Introduction to Reliability Analysis

AE ESP NEW – ESP Training

Instructor: Irene Yu

Schlumberger-Private
Objectives

■ After the session you will be able to:

̶ Calculate ESP performance using different metrics.
̶ Plot simple reliability curves.
̶ Analyze ESP performance based on reliability plots.
̶ Compare ESP performance for different ESP population.
̶ Discuss with client’s reliability engineer and identify client’s requirements and expectations regarding ESP
reliability.

Schlumberger-Private
Introduction to Reliability
Reliability Definition

■ The probability of a product performing without failure

̶ a specified function
̶ under given conditions
̶ for a given period of time.

Quality Control Handbook, Third Edition; McGraw Hill

■ Reliability x Quality
̶ Reliability can be thought as a measure of quality over some time period.

Schlumberger-Private
Introduction to Reliability
Consequences of Low Reliability

■ What would 99% reliability mean?

̶ No electricity for 7 hours each month.
̶ 2 unsafe landings in every airport daily.
̶ 5,000 incorrect surgery operations per week.
̶ 200,000 wrong drug prescription each year.

■ Reliability in Oil and Gas industry

̶ Lost of well control causing accidents and major spills.
̶ Deferred oil production.
̶ Lost in revenue.
̶ Affects reputation.

Schlumberger-Private
Introduction to Reliability
Schlumberger ALS Goals from 2014

■ By year 2020, improving the operational reliability of the Artificial Lift Segment (ALS) by a factor
10.

Schlumberger-Private
Introduction to Reliability
Artificial Lift Quality Indicators

MTTP: Mean Time To Pull ESP Survivability @ 365 Days

Improvement measured against K-M Survival plot at 365 Days
𝑴𝑻𝑻𝑷 =
𝑇𝑜𝑡𝑎𝑙 𝐼𝑛𝑠𝑡𝑎𝑙𝑙𝑒𝑑 𝐷𝑎𝑦𝑠 Justification
𝑇𝑜𝑡𝑎𝑙 𝑃𝑢𝑙𝑙𝑒𝑑 𝐸𝑆𝑃′ 𝑠 ▪ 365 Days Default Warranty period
▪ Current performance is less than expected at 60%
▪ Simplified, estimate equation shown. ▪ Warranty (TRL) failures have measureable financial impact
▪ Actual KPO will use Kaplan Meirer (K-M) method .
▪ MTTX is a Industry standard.
▪ X = Pull or Failure (depending on available data).
▪ Calculation exists in AL Business System.

Schlumberger-Private
Introduction to Reliability
Artificial Lift Quality Indicators – H1 2019

Note:
- 10x reliability target for 2020 is EFr of 5.22
- 2019 KPO: 5.55

Schlumberger-Private
Introduction to Reliability
Artificial Lift Quality Indicators

Early By focusing on disciplined execution, SLB will achieve a 10X improvement in installs per early failure
Failure thereby significantly reducing client total cost of ownership.

Run Life By strengthening the specific focus of the domain on run life improvement, SLB will deliver superior
run life that will maximize production for our customers thereby lowering TCO.

By focusing on Production Lifting Services and Optimization, SLB will leverage the subsurface
Unplanned
Work-Over reservoir knowledge and the power of data analytics to help reduce unplanned work-overs for our
clients.

Schlumberger-Private
Introduction to Reliability
ESP Reliability

■ How is the ESP population performing?

■ Is the performance improving with time?

■ What is a realistic ESP run life expectation?

■ Where should efforts focus on improving performance?

■ Have recent changes actually increased reliability?

Schlumberger-Private
Reliability Metrics
ESP Reliability
■ A meaningful and correct metric is essential.

■ Because we can only improve what we measure.

Schlumberger-Private
Reliability Metrics
Measuring Reliability
System A System B System C
Run Life [Days] Status Run Life [Days] Status Run Life [Days] Status
ESP 1 21 R ESP 1 28 R ESP 1 653 R
ESP 2 649 R ESP 2 15 R ESP 2 781 R
ESP 3 325 F ESP 3 591 R ESP 3 32 F
ESP 4 18 R ESP 4 467 F ESP 4 18 R
ESP 5 45 F ESP 5 3 R ESP 5 45 F
ESP 6 192 F ESP 6 513 F ESP 6 24 F
ESP 7 123 R ESP 7 345 F ESP 7 487 R
ESP 8 549 R ESP 8 549 R
ESP 9 318 R ESP 9 382 F
ESP 10 98 R

