SlideShare ist ein Scribd-Unternehmen logo
1 von 21
svm@student.ethz.ch
Possible System Identification Approaches
PhD Student, Seyed Vahid Moosavi
Professor Ludger Hovestadt
14 June 2012
1
What is System Identification?
Predict the “Smoke Amount”
based on “the building features”
Bridge Displacement
Wind Flow
Stock Market
2
Why System Identification?
• System-State Space Modeling
• Sensitivity analysis
• What-if analysis
• System Design Optimization
3
System Identification In a Symbolic Form
System
(black-Box??)
Or
White box
Model Input
(e.g. load cases)
Some real
Phenomenon
(Model Parameters or
structure)
(e.g. strain or tension
parameters)
Some real observations
(e.g. test cases)
Model Output
(predicted)
(e.g. vertical
displacement of the
bridge)
Error
Threshold
(or confidence interval)
4
Classic System Identification
(We can easily measure any variable of interest)
5
And If we can’t measure enough or we want to Design a new
system….
We Build a Simulation Model
6
Goulet, J.-A., Kripakaran, P., and Smith, I.F.C. (2010). Multimodel structural performance monitoring. Journal of Structural Engineering,
136(10):13091318.
An Example
7
So…
A set Partial
Differential
Equations
(All possible instances of
comprehensive models)
Finite Element Analysis
Approximated
by
A
software
e.g. Ansys
We have
AND considerable Amount
of Parameters as INPUTs for
FEA
Through
Exhaustive
search
8
A sample result
Goulet, J.-A., Kripakaran, P., and Smith, I.F.C. (2010). Multimodel structural performance monitoring. Journal of Structural Engineering,
136(10):13091318.
9
But…
• Normally these simulations are time consuming.
no. of Parameters Np 5 10
No. of possible
values for each
parameter
Nc 5 5
Possible models Nmodels= Nc^Np 5^5=3125 5^10=9,765,625
Total required
time
(Nmodels) X
(time_per_model
)
3125 X 10= 31250
mins
=520 Hours
21 Days
162,7604 H
67,816 Days
10
What can we do?
What is the State of the art?
(Generalize It!!)
Surrogate Models
Meta-modeling
Response Surface Method
• Approximate the input-output of Comprehensive Model (e.g. FE)
with a faster approximation using “Statistical approaches”.
Toward a Black-Box Method
11
First step results… ( A sample case: Data and explanation by James A. Gulet)
Parameters:
1. Plymouth-side support longitudinal stiffness 1E [4, 11] kN/mm
2. Saltash-side support longitudinal stiffness 1E [4, 11] kN/mm
3. Deck expansion joint longitudinal stiffness 1E [4, 11] kN/mm
4. Main-cable initial strain [5E-4, 3E-3] mm/mm
5. Sidespan cable initial strains [5E-4, 3E-3] mm/mm.
The interval of each parameter value is discretized in five
parts to generate a hyper-grid containing 3125 (5^5)
combination of parameters. The result of this process is
an initial model set containing the predicted frequencies
and mode shapes for all 3125 model instances.
12
Sample data
Candidate
Rejected
• Classic approach is “Time Consuming” even for simplified models
• We used a sample of models from FE simulation with their final results.
• We trained a Self Organizing Map (SOM) to see the relation bet.
Different parameters values and the result of the FE model
13
Map interpretation
Normalized
Values of each
parameter
Each map is representing the value of one parameter in our FE model
Each dot in the map shows one possible model
The labels of each dot is either 0 (rejected) or 1 (accepted)
14
The effect of first Parameter
15
The effect of second Parameter
16
The effect of third Parameter
17
The effect of fourth Parameter
18
The effect of fifth Parameter
19
What we got so far
• Faster Model Tuning and sensitivity analysis
• Finding the most important parameters (so lower required
time for model generation)
• We somehow generalized the behavior of FE models
• We can conduct modified sampling method
And Next possible steps
Focusing on Surrogate Models for fast Statistical Models
And Applications in Design-Optimization
20
Thanks!
21

Weitere ähnliche Inhalte

Was ist angesagt?

