You are on page 1of 22

Expert priori knowledge for sizing of

maintenance teams with application in


natural gas stations

10th IMA International Conference on Modeling in


Industrial Maintenance and Reliability

MIMAR
HEITOR FAGNER LOPES DE LIMA

RODRIGO JOSÉ PIRES FERREIRA


FEDERAL UNIVERSITY OF PERNAMBUCO - BRAZIL

JUNE -14- 2018


MANCHESTER-UK
MOTIVATION

 Economic competitiveness scenario

 Cost of human resources in maintenance activities can


account about 58,9% of maintenance costs

 Lack of reliable data from maintenance records

 Need for practical and easy to application methodology


for sizing of maintenance teams
RESEARCH CHALLENGE

Sizing of maintenance teams


using expert priori knowledge
can be better than using
recorded data?

HOW? and WHY?


OBJECTIVES

 General Objective:
 To use a priori knowledge from maintenance experts as the basis for the
sizing of maintenance teams in natural gas custody transfer stations

 Specific Objectives:
 To identify reference criteria for the sizing of maintenance human
resources;

 To expose methodology for data modeling that uses the experience of


repairmen as a basis;

 To carry out a case study with maintenance staff in natural gas transfer
facilities
MAIN CONCEPTS
 Natural gas supply chain in Brazil

Called City Gates


or Delivey Points

State and
Processing Transport municipal
Exploration Transfer distribution
of
Custody

Delivery Points are very important in the natural gas supply chain!
MAIN CONCEPTS

 Definition of criteria Resources allocation in Maintenance


(%)
Corrective Preventive Predictive
Year Others
Maintenance Maintenance Maintenance
 ABRAMAN (2013) - Relationship 2013 30.86 36.55 18.82 13.77
2011 27.40 37.17 18.51 16.92
between Man Hour (MH) of 2009 26.69 40.41 17.81 15.09
2007 25.61 38.78 17.09 18.51
preventive maintenance over 2005 32.11 39.03 16.48 12.38
the total used: 50% 2003
2001
29.98
28.05
35.49
35.67
17.76
18.87
16.77
17.41
1999 27.85 35.84 17.17 19.14
1997 25.53 28.75 18.54 27.18
1995 32.80 35.00 18.64 13.56
 PALMER (2006) - Wrench time
Mh (Maintenance Services) / Mh (Total of work)
(no losses, displacements,
mobilizations, etc.): 50% Resources allocation in Maintenance - Brazilian Companies (ABRAMAN, 2013)

 Target: Attendance of 90% of


the maintenance on time
MAIN CONCEPTS

 Expert priori knowledge – Raiffa Method (1968)


Fractile k Judgmental fractile value ρk
.0625 36
.125 42
.25 50
.50 60
.75 68
.875 75
.9375 80
MAIN CONCEPTS

 Triangular Distribution

If there is a need to ask questions on various different proportions, making many


subdivisions would be very tiring for the specialists, thus we seek the possibility of asking
only three questions and forming a triangular distribution.
PROPOSED METHODOLOGY

HOW?
1. To list the needed maintenances (RCM is recommended);
2. To define the limits where start and finish the execution of each
maintenance task;
3. To identify the experts;
4. To explain to the experts about the methodology;
5. To make questions about the experience of experts and MH;
6. To identify the probability distribution Man-Hour;
7. To calculate the needed Man-Hour using the probability
distribution and the criteria defined.
APPLYING THE METHODOLOGY

 Activities in natural gas installations from RCM

Example Step 1: List the maintenance Example Step 5: Question about


tasks from RCM separated by Specialty experience of experts
and System
How many times have you
Sistem Specialy Activity performed each activity below on
Filtration Automation Maintenance nº1 City Gates? (if it’s never
Filtration
Filtration
Mechanics
Instrumentation
Maintenance nº2
Maintenance nº3
performed, use zero) *
Filtration Electrical Maintenance nº4
Heating Mechanics Maintenance nº5 * Answers:
Heating Automation Maintenance nº6
Heating Instrumentation Maintenance nº7
0 - If you never performed
Heating Electrical Maintenance nº8 1 - If you performed up to 5 times
P. Reduction Mechanics Maintenance nº9 2 – If you performed more than 5 and less than 10 times
P. Reduction Automation Maintenance nº10
P. Reduction Instrumentation Maintenance nº11
3 - If you performed more than 10 times
P. Reduction Electrical Maintenance nº12
Measurement Mechanics Maintenance nº13
Measurement Automation Maintenance nº14
Measurement Instrumentation Maintenance nº15
Measurement Electrical Maintenance nº16
APPLYING THE METHODOLOGY

