SlideShare a Scribd company logo
Saudi Aramco: Company General Use
Jet Engine Remaining Useful
Life (RUL) Prediction
Ali S. Alhamaly
IT Engineering Dept.
10/31/2019
Saudi Aramco: Company General Use
2
The talk focuses on a data driven model to
estimate the remaining time for an engine in
operation
Problem definition, objective, and dataset
From raw data to health index (Training)
Using the model to estimate RUL
33
Problem definition
objective and dataset
1
Saudi Aramco: Company General Use
4
The problem is to predict RUL based on
measurements available from jet engines
TemperaturePressureSpeedFlowRate
Measurements from
engine components
Saudi Aramco: Company General Use
5
The objective to characterize the engine
health based on raw measurements
TemperaturePressureSpeedFlowRate
Measurements from
engine components
Characterized progression
of engine health
Final results :
“You have 35 cycles before your
engine is out of service”
Saudi Aramco: Company General Use
6
Dataset is composed of a simulated run-to-
failure from a fleet of similar engines
Fleet of similar engine
Run-to-failure simulated data
Using C-MAPSS – 21 sensors
Saxena, Abhinav & Goebel, Kai & Simon, Don & Eklund, Neil. (2008). Damage propagation modeling for aircraft engine run-to-failure simulation. International Conference on
Prognostics and Health Management. 10.1109/PHM.2008.4711414.
Trajectories of damage propagation from a
healthy state to failure
77
From raw data to
health index
2
Saudi Aramco: Company General Use
8
The approach is to explore the data, select
sensors, and then fuse them to make a
health index (HI)
Explore
sensors
Select
sensors
Fuse
sensors
and health
index
Saudi Aramco: Company General Use
9
the distribution of almost all sensors is e
skewed Gaussian
Saudi Aramco: Company General Use
10
Raw sensor data for all engines
End of lifeStart operation
Each color represents
different engine data
There is a general
trend toward end of
life
Some sensors lacks
consistent trend
toward end of life
Saudi Aramco: Company General Use
11
Sensors are selected based on magnitude of
linear trend
Number of sensors
used in modeling are
reduced to 6
PCA is done to
further reduce the
dimensionality to 3
only
Sensors that do not
change with time
were removed
Saudi Aramco: Company General Use
12
Sensors need to be fused into one-dimensional
health index (HI)
𝑯𝑰 𝒕 =
𝒊=𝟏
𝑵
𝜽𝒊 𝒙𝒊 𝒕 + 𝜽 𝟎
Healthy =1 End-life=0
Simple linear regression
is used to find 𝜽𝒊
Saudi Aramco: Company General Use
13
Distribution of RUL is used to determine total
life of a “100% healthy engine”
Not so many
engines have total
life longer than
300 cycles Cycles above which an
engine is 100% healthy
Saudi Aramco: Company General Use
14
Fused sensor (HI) as function of time after
obtaining regression coefficients
Sensors are transformed to HI (raw) and then exponential
fitting is applied for HI (raw) for each engine.
Saudi Aramco: Company General Use
15
Raw HI for all engines
The plot signifies the similarities and
variations between engines
1616
Using HI curves to
estimate RUL:
A similarity approach
3
Saudi Aramco: Company General Use
17
Predict RUL by comparing the total life of most
similar “training” engine to the new engine
New
engine
sensor
data
Transform
and fuse
to get
new HI
Compare
new HI to
the model
HI
Similarity
score
Estimate
RUL
Saudi Aramco: Company General Use
18
RUL of a new engine is based on HI model
that has the most similar degradation
pattern
T. Wang, Jianbo Yu, D. Siegel and J. Lee, "A similarity-based prognostics approach for Remaining Useful Life estimation of engineered systems," 2008 International Conference on
Prognostics and Health Management, Denver, CO, 2008, pp. 1-6.
Saudi Aramco: Company General Use
19
Finding similarity score by comparing new
data to each training engine HI curve
𝑺𝑺𝑫 =
𝒋=𝟏
𝑵 𝒏𝒆𝒘
𝑯𝑰 𝒋 𝒏𝒆𝒘 − 𝑯𝑰 𝒋 + 𝒕 𝟎 𝒎𝒐𝒅𝒆𝒍
𝟐
𝒕 𝟎 ∈ 𝟎, 𝑵 𝒅𝒊𝒇𝒇 , 𝑵 𝒅𝒊𝒇𝒇 = 𝑵 𝒎𝒐𝒅𝒆𝒍 − 𝑵 𝒏𝒆𝒘
Find 𝒕 𝟎 which minimizes 𝑺𝑺𝑫 for each
engine in the model  𝒕 𝟎
∗
Repeat process for each engine in
training data
𝑹𝑼𝑳 𝒎 = 𝑵 𝒅𝒊𝒇𝒇 𝒎
− 𝒕 𝟎
∗
𝒎
# data points in
new engine
# data points in
training engine
𝒎 ∈ 𝟏, 𝑴 , 𝑴 𝒕𝒐𝒕𝒂𝒍 𝒏𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝒕𝒓𝒂𝒊𝒏𝒊𝒏𝒈 𝒆𝒏𝒈𝒊𝒏𝒆𝒔
Saudi Aramco: Company General Use
20
Final RUL from several RUL candidate
𝑹𝑼𝑳 =
𝟏
𝑾
𝒎=𝟏
𝑴
𝒘 𝒎 𝑹𝑼𝑳 𝒎
𝑾 =
𝒎=𝟏
𝑴
𝒘 𝒎
• Median of all 𝑹𝑼𝑳 𝒎
• weighted mean based on SSD
• mean of max and min 𝑹𝑼𝑳 𝒎
• Mean of all 𝑹𝑼𝑳 𝒎
Saudi Aramco: Company General Use
21
Accuracy of the model is measured using
asymmetric exponential penalty score
𝒆𝒓𝒓𝒐𝒓 = 𝑹𝑼𝑳 − 𝑹𝑼𝑳
Saudi Aramco: Company General Use
22
Accuracy of the model is measured using
asymmetric exponential penalty score
• M1: Median
• M2: SSD weighted mean
• M3: mean of max and min
• M4: Mean
Saudi Aramco: Company General Use
23
TemperaturePressureSpeedFlowRate
𝑯𝑰 𝒕 =
𝒊=𝟏
𝑵
𝜽𝒊 𝒙𝒊 𝒕 + 𝜽 𝟎
In conclusion, similarity based method is a
simple and effective in estimate RUL

