS AND DISSEMINATION 13

APPENDIX A: REFERENCES 14

VIVACE 1.6/3/EADS-CRC-F/T/05018-2.0 Page: 3/ 14

© 2005 VIVACE Consortium Members. All rights reserved.
VIVACE Model Specification for interval evolution

1 GLOSSARY, ACRONYMS AND ABBREVIATIONS

1.1 GLOSSARY

Term Definition
Failure The inability of an item to perform within previously specified limits.
Failure Cause Why the functional failure occurs.
Function The normal characteristic actions of an item.
Functional Failure Failure of an item to perform its intended function within specified limits.
Hidden Function 1. A function which is normally active and whose cessation will not be evident
to the operating crew during performance of normal duties.
2. A function which is normally inactive and whose readiness to perform, prior
to it being needed, will not be evident to the operating crew during
performance of normal duties.
Interval (Initial - Initial Interval - Interval between the start of service-life and the first task
Repeat) accomplishment
Repeat Interval - The interval (after the initial interval) between successive
accomplishments of a specific maintenance task.
Scheduled Any of the maintenance opportunities which are pre-packaged and are
Maintenance Check accomplished on a regular basis.

1.2 ACRONYMS AND ABBREVIATIONS

Acronyms Designation
A/C Aircraft
EASA European Aviation Safety Agency
FAA Federal Aviation Administration
FC Failure Cause
FF Functional Failure
ISC Industry Steering Committee
MP Maintenance Program
MPD Maintenance Planning Document
MRB Maintenance Review Board
MRBR Maintenance Review Board Report
MSG-3 Maintenance Steering Group 3
MWG Maintenance Working Group
PPH Policy and Procedure Handbook
Serviceability of the function: the probability that the function is serviceable when
S or SF
needed during the period between two scheduled maintenance task.

VIVACE 1.6/3/EADS-CRC-F/T/05018-2.0 Page: 4/ 14

© 2005 VIVACE Consortium Members. All rights reserved.
VIVACE Model Specification for interval evolution

2 EXECUTIVE SUMMARY

Scheduled Maintenance Programme shall be safe, cost-efficient and contribute to improve the
aircraft operational reliability.
The very competitive economic environment, the airlines are confronted with, leads the industry
to consider optimising their Maintenance Programmes (MP) to reduce maintenance costs while
maintaining safety and reliability.
Although some airlines perform this exercise on their own involving their local authorities, the
A/C manufacturer is in charge when sufficient in-service experience can be collected, to
organize a Maintenance Programme Evolution exercise. This exercise is directed by the
relevant Industry Steering Committee, as part of the MRB process. Results of the evolution will
be approved by the MRB and introduced in a revision of the MRBR/MPD.
Evolution exercises are based on in-service data reported by airlines in the form of “Nil
findings/details of findings” for each scheduled maintenance task issued from the MRB process.
Up to now, the evaluation of these data by the Maintenance Working Group (manufacturer and
operators as members and Regulatory Authorities as advisors) was based on the operator
experience, manufacturer expertise and “engineering judgment”.
In order to improve and homogenize this in-service data assessment, models have been
developed to help the decision-making on the maintenance task interval adjustment.
These models have been implemented, tested and this report describes the recommendation
and dissemination of theses models.

VIVACE 1.6/3/EADS-CRC-F/T/05018-2.0 Page: 5/ 14

© 2005 VIVACE Consortium Members. All rights reserved.
VIVACE Model Specification for interval evolution

3 REMINDER OF MAINTENANCE PROGRAMME EVOLUTION

3.1 CONTEXT
3.1.1 MRB PROCESS

EASA and FAA require the A/C Type Certificate holder to prepare and revise the initial minimum
scheduled maintenance requirements that are applicable to a dedicated aircraft (Regulatory
Requirement CS/FAR 25.1529). This document is called the Maintenance Review Board Report
(MRBR), and provides the scheduled maintenance tasks and their frequencies (intervals) for the
aircraft systems (including powerplant), structure and zones. MRBR development is based on
the MSG-3 method.