Schlumberger-Private
Reliability Metrics
Average Run Life

■ Advantages
̶ Familiar concept 𝑛
1
̶ Easy to calculate 𝐴𝑅𝐿 = ෍ 𝑅𝐿𝑖
𝑛
𝑖=1

■ Disadvantages ARL = Average run life

̶ Requires a high number of events RLi = Run life of the ith element
n = total number of elements in the population
̶ Difficult to compare two populations
̶ Uncertainty over which population to use

Schlumberger-Private
Reliability Metrics
Average Run Life
System A System B System C
Run Life [Days] Status Run Life [Days] Status Run Life [Days] Status
ESP 1 21 R ESP 1 28 R ESP 1 653 R
ESP 2 649 R ESP 2 15 R ESP 2 781 R
ESP 3 325 F ESP 3 591 R ESP 3 32 F
ESP 4 18 R ESP 4 467 F ESP 4 18 R
ESP 5 45 F ESP 5 3 R ESP 5 45 F
ESP 6 192 F ESP 6 513 F ESP 6 24 F
ESP 7 123 R ESP 7 345 F ESP 7 487 R
ESP 8 549 R ESP 8 549 R
ESP 9 318 R ESP 9 382 F
ESP 10 98 R

21 + 649 + 325 + 18 + 45 + 192 + 123 + 549 + 318 + 98 28 + 15 + 591 + 467 + 3 + 513 + 345 653 + 781 + 32 + 18 + 45 + 24 + 487 + 549 + 382
𝐴𝑅𝐿 = 𝐴𝑅𝐿 = 𝐴𝑅𝐿 =
10 7 9
𝑨𝑹𝑳 = 𝟐𝟑𝟑. 𝟖 𝑫𝒂𝒚𝒔 𝑨𝑹𝑳 = 𝟐𝟖𝟎. 𝟑 𝑫𝒂𝒚𝒔 𝑨𝑹𝑳 = 𝟑𝟑𝟎. 𝟏 𝑫𝒂𝒚𝒔

Schlumberger-Private
Reliability Metrics
Average Run Life – Failed Population
System A System B System C
Run Life [Days] Status Run Life [Days] Status Run Life [Days] Status
ESP 1 21 R ESP 1 28 R ESP 1 653 R
ESP 2 649 R ESP 2 15 R ESP 2 781 R
ESP 3 325 F ESP 3 591 R ESP 3 32 F
ESP 4 18 R ESP 4 467 F ESP 4 18 R
ESP 5 45 F ESP 5 3 R ESP 5 45 F
ESP 6 192 F ESP 6 513 F ESP 6 24 F
ESP 7 123 R ESP 7 345 F ESP 7 487 R
ESP 8 549 R ESP 8 549 R
ESP 9 318 R ESP 9 382 F
ESP 10 98 R

325 + 45 + 192 467 + 513 + 345 32 + 45 + 24 + 382

𝐴𝑅𝐿 = 𝐴𝑅𝐿 = 𝐴𝑅𝐿 =
3 3 4
𝑨𝑹𝑳 = 𝟏𝟖𝟕. 𝟑 𝑫𝒂𝒚𝒔 𝑨𝑹𝑳 = 𝟒𝟒𝟏. 𝟕 𝑫𝒂𝒚𝒔 𝑨𝑹𝑳 = 𝟏𝟐𝟎. 𝟖 𝑫𝒂𝒚𝒔

Schlumberger-Private
Reliability Metrics
Mean Time to Event (MTTX)
𝑚
■ Event shall be defined clearly. 1
𝑀𝑇𝑇𝐹 = ෍ 𝑅𝐿𝑖
■ Includes individuals without an event in the analysis. 𝑛
𝑖=1
■ Allows indication of population performance with MTTF = Mean Time to Failure
small number of events. RLi = Run life of the ith element
n = total number of failures
■ Most common events in ESP analysis are pull m = total population count
(MTTP) and failure (MTTF).