More Related Content

What's hot

Diagnostic in Adaptive AUTOSAR
Diagnostic in Adaptive AUTOSARDiagnostic in Adaptive AUTOSAR
Diagnostic in Adaptive AUTOSAR
Bernhard Wagner
 
UAV(unmanned aerial vehicle) and its application
UAV(unmanned aerial vehicle) and its application UAV(unmanned aerial vehicle) and its application
UAV(unmanned aerial vehicle) and its application
Joy Karmakar
 
Structural Compliance of Electric Vehicle Battery Pack
Structural Compliance of Electric Vehicle Battery Pack Structural Compliance of Electric Vehicle Battery Pack
Structural Compliance of Electric Vehicle Battery Pack
Arindam Chakraborty, Ph.D., P.E. (CA, TX)
 
Capella Days 2021 | Introduction to CAPELLA/ARCADIA and NASA Systems Engineer...
Capella Days 2021 | Introduction to CAPELLA/ARCADIA and NASA Systems Engineer...Capella Days 2021 | Introduction to CAPELLA/ARCADIA and NASA Systems Engineer...
Capella Days 2021 | Introduction to CAPELLA/ARCADIA and NASA Systems Engineer...
Obeo
 
[Capella Days 2020] Innovating with MBSE – Medical Device Example
[Capella Days 2020] Innovating with MBSE – Medical Device Example[Capella Days 2020] Innovating with MBSE – Medical Device Example
[Capella Days 2020] Innovating with MBSE – Medical Device Example
Obeo
 
Safe commercial airplane with separable cabins
Safe commercial airplane with separable cabinsSafe commercial airplane with separable cabins
Safe commercial airplane with separable cabins
Naser Ziaei
 
Adaptive AUTOSAR - The New AUTOSAR Architecture
Adaptive AUTOSAR - The New AUTOSAR ArchitectureAdaptive AUTOSAR - The New AUTOSAR Architecture
Adaptive AUTOSAR - The New AUTOSAR Architecture
AdaCore
 
Model based design-Hardware in loop-software in loop
Model based design-Hardware in loop-software in loopModel based design-Hardware in loop-software in loop
Model based design-Hardware in loop-software in loop
Mahmoud Hussein
 
Us navy introduction to helicopter aerodynamics workbook cnatra p-401 [us n...
Us navy   introduction to helicopter aerodynamics workbook cnatra p-401 [us n...Us navy   introduction to helicopter aerodynamics workbook cnatra p-401 [us n...
Us navy introduction to helicopter aerodynamics workbook cnatra p-401 [us n...
Mohamed Yaser
 
Assembly_Drawings_Balloons_Parts_Lists.ppt
Assembly_Drawings_Balloons_Parts_Lists.pptAssembly_Drawings_Balloons_Parts_Lists.ppt
Assembly_Drawings_Balloons_Parts_Lists.ppt
HygeniosMirador1
 
Aircraft propulsion by c.fayette taylor
Aircraft propulsion by c.fayette taylorAircraft propulsion by c.fayette taylor
Aircraft propulsion by c.fayette taylor
Aghilesh V
 
Avionics sai
Avionics saiAvionics sai
Avionics sai
Sai Shubhankar
 
State feedback control
State feedback controlState feedback control
State feedback control
Bassam AlGhram
 
The advantages of different types of propellants
The advantages of different types of propellantsThe advantages of different types of propellants
The advantages of different types of propellants
Arya Ramaru
 
Basic Aerodynamics Ii Stability Large
Basic Aerodynamics Ii Stability   LargeBasic Aerodynamics Ii Stability   Large
Basic Aerodynamics Ii Stability Large
lccmechanics
 
Mechanics of Machine lab manual c 20 20ME31P
Mechanics of Machine lab manual c 20 20ME31PMechanics of Machine lab manual c 20 20ME31P
Mechanics of Machine lab manual c 20 20ME31P
THANMAY JS
 
Kinematic Model vs Dynamic Model
Kinematic Model vs Dynamic ModelKinematic Model vs Dynamic Model
Kinematic Model vs Dynamic Model
Hitesh Mohapatra
 