Day2 Verilog HDL Basic
Day2 Verilog HDL BasicDay2 Verilog HDL Basic
Day2 Verilog HDL BasicRon Liu
 
Hamiltonian cycle in data structure 2
Hamiltonian cycle in data structure  2Hamiltonian cycle in data structure  2
Hamiltonian cycle in data structure 2Home
 
Field-programmable gate array
Field-programmable gate arrayField-programmable gate array
Field-programmable gate arrayPrinceArjun1999
 
Code Optimization
Code OptimizationCode Optimization
Code Optimizationguest9f8315
 
Lecture 1 - Learning Dynamical Systems from Demonstrations
Lecture 1 - Learning Dynamical Systems from DemonstrationsLecture 1 - Learning Dynamical Systems from Demonstrations
Lecture 1 - Learning Dynamical Systems from DemonstrationsNadia Barbara
 
Signal flow graph Mason’s Gain Formula
Signal flow graph Mason’s Gain Formula Signal flow graph Mason’s Gain Formula
Signal flow graph Mason’s Gain Formula vishalgohel12195
 
Chapter2 - Linear Time-Invariant System
Chapter2 - Linear Time-Invariant SystemChapter2 - Linear Time-Invariant System
Chapter2 - Linear Time-Invariant SystemAttaporn Ninsuwan
 
Signals and systems-3
Signals and systems-3Signals and systems-3
Signals and systems-3sarun soman
 
Verilog Lecture2 thhts
Verilog Lecture2 thhtsVerilog Lecture2 thhts
Verilog Lecture2 thhtsBéo Tú
 
Controllability and observability
Controllability and observabilityControllability and observability
Controllability and observabilityjawaharramaya
 
Introduction to wavelet transform
Introduction to wavelet transformIntroduction to wavelet transform
Introduction to wavelet transformRaj Endiran
 
Signal flow graph (sfg)
Signal flow graph (sfg)Signal flow graph (sfg)
Signal flow graph (sfg)Dhruv Shah
 
Automatic Test Pattern Generation (Testing of VLSI Design)
Automatic Test Pattern Generation (Testing of VLSI Design)Automatic Test Pattern Generation (Testing of VLSI Design)
Automatic Test Pattern Generation (Testing of VLSI Design)Usha Mehta
 
MODEL TRAIN CONTROLLER.pptx
MODEL TRAIN CONTROLLER.pptxMODEL TRAIN CONTROLLER.pptx
MODEL TRAIN CONTROLLER.pptxjanaki raman
 

Was ist angesagt? (20)

Day2 Verilog HDL Basic
Day2 Verilog HDL BasicDay2 Verilog HDL Basic
Day2 Verilog HDL Basic
 
Loop optimization
Loop optimizationLoop optimization
Loop optimization
 
Hamiltonian cycle in data structure 2
Hamiltonian cycle in data structure  2Hamiltonian cycle in data structure  2
Hamiltonian cycle in data structure 2
 
Field-programmable gate array
Field-programmable gate arrayField-programmable gate array
Field-programmable gate array
 
Code Optimization
Code OptimizationCode Optimization
Code Optimization
 
Lecture 1 - Learning Dynamical Systems from Demonstrations
Lecture 1 - Learning Dynamical Systems from DemonstrationsLecture 1 - Learning Dynamical Systems from Demonstrations
Lecture 1 - Learning Dynamical Systems from Demonstrations
 
Signal flow graph Mason’s Gain Formula
Signal flow graph Mason’s Gain Formula Signal flow graph Mason’s Gain Formula
Signal flow graph Mason’s Gain Formula
 
Chapter2 - Linear Time-Invariant System
Chapter2 - Linear Time-Invariant SystemChapter2 - Linear Time-Invariant System
Chapter2 - Linear Time-Invariant System
 
07 robot arm kinematics
07 robot arm kinematics07 robot arm kinematics
07 robot arm kinematics
 
Solved problems
Solved problemsSolved problems
Solved problems
 
Signals and systems-3
Signals and systems-3Signals and systems-3
Signals and systems-3
 
Verilog Lecture2 thhts
Verilog Lecture2 thhtsVerilog Lecture2 thhts
Verilog Lecture2 thhts
 
Controllability and observability
Controllability and observabilityControllability and observability
Controllability and observability
 
Vlsi design-styles
Vlsi design-stylesVlsi design-styles
Vlsi design-styles
 
Study of vlsi design methodologies and limitations using cad tools for cmos t...
Study of vlsi design methodologies and limitations using cad tools for cmos t...Study of vlsi design methodologies and limitations using cad tools for cmos t...
Study of vlsi design methodologies and limitations using cad tools for cmos t...
 