■ Mean Time to Failure (MTTF) against Mean Time Between Failures (MTBF)

MTBF
MTTF

Schlumberger-Private
Reliability Metrics
Mean Time to Failure
System A System B System C
Run Life [Days] Status Run Life [Days] Status Run Life [Days] Status
ESP 1 21 R ESP 1 28 R ESP 1 653 R
ESP 2 649 R ESP 2 15 R ESP 2 781 R
ESP 3 325 F ESP 3 591 R ESP 3 32 F
ESP 4 18 R ESP 4 467 F ESP 4 18 R
ESP 5 45 F ESP 5 3 R ESP 5 45 F
ESP 6 192 F ESP 6 513 F ESP 6 24 F
ESP 7 123 R ESP 7 345 F ESP 7 487 R
ESP 8 549 R ESP 8 549 R
ESP 9 318 R ESP 9 382 F
ESP 10 98 R

21 + 649 + 325 + 18 + 45 + 192 + 123 + 549 + 318 + 98 28 + 15 + 591 + 467 + 3 + 513 + 345 653 + 781 + 32 + 18 + 45 + 24 + 487 + 549 + 382
𝐴𝑅𝐿 = 𝐴𝑅𝐿 = 𝐴𝑅𝐿 =
3 3 4
𝑴𝑻𝑻𝑭 = 𝟕𝟕𝟗. 𝟑 𝑫𝒂𝒚𝒔 𝑴𝑻𝑻𝑭 = 𝟔𝟓𝟒. 𝟎 𝑫𝒂𝒚𝒔 𝑴𝑻𝑻𝑭 = 𝟕𝟒𝟑. 𝟐 𝑫𝒂𝒚𝒔

Schlumberger-Private
Survival Analysis
Probability, Cumulative, Reliability and Hazard Functions
Histogram Probability Density Function (PDF) f(t) Probability Density Function (PDF) f(t)
30 100% 30 25%
90%
25 25

Cumulative Percentage
80% 20%

20 70% 20

Percentage
Frequency
Frequency

60% 15%
15 50% 15
40% 10%
10 10
30%
5 20% 5 5%
10%
0 0% 0 0%
0 30 60 90 120 150 180 210 240 270 300 330 360 390 420 0 30 60 90 120 150 180 210 240 270 300 330 360 390 420 0 30 60 90 120 150 180 210 240 270 300 330 360 390 420
Days to failure Days to Failure Days to Failure

Cumulative Distribution Function (CDF) F(t) Reliability Function R(t)

100% 100%
90% 90%
𝑅 𝑡 = 1 − 𝐹(𝑡)
Cumulative Probability

80% 80%
70% 70%

Survivability
60% 60%
50% 50%
40% 40%
30% Hazard Function = Instantaneous Failure Rate 30%
20% 20%
10% 𝑓(𝑡) 10%
0% ℎ 𝑡 = 0%
0 30 60 90 120 150 180 210 240 270 300 330 360 390 420 𝑅(𝑡) 0 30 60 90 120 150 180 210 240 270 300 330 360 390 420
Days to failure Days to Failure

Schlumberger-Private
Survival Analysis
Probability, Cumulative, Reliability and Hazard Functions
Histogram
30 100%
90%
25
80%

Cumulative Percentage
70%
20
60%
Frequency

15 50%
40%
10
30%
20%
5
10%
0 0%
0 30 60 90 120 150 180 210 240 270 300 330 360 390 420
Days to failure

Schlumberger-Private
Survival Analysis
Probability, Cumulative, Reliability and Hazard Functions
Probability Density Function (PDF) f(t)
30

25

20
Frequency

15

10

0
0 30 60 90 120 150 180 210 240 270 300 330 360 390 420
Days to Failure

Schlumberger-Private
Survival Analysis
Probability, Cumulative, Reliability and Hazard Functions
Probability Density Function (PDF) f(t)

25%

20%
Percentage

15%

10%

5%

0%
0 30 60 90 120 150 180 210 240 270 300 330 360 390 420
Days to Failure

Schlumberger-Private
Survival Analysis
Probability, Cumulative, Reliability and Hazard Functions
Cumulative Distribution Function (CDF) F(t)
100%
90%
80%
Cumulative Probability

70%
60%
50%
40%
30%
20%
10%
0%
0 30 60 90 120 150 180 210 240 270 300 330 360 390 420
Days to failure

Schlumberger-Private
Survival Analysis
Probability, Cumulative, Reliability and Hazard Functions
Reliability Function R(t)
100%
90%
80%
70%
Survivability

60%
50%
40%
30%
20%
10%
0%
0 30 60 90 120 150 180 210 240 270 300 330 360 390 420
Days to Failure

Schlumberger-Private
Survival Analysis
Probability, Cumulative, Reliability and Hazard Functions
Histogram Probability Density Function (PDF) f(t) Probability Density Function (PDF) f(t)
30 100% 30 25%
90%
25 25