Simulation with Python and MATLAB® in Capella
Simulation with Python and MATLAB® in CapellaSimulation with Python and MATLAB® in Capella
Simulation with Python and MATLAB® in Capella
Obeo
 
Fighter aircraft design adp 1
Fighter aircraft design adp 1Fighter aircraft design adp 1
Fighter aircraft design adp 1
Dudekula Jamal
 
Autosar fundamental
Autosar fundamentalAutosar fundamental
Autosar fundamental
Omkar Rane
 

What's hot (20)

Diagnostic in Adaptive AUTOSAR
Diagnostic in Adaptive AUTOSARDiagnostic in Adaptive AUTOSAR
Diagnostic in Adaptive AUTOSAR
 
UAV(unmanned aerial vehicle) and its application
UAV(unmanned aerial vehicle) and its application UAV(unmanned aerial vehicle) and its application
UAV(unmanned aerial vehicle) and its application
 
Structural Compliance of Electric Vehicle Battery Pack
Structural Compliance of Electric Vehicle Battery Pack Structural Compliance of Electric Vehicle Battery Pack
Structural Compliance of Electric Vehicle Battery Pack
 
Capella Days 2021 | Introduction to CAPELLA/ARCADIA and NASA Systems Engineer...
Capella Days 2021 | Introduction to CAPELLA/ARCADIA and NASA Systems Engineer...Capella Days 2021 | Introduction to CAPELLA/ARCADIA and NASA Systems Engineer...
Capella Days 2021 | Introduction to CAPELLA/ARCADIA and NASA Systems Engineer...
 
[Capella Days 2020] Innovating with MBSE – Medical Device Example
[Capella Days 2020] Innovating with MBSE – Medical Device Example[Capella Days 2020] Innovating with MBSE – Medical Device Example
[Capella Days 2020] Innovating with MBSE – Medical Device Example
 
Safe commercial airplane with separable cabins
Safe commercial airplane with separable cabinsSafe commercial airplane with separable cabins
Safe commercial airplane with separable cabins
 
Adaptive AUTOSAR - The New AUTOSAR Architecture
Adaptive AUTOSAR - The New AUTOSAR ArchitectureAdaptive AUTOSAR - The New AUTOSAR Architecture
Adaptive AUTOSAR - The New AUTOSAR Architecture
 
Model based design-Hardware in loop-software in loop
Model based design-Hardware in loop-software in loopModel based design-Hardware in loop-software in loop
Model based design-Hardware in loop-software in loop
 
Us navy introduction to helicopter aerodynamics workbook cnatra p-401 [us n...
Us navy   introduction to helicopter aerodynamics workbook cnatra p-401 [us n...Us navy   introduction to helicopter aerodynamics workbook cnatra p-401 [us n...
Us navy introduction to helicopter aerodynamics workbook cnatra p-401 [us n...
 
Assembly_Drawings_Balloons_Parts_Lists.ppt
Assembly_Drawings_Balloons_Parts_Lists.pptAssembly_Drawings_Balloons_Parts_Lists.ppt
Assembly_Drawings_Balloons_Parts_Lists.ppt
 
Aircraft propulsion by c.fayette taylor
Aircraft propulsion by c.fayette taylorAircraft propulsion by c.fayette taylor
Aircraft propulsion by c.fayette taylor
 
Avionics sai
Avionics saiAvionics sai
Avionics sai
 
State feedback control
State feedback controlState feedback control
State feedback control
 
The advantages of different types of propellants
The advantages of different types of propellantsThe advantages of different types of propellants
The advantages of different types of propellants
 
Basic Aerodynamics Ii Stability Large
Basic Aerodynamics Ii Stability   LargeBasic Aerodynamics Ii Stability   Large
Basic Aerodynamics Ii Stability Large
 
Mechanics of Machine lab manual c 20 20ME31P
Mechanics of Machine lab manual c 20 20ME31PMechanics of Machine lab manual c 20 20ME31P
Mechanics of Machine lab manual c 20 20ME31P
 
Kinematic Model vs Dynamic Model
Kinematic Model vs Dynamic ModelKinematic Model vs Dynamic Model
Kinematic Model vs Dynamic Model
 
Simulation with Python and MATLAB® in Capella
Simulation with Python and MATLAB® in CapellaSimulation with Python and MATLAB® in Capella
Simulation with Python and MATLAB® in Capella
 
Fighter aircraft design adp 1
Fighter aircraft design adp 1Fighter aircraft design adp 1
Fighter aircraft design adp 1
 
Autosar fundamental
Autosar fundamentalAutosar fundamental
Autosar fundamental
 

Similar to Jet engine remaining useful life prediction

Development of back propagation neural network model to predict performance a...
Development of back propagation neural network model to predict performance a...Development of back propagation neural network model to predict performance a...
Development of back propagation neural network model to predict performance a...
IAEME Publication
 
IRJET- Implementation of Ideal Stop and Go Deactivation Logic for Passeng...
IRJET-  	  Implementation of Ideal Stop and Go Deactivation Logic for Passeng...IRJET-  	  Implementation of Ideal Stop and Go Deactivation Logic for Passeng...
IRJET- Implementation of Ideal Stop and Go Deactivation Logic for Passeng...
IRJET Journal
 