Lecture 23 24-time_response
Lecture 23 24-time_responseLecture 23 24-time_response
Lecture 23 24-time_response
 
Introduction to wavelet transform
Introduction to wavelet transformIntroduction to wavelet transform
Introduction to wavelet transform
 
Signal flow graph (sfg)
Signal flow graph (sfg)Signal flow graph (sfg)
Signal flow graph (sfg)
 
Automatic Test Pattern Generation (Testing of VLSI Design)
Automatic Test Pattern Generation (Testing of VLSI Design)Automatic Test Pattern Generation (Testing of VLSI Design)
Automatic Test Pattern Generation (Testing of VLSI Design)
 
MODEL TRAIN CONTROLLER.pptx
MODEL TRAIN CONTROLLER.pptxMODEL TRAIN CONTROLLER.pptx
MODEL TRAIN CONTROLLER.pptx
 

Ähnlich wie System identification and Surrogate Modeling

Time series and panel data in econometrics
Time series and panel data in econometricsTime series and panel data in econometrics
Time series and panel data in econometricsTorajInteaash2
 
Lecture 1
Lecture 1Lecture 1
Lecture 1eseem
 
ders 6 Panel data analysis.pptx
ders 6 Panel data analysis.pptxders 6 Panel data analysis.pptx
ders 6 Panel data analysis.pptxErgin Akalpler
 
Modeling and Simulation of Electrical Power Systems using OpenIPSL.org and Gr...
Modeling and Simulation of Electrical Power Systems using OpenIPSL.org and Gr...Modeling and Simulation of Electrical Power Systems using OpenIPSL.org and Gr...
Modeling and Simulation of Electrical Power Systems using OpenIPSL.org and Gr...Luigi Vanfretti
 
Design and development of a multi configuration beam vibration test setup
Design and development of a multi configuration beam vibration test setupDesign and development of a multi configuration beam vibration test setup
Design and development of a multi configuration beam vibration test setupIAEME Publication
 
Design and development of a multi configuration beam vibration test setup
Design and development of a multi configuration beam vibration test setupDesign and development of a multi configuration beam vibration test setup
Design and development of a multi configuration beam vibration test setupIAEME Publication
 
Investigation of Effects of impact loads on Framed Structures
Investigation of Effects of impact loads on Framed StructuresInvestigation of Effects of impact loads on Framed Structures
Investigation of Effects of impact loads on Framed StructuresIJMER
 
Role of Simulation in Deep Drawn Cylindrical Part
Role of Simulation in Deep Drawn Cylindrical PartRole of Simulation in Deep Drawn Cylindrical Part
Role of Simulation in Deep Drawn Cylindrical PartIJSRD
 
A WORKSPACE SIMULATION FOR TAL TR-2 ARTICULATED ROBOT
A WORKSPACE SIMULATION FOR TAL TR-2 ARTICULATED ROBOT A WORKSPACE SIMULATION FOR TAL TR-2 ARTICULATED ROBOT
A WORKSPACE SIMULATION FOR TAL TR-2 ARTICULATED ROBOT IAEME Publication
 
AIAA-Aviation-2015-Mehmani
AIAA-Aviation-2015-MehmaniAIAA-Aviation-2015-Mehmani
AIAA-Aviation-2015-MehmaniOptiModel
 
Cross-validation aggregation for forecasting
Cross-validation aggregation for forecastingCross-validation aggregation for forecasting
Cross-validation aggregation for forecastingDevon Barrow
 
An Overview of Performance Evaluation & Simulation
An Overview of Performance Evaluation & SimulationAn Overview of Performance Evaluation & Simulation
An Overview of Performance Evaluation & Simulationdasdfadfdsfsdfasdf
 
introduction to modeling, Types of Models, Classification of mathematical mod...
introduction to modeling, Types of Models, Classification of mathematical mod...introduction to modeling, Types of Models, Classification of mathematical mod...
introduction to modeling, Types of Models, Classification of mathematical mod...Waqas Afzal
 