Systems
MWG(s)
Maintenance Working Group

To provide PPH and To select/adjust

direct MWG activities systems tasks and
intervals

Validation
Zonal ISC MRB
MWG Industry Steering
Maintenance Review Board

Maintenance Committee
Working Group

Approval
PPH letter
Policy and
Procedures
handbook
MRB

Structure
MWG
Maintenance Working Group

Figure 1: Maintenance Review Board Process

VIVACE 1.6/3/EADS-CRC-F/T/05018-2.0 Page: 6/ 14

© 2005 VIVACE Consortium Members. All rights reserved.
VIVACE Model Specification for interval evolution

3.1.2 MAINTENANCE PROGRAM EVOLUTION PROCESS

The objective of maintenance program evolution is to maintain safety, reliability, at minimum

cost. The main means to achieve this is to adjust the initial/current selected intervals. Such
exercise is launched when the A/C manufacturer and the ISC consider that sufficient in-service
experience is collected, supported by the MRB as advisor.
Airlines report in-service experience in the form of “Nil findings/details of findings” reports for
each task. Then all tasks are reviewed individually in MWG meetings.

k
O MP Tas
MRBR
d
car
ISC MPD Job
Industry
Steering
Committee Maintenance Review Operator Maintenance
Maintenance
Board Report Program
Planning Document

Interval
adjustment
proposal Task
report
MWG
Manufacturer data
Maintenance compilation and analysis
Working Group

Figure 2 : Maintenance Program Evolution process

3.2 VIVACE MP OBJECTIVE

Up to now the evaluation of in-service data by the MWG during MP Evolution was based on the
operator experience, manufacturer expertise and “engineering judgment”.
This evaluation could be found sometimes conservative.

VIVACE MP objective was to help the decision-making by developing models and tools.

This would improve in-service data assessment and bring consistency all along the process
(several MWG meetings may be needed), between MWG attendees (different representative of
operators and manufacturers may attend different MWG meetings) and homogenise the
evolution results.

VIVACE 1.6/3/EADS-CRC-F/T/05018-2.0 Page: 7/ 14

© 2005 VIVACE Consortium Members. All rights reserved.
VIVACE Model Specification for interval evolution

3.3 MODELS DEVELOPED

Different models have been developed and prototyped:

- Model 1x based on data from scheduled maintenance,

- Model 2x based on data from scheduled and unscheduled maintenance.

3.3.1 MODEL 1X
3.3.1.1 PRINCIPLES

The model 1.X is based on the following principle

1) The sample of in-service data is first classed in three categories:

In service data considered statistically irrelevant;

Data are considered moderately representative of the whole fleet. In this
case, “Engineering judgment” is necessary to confirm the trend and make the
final decision;
Data are considered representative of the fleet and a decision can be made.

2) For data that are considered representative, interval adjustment is distributed in 3

categories
increase interval
decrease interval
keep interval as it is)

The above has lead to develop the following categories:

A in-service data supports target values of the new interval framework, possibly even a
higher interval can be justified
B In-service data supports target values of the new interval framework

C in-service data supports target values of the new interval framework, however other
inputs (e.g. engineering) should have a higher weight when making a final interval
decision
D in-service data does not support target values of the new interval framework, however
other inputs (e.g. engineering) should have a higher weight when making a final interval
decision
E in-service data does not support target values of the new interval framework, task
interval should possibly be decreased
F In-service data statistically irrelevant. However the MWG should not necessarily fully
disregard the reported data.
G current data does not support an escalation. Interval should remain as it is.

VIVACE 1.6/3/EADS-CRC-F/T/05018-2.0 Page: 8/ 14

© 2005 VIVACE Consortium Members. All rights reserved.
VIVACE Model Specification for interval evolution

3.3.1.2 STATISTICAL MODEL

Model 1x is a statistical model developed to estimate serviceability (S) of hidden function (e.g.
probability that the hidden function is serviceable when a trigger event occurs).

Model 1x principle is as follow:

Not OK
Trigger “Hidden” function function
event unserviceable
e.g. loss of system 1 or e.g. system 2 (or back-up)
need of back-up system OK is not able to take over

function
serviceable
e.g. system 2 (or back-up)
takes over

This model includes a statistical risk, which introduces an uncertainty in the computation of the
serviceability.

The defined categories (A, B, C, D, E, F, G) are linked to the model as follows:

o Low uncertainty and high serviceability (low percentage of findings): category A and B
o Low uncertainty and medium serviceability (medium percentage of findings): category G
o Low uncertainty and low serviceability (high percentage of findings): category E
o Medium uncertainty and medium serviceability (low percentage of findings): category C
o Medium uncertainty and low serviceability (high percentage of findings): category D
o High uncertainty: category F

The result is an evolution guidance ([ρ, NT] table) as follow:

Number of reported checks NT

A
B
% of findings

C
G
F D
ρ E

Placing the number of check and the percentage of finding on the above table, the MWG gets
the category (A, B, C…) defined earlier for the task under consideration.

Using this table, the MWG is helped to better evaluate the right adjustment of the interval,
according to the in-service experience.

VIVACE 1.6/3/EADS-CRC-F/T/05018-2.0 Page: 9/ 14

© 2005 VIVACE Consortium Members. All rights reserved.
VIVACE Model Specification for interval evolution

A prototype has been implemented within MS EXCEL:

For further details on model 1x, please refer to D1.6.3.3 issue 02 report.