Speed Control of PMDCM Based GA and DS Techniques
Speed Control of PMDCM Based GA and DS TechniquesSpeed Control of PMDCM Based GA and DS Techniques
Speed Control of PMDCM Based GA and DS Techniques
International Journal of Power Electronics and Drive Systems
 
Experimental Investigation on Performance of Turbo-matching of Turbocharger A...
Experimental Investigation on Performance of Turbo-matching of Turbocharger A...Experimental Investigation on Performance of Turbo-matching of Turbocharger A...
Experimental Investigation on Performance of Turbo-matching of Turbocharger A...
IRJET Journal
 
IRJET- A Review of Testing of Multi Cylinder S.I. Petrol Engine
IRJET-  	  A Review of Testing of Multi Cylinder S.I. Petrol EngineIRJET-  	  A Review of Testing of Multi Cylinder S.I. Petrol Engine
IRJET- A Review of Testing of Multi Cylinder S.I. Petrol Engine
IRJET Journal
 
Functional size measurement for processor load estimation hassan soubra
Functional size measurement for processor load estimation   hassan soubraFunctional size measurement for processor load estimation   hassan soubra
Functional size measurement for processor load estimation hassan soubra
IWSM Mensura
 
STUDY AND ANALYSIS OF DIFFERENTIAL CONTROLLED CONTINUOUSLY VARIABLE DRIVE
STUDY AND ANALYSIS OF DIFFERENTIAL CONTROLLED CONTINUOUSLY VARIABLE DRIVESTUDY AND ANALYSIS OF DIFFERENTIAL CONTROLLED CONTINUOUSLY VARIABLE DRIVE
STUDY AND ANALYSIS OF DIFFERENTIAL CONTROLLED CONTINUOUSLY VARIABLE DRIVE
Journal For Research
 
IRJET- Performance Analysis of ACO based PID Controller in AVR System: A ...
IRJET-  	  Performance Analysis of ACO based PID Controller in AVR System: A ...IRJET-  	  Performance Analysis of ACO based PID Controller in AVR System: A ...
IRJET- Performance Analysis of ACO based PID Controller in AVR System: A ...
IRJET Journal
 
Engine plant model_development for hil system and application to on-board dia...
Engine plant model_development for hil system and application to on-board dia...Engine plant model_development for hil system and application to on-board dia...
Engine plant model_development for hil system and application to on-board dia...
Yixin Chen
 
IRJET- A Review Paper on Pneumatic Controlled ABS
IRJET- A Review Paper on Pneumatic Controlled ABSIRJET- A Review Paper on Pneumatic Controlled ABS
IRJET- A Review Paper on Pneumatic Controlled ABS
IRJET Journal
 
Study and Analysis of Nonlinear Constrained Components A Study of Plug-in Hyb...
Study and Analysis of Nonlinear Constrained Components A Study of Plug-in Hyb...Study and Analysis of Nonlinear Constrained Components A Study of Plug-in Hyb...
Study and Analysis of Nonlinear Constrained Components A Study of Plug-in Hyb...
ijtsrd
 
Speed_Control_of_DC_Motor_Using_Particle_Swarm_Opt.pdf
Speed_Control_of_DC_Motor_Using_Particle_Swarm_Opt.pdfSpeed_Control_of_DC_Motor_Using_Particle_Swarm_Opt.pdf
Speed_Control_of_DC_Motor_Using_Particle_Swarm_Opt.pdf
ThienMai14
 
Data Mining Presentation.pptx
Data Mining Presentation.pptxData Mining Presentation.pptx
Data Mining Presentation.pptx
gina458018
 
2015_EEMODS_SynRM_Presentation_Final_09142015
2015_EEMODS_SynRM_Presentation_Final_091420152015_EEMODS_SynRM_Presentation_Final_09142015
2015_EEMODS_SynRM_Presentation_Final_09142015
Baskar Vairamohan, P.E., C.E.M.
 
IRJET- Analysis of Emission Data by using Testbed for Euro VI Norms
IRJET- Analysis of Emission Data by using Testbed for Euro VI NormsIRJET- Analysis of Emission Data by using Testbed for Euro VI Norms
IRJET- Analysis of Emission Data by using Testbed for Euro VI Norms
IRJET Journal
 
DTE_Industrial_Motors_Market_Research_Presentation.pdf
DTE_Industrial_Motors_Market_Research_Presentation.pdfDTE_Industrial_Motors_Market_Research_Presentation.pdf
DTE_Industrial_Motors_Market_Research_Presentation.pdf
AnkitJoshi309913
 
Fuel efficiency in construction equipment – optimize the machine as one system
Fuel efficiency in construction equipment – optimize the machine as one systemFuel efficiency in construction equipment – optimize the machine as one system
Fuel efficiency in construction equipment – optimize the machine as one system
Reno Filla
 
Performance Optimization of Diesel Engine with Chicken Waste Bio-diesel Blend...
Performance Optimization of Diesel Engine with Chicken Waste Bio-diesel Blend...Performance Optimization of Diesel Engine with Chicken Waste Bio-diesel Blend...
Performance Optimization of Diesel Engine with Chicken Waste Bio-diesel Blend...
IJERD Editor
 
Rbi masterclass fleming summit_ss
Rbi masterclass fleming summit_ssRbi masterclass fleming summit_ss
Rbi masterclass fleming summit_ss
Ganeshraju
 
IRJET- Design and Development of Centralized Tyre Pressure System
IRJET-  	  Design and Development of Centralized Tyre Pressure SystemIRJET-  	  Design and Development of Centralized Tyre Pressure System
IRJET- Design and Development of Centralized Tyre Pressure System
IRJET Journal
 

Similar to Jet engine remaining useful life prediction (20)

Development of back propagation neural network model to predict performance a...
Development of back propagation neural network model to predict performance a...Development of back propagation neural network model to predict performance a...
Development of back propagation neural network model to predict performance a...
 