IRJET- A Study on Some Repairable Systems
IRJET- A Study on Some Repairable SystemsIRJET- A Study on Some Repairable Systems
IRJET- A Study on Some Repairable SystemsIRJET Journal
 
One-Sample Face Recognition Using HMM Model of Fiducial Areas
One-Sample Face Recognition Using HMM Model of Fiducial AreasOne-Sample Face Recognition Using HMM Model of Fiducial Areas
One-Sample Face Recognition Using HMM Model of Fiducial AreasCSCJournals
 

Ähnlich wie System identification and Surrogate Modeling (20)

Time series and panel data in econometrics
Time series and panel data in econometricsTime series and panel data in econometrics
Time series and panel data in econometrics
 
Benson IMDC2012
Benson IMDC2012Benson IMDC2012
Benson IMDC2012
 
Variation response method CAE simulation suite
Variation response method CAE simulation suiteVariation response method CAE simulation suite
Variation response method CAE simulation suite
 
Lecture 1
Lecture 1Lecture 1
Lecture 1
 
ders 6 Panel data analysis.pptx
ders 6 Panel data analysis.pptxders 6 Panel data analysis.pptx
ders 6 Panel data analysis.pptx
 
5626
56265626
5626
 
lecture 1.pptx
lecture 1.pptxlecture 1.pptx
lecture 1.pptx
 
Modeling and Simulation of Electrical Power Systems using OpenIPSL.org and Gr...
Modeling and Simulation of Electrical Power Systems using OpenIPSL.org and Gr...Modeling and Simulation of Electrical Power Systems using OpenIPSL.org and Gr...
Modeling and Simulation of Electrical Power Systems using OpenIPSL.org and Gr...
 
Design and development of a multi configuration beam vibration test setup
Design and development of a multi configuration beam vibration test setupDesign and development of a multi configuration beam vibration test setup
Design and development of a multi configuration beam vibration test setup
 
Design and development of a multi configuration beam vibration test setup
Design and development of a multi configuration beam vibration test setupDesign and development of a multi configuration beam vibration test setup
Design and development of a multi configuration beam vibration test setup
 
Investigation of Effects of impact loads on Framed Structures
Investigation of Effects of impact loads on Framed StructuresInvestigation of Effects of impact loads on Framed Structures
Investigation of Effects of impact loads on Framed Structures
 
Role of Simulation in Deep Drawn Cylindrical Part
Role of Simulation in Deep Drawn Cylindrical PartRole of Simulation in Deep Drawn Cylindrical Part
Role of Simulation in Deep Drawn Cylindrical Part
 
A WORKSPACE SIMULATION FOR TAL TR-2 ARTICULATED ROBOT
A WORKSPACE SIMULATION FOR TAL TR-2 ARTICULATED ROBOT A WORKSPACE SIMULATION FOR TAL TR-2 ARTICULATED ROBOT
A WORKSPACE SIMULATION FOR TAL TR-2 ARTICULATED ROBOT
 
AIAA-Aviation-2015-Mehmani
AIAA-Aviation-2015-MehmaniAIAA-Aviation-2015-Mehmani
AIAA-Aviation-2015-Mehmani
 
Cross-validation aggregation for forecasting
Cross-validation aggregation for forecastingCross-validation aggregation for forecasting
Cross-validation aggregation for forecasting
 
An Overview of Performance Evaluation & Simulation
An Overview of Performance Evaluation & SimulationAn Overview of Performance Evaluation & Simulation
An Overview of Performance Evaluation & Simulation
 
introduction to modeling, Types of Models, Classification of mathematical mod...
introduction to modeling, Types of Models, Classification of mathematical mod...introduction to modeling, Types of Models, Classification of mathematical mod...
introduction to modeling, Types of Models, Classification of mathematical mod...
 