VIVACE 1.6/3/EADS-CRC-F/T/05018-2.0 Page: 10/ 14

© 2005 VIVACE Consortium Members. All rights reserved.
VIVACE Model Specification for interval evolution

3.3.2 MODEL 2X

3.3.2.1 PRINCIPLES

Model 2x is an evolution of model 1x taking into account:

- The impact of unscheduled maintenance, i.e. “hidden” functional failure may become
evident to the flight crew: upon multiple failures or to the ground personnel: during a line
maintenance check,

1- Finding Task interval

0- Nil-finding

In In+1 In+2 In+3 time

Events reported from scheduled maintenance task

Events reported from Operational Interruptions
Events reported from Removals

- The modelling of MSG3 functional failure which is defined as a series of Failure Cause
because in MSG3 method, the failure of any cause implies the failure of the function,

Functional
Failure

Failure Cause 1 Failure Cause 2 Failure Cause 3

- The complex Failure Cause due to complex equipment, performing many functions and
then can lead to both hidden and evident functional failure,

Evident Evident
Functional Evident
part part
part
Failure Hidden Hidden
part
Hidden
part
part

Failure Cause 1 Failure Cause 2 Failure Cause 3

3.3.2.2 STATISTICAL MODEL

Model 2x estimates serviceability (S) of the function i.e. the probability that the function is
serviceable when needed. Model 2x principle is as follow:

Trigger
events?

Evident Evident
Functional Evident
Function Not
Failure
part
Hidden
part
Hidden
part
Hidden Function
part part part state?
Failure Cause 1 Failure Cause 2 Failure Cause 3 OK unserviceable

OK

Function
serviceable

VIVACE 1.6/3/EADS-CRC-F/T/05018-2.0 Page: 11/ 14

© 2005 VIVACE Consortium Members. All rights reserved.
VIVACE Model Specification for interval evolution

Model 2x results are computed for each failure cause 1, 2, 3 and then compiled for the
functional failure to determine the function state:

S F = ∏ S FC i
i

A prototype has been implemented within MS EXCEL:

For further details on model 2x, please refer to D1.6.3.3 issue 02 report.

VIVACE 1.6/3/EADS-CRC-F/T/05018-2.0 Page: 12/ 14

© 2005 VIVACE Consortium Members. All rights reserved.
VIVACE Model Specification for interval evolution

4 RECOMMENDATIONS AND DISSEMINATION

Up to now, the analysis of in-service data for the purpose of maintenance programme evolution
was based on the operator experience, manufacturer expertise and “engineering judgment”. In
order to improve and homogenize this in-service data assessment, models have been
implemented to help the decision-making on the adjustment of the maintenance task intervals.
Using in-service feedback experience and information from MSG-3 analysis, we were able to
build different mathematical models.
All models are developed within a certain environment. It means that before applying these
models, the user should be aware of the hypothesis, applicability and limitations (described in
deliverable D1.6.3.3 issue 02).

Model 2x is innovative, very close to MSG-3 methodology. It should be taken as functional

failure oriented model and both scheduled and unscheduled maintenance are considered.
Use-case has been carried out on model 2x and it remains more complex than model 1x, to
obtain the input data.

Model 1x should be taken as a hidden function oriented model and it deals with scheduled
maintenance data only. It:
- proposes to use several categories (A, B, C…) for interval adjustment
- indicates how to choose between these categories, using a [ρ, NT] table
- is easy-to-use and homogeneous for MWG activities.
- provides the shapes of categories zones on the [ρ, NT] table.

Once the Serviceability boundaries are defined according to the targeted percentage of
escalation, the MWG is supported, using the [ρ, NT] table, to better evaluate the right
adjustment of the interval, according to the in-service experience.

VIVACE 1.6/3/EADS-CRC-F/T/05018-2.0 Page: 13/ 14

© 2005 VIVACE Consortium Members. All rights reserved.
VIVACE Model Specification for interval evolution

APPENDIX A: REFERENCES

[1]: D1.6.3.1 – “Report on maintenance programme user requirements and state of the art
capture”

[2]: D1.6.3.2 – “Maintenance programme – in-service data analysis & model specification”

[3]: D1.6.3.3 issue 02 – “Maintenance programme: model specification for interval evolution”

[4]: Maintenance Programme Evolution (MPE) demonstrator presented in the VIVACE use
forum2.

VIVACE 1.6/3/EADS-CRC-F/T/05018-2.0 Page: 14/ 14

© 2005 VIVACE Consortium Members. All rights reserved.