IRJET- Implementation of Ideal Stop and Go Deactivation Logic for Passeng...
IRJET-  	  Implementation of Ideal Stop and Go Deactivation Logic for Passeng...IRJET-  	  Implementation of Ideal Stop and Go Deactivation Logic for Passeng...
IRJET- Implementation of Ideal Stop and Go Deactivation Logic for Passeng...
 
Speed Control of PMDCM Based GA and DS Techniques
Speed Control of PMDCM Based GA and DS TechniquesSpeed Control of PMDCM Based GA and DS Techniques
Speed Control of PMDCM Based GA and DS Techniques
 
Experimental Investigation on Performance of Turbo-matching of Turbocharger A...
Experimental Investigation on Performance of Turbo-matching of Turbocharger A...Experimental Investigation on Performance of Turbo-matching of Turbocharger A...
Experimental Investigation on Performance of Turbo-matching of Turbocharger A...
 
IRJET- A Review of Testing of Multi Cylinder S.I. Petrol Engine
IRJET-  	  A Review of Testing of Multi Cylinder S.I. Petrol EngineIRJET-  	  A Review of Testing of Multi Cylinder S.I. Petrol Engine
IRJET- A Review of Testing of Multi Cylinder S.I. Petrol Engine
 
Functional size measurement for processor load estimation hassan soubra
Functional size measurement for processor load estimation   hassan soubraFunctional size measurement for processor load estimation   hassan soubra
Functional size measurement for processor load estimation hassan soubra
 
STUDY AND ANALYSIS OF DIFFERENTIAL CONTROLLED CONTINUOUSLY VARIABLE DRIVE
STUDY AND ANALYSIS OF DIFFERENTIAL CONTROLLED CONTINUOUSLY VARIABLE DRIVESTUDY AND ANALYSIS OF DIFFERENTIAL CONTROLLED CONTINUOUSLY VARIABLE DRIVE
STUDY AND ANALYSIS OF DIFFERENTIAL CONTROLLED CONTINUOUSLY VARIABLE DRIVE
 
IRJET- Performance Analysis of ACO based PID Controller in AVR System: A ...
IRJET-  	  Performance Analysis of ACO based PID Controller in AVR System: A ...IRJET-  	  Performance Analysis of ACO based PID Controller in AVR System: A ...
IRJET- Performance Analysis of ACO based PID Controller in AVR System: A ...
 
Engine plant model_development for hil system and application to on-board dia...
Engine plant model_development for hil system and application to on-board dia...Engine plant model_development for hil system and application to on-board dia...
Engine plant model_development for hil system and application to on-board dia...
 
IRJET- A Review Paper on Pneumatic Controlled ABS
IRJET- A Review Paper on Pneumatic Controlled ABSIRJET- A Review Paper on Pneumatic Controlled ABS
IRJET- A Review Paper on Pneumatic Controlled ABS
 
Study and Analysis of Nonlinear Constrained Components A Study of Plug-in Hyb...
Study and Analysis of Nonlinear Constrained Components A Study of Plug-in Hyb...Study and Analysis of Nonlinear Constrained Components A Study of Plug-in Hyb...
Study and Analysis of Nonlinear Constrained Components A Study of Plug-in Hyb...
 
Speed_Control_of_DC_Motor_Using_Particle_Swarm_Opt.pdf
Speed_Control_of_DC_Motor_Using_Particle_Swarm_Opt.pdfSpeed_Control_of_DC_Motor_Using_Particle_Swarm_Opt.pdf
Speed_Control_of_DC_Motor_Using_Particle_Swarm_Opt.pdf
 
Data Mining Presentation.pptx
Data Mining Presentation.pptxData Mining Presentation.pptx
Data Mining Presentation.pptx
 
2015_EEMODS_SynRM_Presentation_Final_09142015
2015_EEMODS_SynRM_Presentation_Final_091420152015_EEMODS_SynRM_Presentation_Final_09142015
2015_EEMODS_SynRM_Presentation_Final_09142015
 
IRJET- Analysis of Emission Data by using Testbed for Euro VI Norms
IRJET- Analysis of Emission Data by using Testbed for Euro VI NormsIRJET- Analysis of Emission Data by using Testbed for Euro VI Norms
IRJET- Analysis of Emission Data by using Testbed for Euro VI Norms
 
DTE_Industrial_Motors_Market_Research_Presentation.pdf
DTE_Industrial_Motors_Market_Research_Presentation.pdfDTE_Industrial_Motors_Market_Research_Presentation.pdf
DTE_Industrial_Motors_Market_Research_Presentation.pdf
 
Fuel efficiency in construction equipment – optimize the machine as one system
Fuel efficiency in construction equipment – optimize the machine as one systemFuel efficiency in construction equipment – optimize the machine as one system
Fuel efficiency in construction equipment – optimize the machine as one system
 
Performance Optimization of Diesel Engine with Chicken Waste Bio-diesel Blend...
Performance Optimization of Diesel Engine with Chicken Waste Bio-diesel Blend...Performance Optimization of Diesel Engine with Chicken Waste Bio-diesel Blend...
Performance Optimization of Diesel Engine with Chicken Waste Bio-diesel Blend...
 