 Questions using Raiffa Method to form a triangular


distribution

What is the maximum MH (slowest) used to


perform the job?

What is the minimum MH (faster) used to perform


the job?

What is the most likely MH to occur used to


perform the job?
RESULTS

 Man Hour limits and experience obtained


Repairman 1 Repairman 2 Repairman 3 Repairman 4 Repairman 5 Repairman 6 Repairman 7
Result Experience Experience Experience Experience Experience Experience Experience
Total experience 0,8 2,2 1,6 1,4 1,5 2,0 2,3

Instrumentation experience 0,7 2,9 2,1 2,0 2,2 1,9 2,1


Mechanics experience 0,3 2,4 2,3 2,1 1,8 2,0 2,3
Eletrical experience 2,4 1,8 0,6 0,0 0,0 1,6 2,8
Automation experience 0,7 0,7 0,0 0,0 0,2 2,7 2,7

For each maintenance task, was


Minumum Medium Maximum
Sistem Specialy Activity
MH MH MH

Filtration
Filtration
Automation
Mechanics
Maintenance nº1
Maintenance nº2
0,5
1,0
0,5
1,4
1,0
1,8
listed the points of triangular
Filtration
Filtration
Instrumentation
Electrical
Maintenance nº3
Maintenance nº4
0,7
0,7
0,8
0,9
1,3
1,4
distribution of Man-Hour needed
Heating Mechanics Maintenance nº5 0,9 1,3 2,6 individually and per month.
Heating Automation Maintenance nº6 1,2 2,1 3,6
Heating Instrumentation Maintenance nº7 0,5 0,8 1,2
Heating Electrical Maintenance nº8 0,9 1,4 2,5
P. Reduction Mechanics Maintenance nº9 2,3 3,4 5,0 The experience was used as a
P. Reduction Automation Maintenance nº10 2,7 4,0 6,2
P. Reduction Instrumentation Maintenance nº11 0,6 0,7 1,0 weight for the answers.
P. Reduction Electrical Maintenance nº12 1,9 2,7 4,0
Measurement Mechanics Maintenance nº13 0,5 0,5 1,0
Measurement Automation Maintenance nº14 1,2 2,1 3,6
Measurement Instrumentation Maintenance nº15 2,9 3,6 5,3
Measurement Electrical Maintenance nº16 1,5 2,1 3,3
RESULTS

 Man Hour limits obtained


MH per speciality INSTRUMENTATION AUTOMATION MECHANICS ELECTRICAL SUM
Minimum MH (per month) 20,8 41,0 23,2 8,2 93,2
MH for planned
maintenance
Medium MH (per month) 30,5 54,8 36,8 12,1 134,2
Maximum MH (per month) 47,7 105,4 54,2 17,7 225,1

Minimum MH (per month) 41,7 82,0 46,3 16,4 186,4


Total MH necessary Medium MH (per month) 60,9 109,7 73,6 24,2 268,4
Maximum MH (per month) 95,4 210,9 108,5 35,3 450,1

 Man Hour available per month


Mh available. = (m / 12) x d x h x WT = (11 / 12) x 21 x 8 x 0,5 = 77

Where:
• m – number of months available to work (12 taking one month to vacation)
• d – number of days available to work per month
• h – number of hours available to work per day
• WT – Wrench Time (productivity factor adopted)
RESULTS