IRJET- A Study on Some Repairable Systems
IRJET- A Study on Some Repairable SystemsIRJET- A Study on Some Repairable Systems
IRJET- A Study on Some Repairable Systems
 
Simulation
SimulationSimulation
Simulation
 
One-Sample Face Recognition Using HMM Model of Fiducial Areas
One-Sample Face Recognition Using HMM Model of Fiducial AreasOne-Sample Face Recognition Using HMM Model of Fiducial Areas
One-Sample Face Recognition Using HMM Model of Fiducial Areas
 

Mehr von Vahid Moosavi

On Optimization Problems
On Optimization ProblemsOn Optimization Problems
On Optimization ProblemsVahid Moosavi
 
Toward of a Theory of Modeling
Toward of a Theory of ModelingToward of a Theory of Modeling
Toward of a Theory of ModelingVahid Moosavi
 
Markovian Modeling of Urban Traffic Flows in Coexistence With Urban Data Streams
Markovian Modeling of Urban Traffic Flows in Coexistence With Urban Data StreamsMarkovian Modeling of Urban Traffic Flows in Coexistence With Urban Data Streams
Markovian Modeling of Urban Traffic Flows in Coexistence With Urban Data StreamsVahid Moosavi
 
Finding Candidate Locations for Aerosol Pollution Monitoring at Street Level ...
Finding Candidate Locations for Aerosol Pollution Monitoring at Street Level ...Finding Candidate Locations for Aerosol Pollution Monitoring at Street Level ...
Finding Candidate Locations for Aerosol Pollution Monitoring at Street Level ...Vahid Moosavi
 
Modeling in Coexistence with Data: Toward a Generic Notion of Space
Modeling in Coexistence with Data: Toward a Generic Notion of SpaceModeling in Coexistence with Data: Toward a Generic Notion of Space
Modeling in Coexistence with Data: Toward a Generic Notion of SpaceVahid Moosavi
 
Data Driven Modeling Beyond Idealization
Data Driven Modeling Beyond IdealizationData Driven Modeling Beyond Idealization
Data Driven Modeling Beyond IdealizationVahid Moosavi
 
Pre-Specific Modeling - Computational Machines in Coexistence with Urban Data...
Pre-Specific Modeling - Computational Machines in Coexistence with Urban Data...Pre-Specific Modeling - Computational Machines in Coexistence with Urban Data...
Pre-Specific Modeling - Computational Machines in Coexistence with Urban Data...Vahid Moosavi
 
Bi g data_urban modeling_applications_23092013
Bi g data_urban modeling_applications_23092013Bi g data_urban modeling_applications_23092013
Bi g data_urban modeling_applications_23092013Vahid Moosavi
 
Bi g data_urban modeling_21082013
Bi g data_urban modeling_21082013Bi g data_urban modeling_21082013
Bi g data_urban modeling_21082013Vahid Moosavi
 
Cluster labeling fcl_weeklymeeting30102013
Cluster labeling fcl_weeklymeeting30102013Cluster labeling fcl_weeklymeeting30102013
Cluster labeling fcl_weeklymeeting30102013Vahid Moosavi
 

Mehr von Vahid Moosavi (10)

On Optimization Problems
On Optimization ProblemsOn Optimization Problems
On Optimization Problems
 
Toward of a Theory of Modeling
Toward of a Theory of ModelingToward of a Theory of Modeling
Toward of a Theory of Modeling
 
Markovian Modeling of Urban Traffic Flows in Coexistence With Urban Data Streams
Markovian Modeling of Urban Traffic Flows in Coexistence With Urban Data StreamsMarkovian Modeling of Urban Traffic Flows in Coexistence With Urban Data Streams
Markovian Modeling of Urban Traffic Flows in Coexistence With Urban Data Streams
 
Finding Candidate Locations for Aerosol Pollution Monitoring at Street Level ...
Finding Candidate Locations for Aerosol Pollution Monitoring at Street Level ...Finding Candidate Locations for Aerosol Pollution Monitoring at Street Level ...
Finding Candidate Locations for Aerosol Pollution Monitoring at Street Level ...
 
Modeling in Coexistence with Data: Toward a Generic Notion of Space
Modeling in Coexistence with Data: Toward a Generic Notion of SpaceModeling in Coexistence with Data: Toward a Generic Notion of Space
Modeling in Coexistence with Data: Toward a Generic Notion of Space
 
Data Driven Modeling Beyond Idealization
Data Driven Modeling Beyond IdealizationData Driven Modeling Beyond Idealization
Data Driven Modeling Beyond Idealization
 
Pre-Specific Modeling - Computational Machines in Coexistence with Urban Data...
Pre-Specific Modeling - Computational Machines in Coexistence with Urban Data...Pre-Specific Modeling - Computational Machines in Coexistence with Urban Data...
Pre-Specific Modeling - Computational Machines in Coexistence with Urban Data...
 