Rbi masterclass fleming summit_ss
Rbi masterclass fleming summit_ssRbi masterclass fleming summit_ss
Rbi masterclass fleming summit_ss
 
IRJET- Design and Development of Centralized Tyre Pressure System
IRJET-  	  Design and Development of Centralized Tyre Pressure SystemIRJET-  	  Design and Development of Centralized Tyre Pressure System
IRJET- Design and Development of Centralized Tyre Pressure System
 

Recently uploaded

CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURSCompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
RamonNovais6
 
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
Yasser Mahgoub
 
Rainfall intensity duration frequency curve statistical analysis and modeling...
Rainfall intensity duration frequency curve statistical analysis and modeling...Rainfall intensity duration frequency curve statistical analysis and modeling...
Rainfall intensity duration frequency curve statistical analysis and modeling...
bijceesjournal
 
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
Gino153088
 
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student MemberIEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
VICTOR MAESTRE RAMIREZ
 
cnn.pptx Convolutional neural network used for image classication
cnn.pptx Convolutional neural network used for image classicationcnn.pptx Convolutional neural network used for image classication
cnn.pptx Convolutional neural network used for image classication
SakkaravarthiShanmug
 
Material for memory and display system h
Material for memory and display system hMaterial for memory and display system h
Material for memory and display system h
gowrishankartb2005
 
Design and optimization of ion propulsion drone
Design and optimization of ion propulsion droneDesign and optimization of ion propulsion drone
Design and optimization of ion propulsion drone
bjmsejournal
 
artificial intelligence and data science contents.pptx
artificial intelligence and data science contents.pptxartificial intelligence and data science contents.pptx
artificial intelligence and data science contents.pptx
GauravCar
 
原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样
原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样
原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样
ydzowc
 
AI assisted telemedicine KIOSK for Rural India.pptx
AI assisted telemedicine KIOSK for Rural India.pptxAI assisted telemedicine KIOSK for Rural India.pptx
AI assisted telemedicine KIOSK for Rural India.pptx
architagupta876
 
Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...
IJECEIAES
 
哪里办理(csu毕业证书)查尔斯特大学毕业证硕士学历原版一模一样
哪里办理(csu毕业证书)查尔斯特大学毕业证硕士学历原版一模一样哪里办理(csu毕业证书)查尔斯特大学毕业证硕士学历原版一模一样
哪里办理(csu毕业证书)查尔斯特大学毕业证硕士学历原版一模一样
insn4465
 
Comparative analysis between traditional aquaponics and reconstructed aquapon...
Comparative analysis between traditional aquaponics and reconstructed aquapon...Comparative analysis between traditional aquaponics and reconstructed aquapon...
Comparative analysis between traditional aquaponics and reconstructed aquapon...
bijceesjournal
 
spirit beverages ppt without graphics.pptx
spirit beverages ppt without graphics.pptxspirit beverages ppt without graphics.pptx
spirit beverages ppt without graphics.pptx
Madan Karki
 
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
171ticu
 
Computational Engineering IITH Presentation
Computational Engineering IITH PresentationComputational Engineering IITH Presentation
Computational Engineering IITH Presentation
co23btech11018
 
Certificates - Mahmoud Mohamed Moursi Ahmed
Certificates - Mahmoud Mohamed Moursi AhmedCertificates - Mahmoud Mohamed Moursi Ahmed
Certificates - Mahmoud Mohamed Moursi Ahmed
Mahmoud Morsy
 
Welding Metallurgy Ferrous Materials.pdf
Welding Metallurgy Ferrous Materials.pdfWelding Metallurgy Ferrous Materials.pdf
Welding Metallurgy Ferrous Materials.pdf
AjmalKhan50578
 
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Sinan KOZAK
 

Recently uploaded (20)

CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURSCompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
 
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
 
Rainfall intensity duration frequency curve statistical analysis and modeling...
Rainfall intensity duration frequency curve statistical analysis and modeling...Rainfall intensity duration frequency curve statistical analysis and modeling...
Rainfall intensity duration frequency curve statistical analysis and modeling...
 
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
 
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student MemberIEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
 
cnn.pptx Convolutional neural network used for image classication
cnn.pptx Convolutional neural network used for image classicationcnn.pptx Convolutional neural network used for image classication
cnn.pptx Convolutional neural network used for image classication
 
Material for memory and display system h
Material for memory and display system hMaterial for memory and display system h
Material for memory and display system h
 
Design and optimization of ion propulsion drone
Design and optimization of ion propulsion droneDesign and optimization of ion propulsion drone
Design and optimization of ion propulsion drone
 
artificial intelligence and data science contents.pptx
artificial intelligence and data science contents.pptxartificial intelligence and data science contents.pptx
artificial intelligence and data science contents.pptx
 
原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样
原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样
原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样
 
AI assisted telemedicine KIOSK for Rural India.pptx
AI assisted telemedicine KIOSK for Rural India.pptxAI assisted telemedicine KIOSK for Rural India.pptx
AI assisted telemedicine KIOSK for Rural India.pptx
 
Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...
 