Cumulative Percentage
80% 20%

20 70% 20

Percentage
Frequency
Frequency

60% 15%
15 50% 15
40% 10%
10 10
30%
5 20% 5 5%
10%
0 0% 0 0%
0 30 60 90 120 150 180 210 240 270 300 330 360 390 420 0 30 60 90 120 150 180 210 240 270 300 330 360 390 420 0 30 60 90 120 150 180 210 240 270 300 330 360 390 420
Days to failure Days to Failure Days to Failure

Cumulative Distribution Function (CDF) F(t) Reliability Function R(t)

100% 100%
90% 90%
𝑅 𝑡 = 1 − 𝐹(𝑡)
Cumulative Probability

80% 80%
70% 70%

Survivability
60% 60%
50% 50%
40% 40%
30% Hazard Function = Instantaneous Failure Rate 30%
20% 20%
10% 𝑓(𝑡) 10%
0% ℎ 𝑡 = 0%
0 30 60 90 120 150 180 210 240 270 300 330 360 390 420 𝑅(𝑡) 0 30 60 90 120 150 180 210 240 270 300 330 360 390 420
Days to failure Days to Failure

Schlumberger-Private
Reliability Plots
Parametric Distributions

■ An equation is derived to best fit the actual data.

■ This is the concept of parametric plotting method or “fitting a parametric distribution”.
■ Examples of parametric distributions:
Reliability Function
̶ Normal 100%
̶ Lognormal 80%
̶ Exponential

Survivability
y = 9E-06x2 - 0.0059x + 1.0147
60%
̶ Gamma 40%
̶ Weibull 20%
̶ Etc… 0%
0 100 200 300 400 500
Days

Schlumberger-Private
Parametric Method
Exponential Distribution
CDF PDF Reliability Function Hazard Function
λ=1.5 λ=1.5
λ=1.5
λ=1.0
λ=0.5
λ=1.0

R(t)

h(t)
F(t)

f(t)
λ=1.0 λ=0.5
λ=0.5
λ=1.0
λ=0.5 λ=1.5
t t t t

𝑓 𝑡 = 𝜆𝑒 −𝜆𝑡 R 𝑡 = 𝑒 −𝜆𝑡 𝑓(𝑡) 𝜆𝑒 −𝜆𝑡

F 𝑡 =1− 𝑒 −𝜆𝑡 ℎ 𝑡 = = −𝜆𝑡 = 𝜆
𝑅(𝑡) 𝑒

■ λ is the failure rate of the population. This is the chance of an event in a certain time frame.
1
𝜆=
𝑀𝑇𝑇𝐹

Schlumberger-Private
Parametric Method
Exponential Distribution
System A System B System C
MTTF [Days] 779.3 654.0 742.8
Failure Rate (λ) 0.00128 0.00153 0.00135

Exponential Distribution
1
0.9
0.8
0.7
Survivability

0.6
0.5
0.4
0.3
0.2
0.1
0
0
25
50
75

550
100
125
150
175
200
225
250
275
300
325
350
375
400
425
450
475
500
525

575
600
625
650
675
700
725
750
775
Days

System A System B System C

Schlumberger-Private
Parametric Method
MTTF and Exponential Distribution

■ Easy to calculate.

■ Population is assumed to perform under the exponential distribution.

■ Be careful! Calculated value can be misleading.

̶ MTTF and Average run life are two completely different concepts.

■ Constant failure rate in ESP operation can be questionable.

■ The model gets more accurate as the number of events increase.

Schlumberger-Private
Reliability Plots
Non-Parametric Method

■ No specific distribution is considered.

Reliability Function
100%
90%
■ The actual shape of the survivability curve is considered. 80%
70%

Survivability
60%
50%

■ The most common non-parametric approach used in reliability is 40%

30%
the Kaplan-Meier method. 20%
10%
0%
0 100 200 300 400 500

■ Semi-parametric approaches are available however they involve Days

complex calculations (E.g. Cox’s Proportional Hazards or Cox’s

Regression Model).

Schlumberger-Private
Non-Parametric Method
Kaplan-Meier Method

■ Method developed in 1958 by Edward Kaplan and Paul

100%
Meier. 90%
80%
■ First developed for usage in medical researches. 70%

Survivability
60%
■ Nowadays the method is widely used in different 50%
40%
applications including reliability analysis. 30%
20%
■ Each step in Kaplan-Meier plot represents one event. 10%
0%
The size of the step depends on the number of events at 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0
Time
a certain period and the number of individuals still in risk
passed the period.