 Repairmen limits obtained


MH per speciality INSTRUMENTATION AUTOMATION MECHANICS ELECTRICAL SUM
Minimum MH (per month) 20,8 41,0 23,2 8,2 93,2
MH for planned
maintenance
Medium MH (per month) 30,5 54,8 36,8 12,1 134,2
Maximum MH (per month) 47,7 105,4 54,2 17,7 225,1

Minimum MH (per month) 41,7 82,0 46,3 16,4 186,4


Total MH necessary Medium MH (per month) 60,9 109,7 73,6 24,2 268,4
Maximum MH (per month) 95,4 210,9 108,5 35,3 450,1

Repairmen limits Minimum MH (per month) 0,5 1,1 0,6 0,2 2,4
(77h available per Medium MH (per month) 0,8 1,4 1,0 0,3 3,5
month) Maximum MH (per month) 1,2 2,7 1,4 0,5 5,8

Thus, when the maintenance tasks lasts as long as possible,


will be necessary 6 repairmen.
RESULTS

 Man Hour and number of repairmen obtained


MH P.D. A.P.D.
HH minimum 186,40 0,0% 0,0%
HH medium 268,43 0,8% 31,1%
HH maximum 450,12 0,0% 100,0%
Hypothesis Test 380,92 0,3% 90%
Qty of repairmen (90%) 4,9

The unit needs 5 repairmen


to execute the maintenance

At the time of the study, the unit had 8 repairmen. The number of repairmen was
reduced from 8 to 5 approximately one year ago and the unit continued to meet
maintenance needs. The total maintenance cost was reduced approximately 22%
TEST USING REAL COLLECTED DATA

 Calculation based on collected MH

Quant. Months 12
y -20,13
β 3,38
R² 0,9241
η 383,53
Target 90,0%
Mh needed 490,55
Quant. Repairmen 6,4
RESULTS

A real case study indicated that the sizing of a


maintenance team using a priori knowledge
resulted in 5 repairmen and using recorded data
7 repairmen

WHY?
RESULTS

Companies use all available Man Hour.

“Too much human resources can hide the real


needs to perform maintenance tasks”

“The real data cannot reveal the real needs”

For example, three people can perform a specific job


instead of two needed, simply because three people are
available. The real recorded MH will represents three
people, because that is what was used.
RESULTS

“Expert Priori knowledge can


support this kind of decision ”
CONCLUSIONS

 Easy-to-apply and practical methodology for sizing


of maintenance teams

 Number of technicians suggested by the


methodology was confirmed in practice

 Analysis according to MH recorded is impracticable


REFERENCES

NASCIF, Julio e KARDEC, Alan. Manutenção função estratégica. Qualitymark editora. 3ª edição, 2009.

LOPES, Paulo Roberto Pinto. Uso de expansores para geração de energia elétrica em estações de redução
de pressão de gás natural. Rio de Janeiro. Originalmente apresentado como dissertação de mestrado, 2009.

FERREIRA, Heldemarcio Leite. Uso do Conhecimento a Priori de Especialista no dimensionamento de


Sobressalentes para a Manutenção do Sistema na CELPE. Recife. Originalmente apresentado como
dissertação de mestrado, 2001.

FERREIRA, Rodrigo José Pires. Modelos de decisão para dimensionamento de equipes de manutenção.
Recife. Originalmente apresentado como dissertação de mestrado, 2006.

NASCIF, Julio e DORIGO, Luiz Carlos. Manutenção orientada para resultados. Qualitymark editora. 1ª edição,
2010.

PALMER, Richard Doc. Maintenance Planning and Scheduling Handbook. Mcgraw-hill. 2ª edição, 2006.

ASSOCIAÇÃO BRASILEIRA DE MANUTENÇÃO E GESTÃO DE ATIVOS, ABRAMAN. Situação da Manutenção


no Brasil - Documento Nacional 2013. Disponível em: < http://www.abraman.org.br/sidebar/documento-
nacional/resultado-2013>. Acessado em 27/08/2017.

RAIFFA, Howard. Decision Analysis: Introductory Lectures on Choices Under Uncertainty. Addison-
Wesley. 2ª edição, 1968.
THANK YOU!

You might also like