哪里办理(csu毕业证书)查尔斯特大学毕业证硕士学历原版一模一样
哪里办理(csu毕业证书)查尔斯特大学毕业证硕士学历原版一模一样哪里办理(csu毕业证书)查尔斯特大学毕业证硕士学历原版一模一样
哪里办理(csu毕业证书)查尔斯特大学毕业证硕士学历原版一模一样
 
Comparative analysis between traditional aquaponics and reconstructed aquapon...
Comparative analysis between traditional aquaponics and reconstructed aquapon...Comparative analysis between traditional aquaponics and reconstructed aquapon...
Comparative analysis between traditional aquaponics and reconstructed aquapon...
 
spirit beverages ppt without graphics.pptx
spirit beverages ppt without graphics.pptxspirit beverages ppt without graphics.pptx
spirit beverages ppt without graphics.pptx
 
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
 
Computational Engineering IITH Presentation
Computational Engineering IITH PresentationComputational Engineering IITH Presentation
Computational Engineering IITH Presentation
 
Certificates - Mahmoud Mohamed Moursi Ahmed
Certificates - Mahmoud Mohamed Moursi AhmedCertificates - Mahmoud Mohamed Moursi Ahmed
Certificates - Mahmoud Mohamed Moursi Ahmed
 
Welding Metallurgy Ferrous Materials.pdf
Welding Metallurgy Ferrous Materials.pdfWelding Metallurgy Ferrous Materials.pdf
Welding Metallurgy Ferrous Materials.pdf
 
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
 

Jet engine remaining useful life prediction

  • 1. Saudi Aramco: Company General Use Jet Engine Remaining Useful Life (RUL) Prediction Ali S. Alhamaly IT Engineering Dept. 10/31/2019
  • 2. Saudi Aramco: Company General Use 2 The talk focuses on a data driven model to estimate the remaining time for an engine in operation Problem definition, objective, and dataset From raw data to health index (Training) Using the model to estimate RUL
  • 4. Saudi Aramco: Company General Use 4 The problem is to predict RUL based on measurements available from jet engines TemperaturePressureSpeedFlowRate Measurements from engine components
  • 5. Saudi Aramco: Company General Use 5 The objective to characterize the engine health based on raw measurements TemperaturePressureSpeedFlowRate Measurements from engine components Characterized progression of engine health Final results : “You have 35 cycles before your engine is out of service”
  • 6. Saudi Aramco: Company General Use 6 Dataset is composed of a simulated run-to- failure from a fleet of similar engines Fleet of similar engine Run-to-failure simulated data Using C-MAPSS – 21 sensors Saxena, Abhinav & Goebel, Kai & Simon, Don & Eklund, Neil. (2008). Damage propagation modeling for aircraft engine run-to-failure simulation. International Conference on Prognostics and Health Management. 10.1109/PHM.2008.4711414. Trajectories of damage propagation from a healthy state to failure
  • 7. 77 From raw data to health index 2
  • 8. Saudi Aramco: Company General Use 8 The approach is to explore the data, select sensors, and then fuse them to make a health index (HI) Explore sensors Select sensors Fuse sensors and health index
  • 9. Saudi Aramco: Company General Use 9 the distribution of almost all sensors is e skewed Gaussian
  • 10. Saudi Aramco: Company General Use 10 Raw sensor data for all engines End of lifeStart operation Each color represents different engine data There is a general trend toward end of life Some sensors lacks consistent trend toward end of life
  • 11. Saudi Aramco: Company General Use 11 Sensors are selected based on magnitude of linear trend Number of sensors used in modeling are reduced to 6 PCA is done to further reduce the dimensionality to 3 only Sensors that do not change with time were removed
  • 12. Saudi Aramco: Company General Use 12 Sensors need to be fused into one-dimensional health index (HI) 𝑯𝑰 𝒕 = 𝒊=𝟏 𝑵 𝜽𝒊 𝒙𝒊 𝒕 + 𝜽 𝟎 Healthy =1 End-life=0 Simple linear regression is used to find 𝜽𝒊
  • 13. Saudi Aramco: Company General Use 13 Distribution of RUL is used to determine total life of a “100% healthy engine” Not so many engines have total life longer than 300 cycles Cycles above which an engine is 100% healthy
  • 14. Saudi Aramco: Company General Use 14 Fused sensor (HI) as function of time after obtaining regression coefficients Sensors are transformed to HI (raw) and then exponential fitting is applied for HI (raw) for each engine.
  • 15. Saudi Aramco: Company General Use 15 Raw HI for all engines The plot signifies the similarities and variations between engines
  • 16. 1616 Using HI curves to estimate RUL: A similarity approach 3
  • 17. Saudi Aramco: Company General Use 17 Predict RUL by comparing the total life of most similar “training” engine to the new engine New engine sensor data Transform and fuse to get new HI Compare new HI to the model HI Similarity score Estimate RUL
  • 18. Saudi Aramco: Company General Use 18 RUL of a new engine is based on HI model that has the most similar degradation pattern T. Wang, Jianbo Yu, D. Siegel and J. Lee, "A similarity-based prognostics approach for Remaining Useful Life estimation of engineered systems," 2008 International Conference on Prognostics and Health Management, Denver, CO, 2008, pp. 1-6.
  • 19. Saudi Aramco: Company General Use 19 Finding similarity score by comparing new data to each training engine HI curve 𝑺𝑺𝑫 = 𝒋=𝟏 𝑵 𝒏𝒆𝒘 𝑯𝑰 𝒋 𝒏𝒆𝒘 − 𝑯𝑰 𝒋 + 𝒕 𝟎 𝒎𝒐𝒅𝒆𝒍 𝟐 𝒕 𝟎 ∈ 𝟎, 𝑵 𝒅𝒊𝒇𝒇 , 𝑵 𝒅𝒊𝒇𝒇 = 𝑵 𝒎𝒐𝒅𝒆𝒍 − 𝑵 𝒏𝒆𝒘 Find 𝒕 𝟎 which minimizes 𝑺𝑺𝑫 for each engine in the model  𝒕 𝟎 ∗ Repeat process for each engine in training data 𝑹𝑼𝑳 𝒎 = 𝑵 𝒅𝒊𝒇𝒇 𝒎 − 𝒕 𝟎 ∗ 𝒎 # data points in new engine # data points in training engine 𝒎 ∈ 𝟏, 𝑴 , 𝑴 𝒕𝒐𝒕𝒂𝒍 𝒏𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝒕𝒓𝒂𝒊𝒏𝒊𝒏𝒈 𝒆𝒏𝒈𝒊𝒏𝒆𝒔
  • 20. Saudi Aramco: Company General Use 20 Final RUL from several RUL candidate 𝑹𝑼𝑳 = 𝟏 𝑾 𝒎=𝟏 𝑴 𝒘 𝒎 𝑹𝑼𝑳 𝒎 𝑾 = 𝒎=𝟏 𝑴 𝒘 𝒎 • Median of all 𝑹𝑼𝑳 𝒎 • weighted mean based on SSD • mean of max and min 𝑹𝑼𝑳 𝒎 • Mean of all 𝑹𝑼𝑳 𝒎
  • 21. Saudi Aramco: Company General Use 21 Accuracy of the model is measured using asymmetric exponential penalty score 𝒆𝒓𝒓𝒐𝒓 = 𝑹𝑼𝑳 − 𝑹𝑼𝑳
  • 22. Saudi Aramco: Company General Use 22 Accuracy of the model is measured using asymmetric exponential penalty score • M1: Median • M2: SSD weighted mean • M3: mean of max and min • M4: Mean
  • 23. Saudi Aramco: Company General Use 23 TemperaturePressureSpeedFlowRate 𝑯𝑰 𝒕 = 𝒊=𝟏 𝑵 𝜽𝒊 𝒙𝒊 𝒕 + 𝜽 𝟎 In conclusion, similarity based method is a simple and effective in estimate RUL