Schlumberger-Private
Non-Parametric Method 1

Kaplan-Meier Example 0.8

Survivability
0.6

0.4
■ Plot Kaplan-Meier Survival graph for the following data: 0.2

0
0 50 100 150 200 250 300 350
Time [Days]

Run Life Run Life End Number of Units at risk at ∏(nt – rt)/nt
(nt – rt)/nt
(Days) (Days) time [t] Failures, [rt] the time t, [nt]
ESP 1 244 Sort the run life ESP 2 30 0 0 7 1 1
ESP 2 information in 30 1 7 0.857143 0.857143
30 ESP 3 172
ascending order
ESP 3 172 ESP 7 174 172 1 6 0.833333 0.714286
ESP 4 303 ESP 6 224 174 1 5 0.8 0.571429
ESP 5 253 ESP 1 244 224 1 4 0.75 0.428571
ESP 6 224 ESP 5 253 244 1 3 0.666667 0.285714
ESP 7 174 ESP 4 303 253 1 2 0.5 0.142857
303 1 1 0 0

Schlumberger-Private
Survival Analysis
Complete and Censored Data

■ Complete data means that the value for each

individual is known.

■ In many cases, not all the individuals in the

population experienced an event at the time of the
analysis.

■ Right-censored data is widely used in ESP survival

analysis.

Schlumberger-Private
 Non-Parametric Method 1
0.9
0.8

Kaplan-Meier Example with Censored Data 0.7

Survivability
0.6
0.5
0.4

■ Plot Kaplan-Meier Survival graph for the following data:

0.3
0.2
0.1
0
Run Life Status Run Life 0 50 100 150 200 250 300 350
Status
(Days) (Days) Time [Days]

ESP 1 244 ESP Related

Running ESP 2 30
Failure End Number of Units at risk at ∏(nt – rt)/nt
ESP Related (nt – rt)/nt
ESP 2 30 ESP 3 172 Running time [t] Failures, [rt] the time t, [nt]
Failure
ESP 3 172 ESP Related 0 0 7 1 1
Running ESP 7 174
Failure
ESP 4 303 30 1 7 0.857143 0.857143
Running ESP 6 224 Running
ESP 5 253 172 0 6 1 0.857143
Running ESP 1 244 Running
ESP 6 224 174 1 5 0.8 0.685714
Running ESP 5 253 Running
ESP Related 224 0 4 1 0.685714
ESP 7 174 ESP 4 303 Running
Failure
244 0 3 1 0.685714
253 0 2 1 0.685714
303 0 1 1 0.685714

Schlumberger-Private
Non-Parametric Method
Kaplan-Meier Estimator
1
0.9
0.8
■ Similar to parametric distributions, it is possible to 0.7

Survivability
0.6
calculate the Mean-time to Failure using Kaplan-Meier 0.5

method. 0.4
0.3
0.2
0.1

■ MTTF is obtained by calculating the area under the

0
0 50 100 150 200 250 300 350

reliability curve.
Time [Days]

∞
𝑀𝑇𝑇𝐹 = න 𝑅 𝑡 . 𝑑𝑡
𝑀𝑇𝑇𝐹 = 30 ∗ 1 + 174 − 30 ∗ 0.857 + 303 − 174 ∗ 0.686
𝑡=0

𝑀𝑇𝑇𝐹 = 241 𝐷𝑎𝑦𝑠

Schlumberger-Private
Non-Parametric Method
Kaplan-Meier Plot
KM Right Censored
1
ESP Population Kaplan Meier Estimator
0.9
System A 417.9 Days
0.8
System B 479.0 Days
0.7
System C 450.9 Days
Survivability

0.6
0.5
0.4
0.3
0.2
0.1
0
0
25
50
75
100

675
125
150
175
200
225
250
275
300
325
350
375
400
425
450
475
500
525
550
575
600
625
650

700
725
750
775
Days

System A System B System C

Schlumberger-Private
 Non-Parametric Method
Kaplan-Meier Plot
KM Right Censored Exponential Distribution
1 1
0.9 0.9
0.8 0.8
0.7 0.7
Survivability

Survivability
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0

0
30
60
90

330

30
60
90

690
120
150
180
210
240
270
300

360
390
420
450
480
510
540
570
600
630
660
690
720
750
780

120
150
180
210
240
270
300
330
360
390
420
450
480
510
540
570
600
630
660

720
750
780
Days Days

System A System B System C System A System B System C

Schlumberger-Private
Survival Analysis
The Bathtub Curve
■ Infant Mortality
̶ High failure rate however decreasing with time.
̶ Caused by wrong product specification, flaws in the
installation, manufacturing defects, improper handling.
■ Normal Life
̶ Constant failure rate.
̶ ESP failures happen randomly.
■ End of Life Period
̶ Failure rate increases over time.
̶ Components start to fatigue or wear-out.

Schlumberger-Private
Parametric Method
Weibull Distribution

■ Parametric distribution defined in 1951 by Waloddi Weibull.

■ Widely used in reliability engineering and survival analysis.
■ Allows to model the survivability plot with time-dependant failure rate.
■ The shape of the distribution is defined by 2 parameters:
̶ β which is called shape parameter
̶ η which is called scale parameter or characteristic life
■ Both parameters are obtained based on run life data.
■ Once β and η are known, the reliability for any time t can be determined.

Schlumberger-Private
Parametric Method
Weibull Distribution
η=1 Hazard Function
η=1.75
Cumulative Distribution
𝑡 𝛽
Function (CDF) β=0.5 𝛽 𝑡 𝛽−1 − 𝜂
𝑒 𝛽 𝑡 𝛽−1

F(t)

F(t)
η=2.5 𝑓(𝑡) 𝜂 𝜂
𝑡 𝛽 H 𝑡 = = 𝑡 𝛽
=
− 𝑅(𝑡) − 𝜂
𝜂 𝜂
F 𝑡 =1− 𝑒 𝜂 β=1 𝑒
β=3
t t
β<1 β =1 β >1
η=1 β=3

Probability Distribution β=0.5

Function (PDF) η=1.75

H(t)
β=3 β=1
f(t)

f(t)
β=1 η=2.5
𝛽−1 𝑡 𝛽
𝛽 𝑡 − 𝜂
β=0.5
𝑓 𝑡 = 𝑒
𝜂 𝜂 t
t t

β=3
Mean Time to Failure
Reliability Function β=1
η=2.5 ∞
R(t)

R(t)

𝑡 𝛽 1
− β=0.5
𝑀𝑇𝑇𝐹 = 𝜂Γ +1 Γ(𝑥) = න 𝑒 −𝑡 𝑡 𝑥−1 𝑑𝑡
R 𝑡 =𝑒 𝜂
𝛽 𝑡=0
η=1.75
η=1
t t

Schlumberger-Private
Reliability Plots
Weibull Plot

Weibull Distribution
1.0
Weibull 0.9
ESP Population MTTF [Days]
Parameters 0.8
β = 0.98 0.7

Survivability
System A 672 0.6
η = 666 0.5
β = 5.14 0.4
System B 479 0.3
η = 521 0.2
β = 0.33 0.1
System C 10,725 0.0
η = 1726

90
30
60

480
120
150
180
210
240
270
300
330
360
390
420
450

510
540
570
600
630
660
690
720
750
780
Days

System A System B System C

Schlumberger-Private
 Reliability Plots
Methods Comparison
Weibull Distribution KM Right Censored Exponential Distribution
1.0 1 1

0.9 0.9 0.9

0.8 0.8 0.8

0.7 0.7 0.7

0.6 0.6 0.6

Survivability

0.5 0.5 0.5

0.4 0.4 0.4

0.3 0.3 0.3

0.2 0.2 0.2

0.1 0.1 0.1

0.0 0 0

40
80
0

120
160
200
240
280
320
360
400
440
480
520
560
600
640
680
720
760
40
80

40
80
0

0
120
160
200
240
280
320
360
400
440
480
520
560
600
640
680
720
760

120
160
200
240
280
320
360
400
440
480
520
560
600
640
680
720
760
Days Days Days

System A System B System C System A System B System C System A System B System C

Schlumberger-Private
Questions

Schlumberger-Private
Summary

■ ESP reliability metrics

̶ Average run life.
̶ Mean time to event.
■ Advantages and disadvantages of each metric.
■ Survival Analysis
̶ Parametric and non-parametric approaches.
■ Bathtub curve
̶ Influence of early failures and wear-out failures in the reliability plots.

Schlumberger-Private