Editor's Notes

  1. Data simulated using C-MAPSS (Commercial Modular Aero-Propulsion System Simulation) The simulated data generated were used as challenge data for the 1st Prognostics and Health Management (PHM) data competition at PHM’08.
  2. from the above distribution plot, we can see that some sensors do not change and hence they can be removed from the analysis also the main takeaways: 1. the distribution of almost all variables is single skewed gaussian 2. op_cond_2 and sn_17 seem to be discrete variables and not continuous ( maybe these need to be deleted to keep only continuously varying variables for modeling) 3. the observations above holds when plotting all engines and when the plot is made for a specific engine. (means all engine are very similar in their output response)
  3. From the time series plot of all engine, we can find the below takeaways columns [op_cond_1, op_cond_2] do not have an apparent trend toward the end life of the engine. they are just random noise. so with great confidence, I can say that these two columns cannot help a predictive model that is based on the trend of the series to discover relevant information regarding estimating the RUL columns [s n_9, sn_14] indicates that the trend depends on the specific engine. some engines at the end of life tend to increase in these two columns while others tend to decrease. what is common about these two sensors is that the magnitude at the end life gets amplified. it is good to see that all other columns show an apparent trend as the fault propagate throughout the engine cycles and cause them to fail. [ this helps the model that will try to use the data to predict RUL :)
  4. based on the analysis of linear trend, the top 6 sensors are chosen based on the magnitude of their linear trend, i.e. the magnitude of their linear regression slope. it looks that based on these 6 sensors, taking 3 first principle components s captures about 90% of the data variability. hence the further reduction in dimensionality comes at a low loss of information. the sensors that do not change with time ( do not have variation with engine operational cycles) are dropped since they do not offer any information toward prediction the end of life the sensors that do not have apparent trend (looks like noise only, or do not have a trend toward the end of life) are dropped as well. this contains the sensors that behave differently for different engines ( since these will confuse the learning algorithm and can cause large testing errors since their behavior are not universal concerning all engines) based on linear regression of the remain sensor data with RUL, the highest 6 sensors in terms of the absolute values of the slopes are kept only. these sensors change predictably at the end of life for the engines. further, reduce the dimensionality by taking the first 3 principal components for the data the remaining 3 components of the data will be fused to make a Health Index (HI) function with RUL for each engine
  5. Based on the predictive health degradation curve (hp) from an offline system unit, an optimum fitting is conducted to determine a time-scale initial health condition (T0) that minimizes the sum of squared differences SSD between the online and offline health index