Bi g data_urban modeling_applications_23092013
Bi g data_urban modeling_applications_23092013Bi g data_urban modeling_applications_23092013
Bi g data_urban modeling_applications_23092013
 
Bi g data_urban modeling_21082013
Bi g data_urban modeling_21082013Bi g data_urban modeling_21082013
Bi g data_urban modeling_21082013
 
Cluster labeling fcl_weeklymeeting30102013
Cluster labeling fcl_weeklymeeting30102013Cluster labeling fcl_weeklymeeting30102013
Cluster labeling fcl_weeklymeeting30102013
 

Kürzlich hochgeladen

Boost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivityBoost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivityPrincipled Technologies
 
A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?Igalia
 
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsJoaquim Jorge
 
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsHandwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsMaria Levchenko
 
Data Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt RobisonData Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt RobisonAnna Loughnan Colquhoun
 
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Drew Madelung
 
CNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of ServiceCNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of Servicegiselly40
 
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptx
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptxEIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptx
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptxEarley Information Science
 
Tata AIG General Insurance Company - Insurer Innovation Award 2024
Tata AIG General Insurance Company - Insurer Innovation Award 2024Tata AIG General Insurance Company - Insurer Innovation Award 2024
Tata AIG General Insurance Company - Insurer Innovation Award 2024The Digital Insurer
 
Exploring the Future Potential of AI-Enabled Smartphone Processors
Exploring the Future Potential of AI-Enabled Smartphone ProcessorsExploring the Future Potential of AI-Enabled Smartphone Processors
Exploring the Future Potential of AI-Enabled Smartphone Processorsdebabhi2
 
08448380779 Call Girls In Civil Lines Women Seeking Men
08448380779 Call Girls In Civil Lines Women Seeking Men08448380779 Call Girls In Civil Lines Women Seeking Men
08448380779 Call Girls In Civil Lines Women Seeking MenDelhi Call girls
 
Boost Fertility New Invention Ups Success Rates.pdf
Boost Fertility New Invention Ups Success Rates.pdfBoost Fertility New Invention Ups Success Rates.pdf
Boost Fertility New Invention Ups Success Rates.pdfsudhanshuwaghmare1
 
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc
 
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptxHampshireHUG
 
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...Martijn de Jong
 
Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...
Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...
Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...apidays
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreternaman860154
 
Scaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationScaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationRadu Cotescu
 
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...Igalia
 
Advantages of Hiring UIUX Design Service Providers for Your Business
Advantages of Hiring UIUX Design Service Providers for Your BusinessAdvantages of Hiring UIUX Design Service Providers for Your Business
Advantages of Hiring UIUX Design Service Providers for Your BusinessPixlogix Infotech
 

Kürzlich hochgeladen (20)

Boost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivityBoost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivity
 
A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?
 
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
 
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsHandwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed texts
 
Data Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt RobisonData Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt Robison
 
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
 
CNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of ServiceCNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of Service
 
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptx
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptxEIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptx
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptx
 
Tata AIG General Insurance Company - Insurer Innovation Award 2024
Tata AIG General Insurance Company - Insurer Innovation Award 2024Tata AIG General Insurance Company - Insurer Innovation Award 2024
Tata AIG General Insurance Company - Insurer Innovation Award 2024
 
Exploring the Future Potential of AI-Enabled Smartphone Processors
Exploring the Future Potential of AI-Enabled Smartphone ProcessorsExploring the Future Potential of AI-Enabled Smartphone Processors
Exploring the Future Potential of AI-Enabled Smartphone Processors
 
08448380779 Call Girls In Civil Lines Women Seeking Men
08448380779 Call Girls In Civil Lines Women Seeking Men08448380779 Call Girls In Civil Lines Women Seeking Men
08448380779 Call Girls In Civil Lines Women Seeking Men
 
Boost Fertility New Invention Ups Success Rates.pdf
Boost Fertility New Invention Ups Success Rates.pdfBoost Fertility New Invention Ups Success Rates.pdf
Boost Fertility New Invention Ups Success Rates.pdf
 
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
 
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
 
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...
 
Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...
Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...
Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreter
 
Scaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationScaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organization
 
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
 
Advantages of Hiring UIUX Design Service Providers for Your Business
Advantages of Hiring UIUX Design Service Providers for Your BusinessAdvantages of Hiring UIUX Design Service Providers for Your Business
Advantages of Hiring UIUX Design Service Providers for Your Business
 

System identification and Surrogate Modeling

  • 1. svm@student.ethz.ch Possible System Identification Approaches PhD Student, Seyed Vahid Moosavi Professor Ludger Hovestadt 14 June 2012 1
  • 2. What is System Identification? Predict the “Smoke Amount” based on “the building features” Bridge Displacement Wind Flow Stock Market 2
  • 3. Why System Identification? • System-State Space Modeling • Sensitivity analysis • What-if analysis • System Design Optimization 3
  • 4. System Identification In a Symbolic Form System (black-Box??) Or White box Model Input (e.g. load cases) Some real Phenomenon (Model Parameters or structure) (e.g. strain or tension parameters) Some real observations (e.g. test cases) Model Output (predicted) (e.g. vertical displacement of the bridge) Error Threshold (or confidence interval) 4
  • 5. Classic System Identification (We can easily measure any variable of interest) 5
  • 6. And If we can’t measure enough or we want to Design a new system…. We Build a Simulation Model 6
  • 7. Goulet, J.-A., Kripakaran, P., and Smith, I.F.C. (2010). Multimodel structural performance monitoring. Journal of Structural Engineering, 136(10):13091318. An Example 7
  • 8. So… A set Partial Differential Equations (All possible instances of comprehensive models) Finite Element Analysis Approximated by A software e.g. Ansys We have AND considerable Amount of Parameters as INPUTs for FEA Through Exhaustive search 8
  • 9. A sample result Goulet, J.-A., Kripakaran, P., and Smith, I.F.C. (2010). Multimodel structural performance monitoring. Journal of Structural Engineering, 136(10):13091318. 9
  • 10. But… • Normally these simulations are time consuming. no. of Parameters Np 5 10 No. of possible values for each parameter Nc 5 5 Possible models Nmodels= Nc^Np 5^5=3125 5^10=9,765,625 Total required time (Nmodels) X (time_per_model ) 3125 X 10= 31250 mins =520 Hours 21 Days 162,7604 H 67,816 Days 10
  • 11. What can we do? What is the State of the art? (Generalize It!!) Surrogate Models Meta-modeling Response Surface Method • Approximate the input-output of Comprehensive Model (e.g. FE) with a faster approximation using “Statistical approaches”. Toward a Black-Box Method 11
  • 12. First step results… ( A sample case: Data and explanation by James A. Gulet) Parameters: 1. Plymouth-side support longitudinal stiffness 1E [4, 11] kN/mm 2. Saltash-side support longitudinal stiffness 1E [4, 11] kN/mm 3. Deck expansion joint longitudinal stiffness 1E [4, 11] kN/mm 4. Main-cable initial strain [5E-4, 3E-3] mm/mm 5. Sidespan cable initial strains [5E-4, 3E-3] mm/mm. The interval of each parameter value is discretized in five parts to generate a hyper-grid containing 3125 (5^5) combination of parameters. The result of this process is an initial model set containing the predicted frequencies and mode shapes for all 3125 model instances. 12
  • 13. Sample data Candidate Rejected • Classic approach is “Time Consuming” even for simplified models • We used a sample of models from FE simulation with their final results. • We trained a Self Organizing Map (SOM) to see the relation bet. Different parameters values and the result of the FE model 13
  • 14. Map interpretation Normalized Values of each parameter Each map is representing the value of one parameter in our FE model Each dot in the map shows one possible model The labels of each dot is either 0 (rejected) or 1 (accepted) 14
  • 15. The effect of first Parameter 15
  • 16. The effect of second Parameter 16
  • 17. The effect of third Parameter 17
  • 18. The effect of fourth Parameter 18
  • 19. The effect of fifth Parameter 19
  • 20. What we got so far • Faster Model Tuning and sensitivity analysis • Finding the most important parameters (so lower required time for model generation) • We somehow generalized the behavior of FE models • We can conduct modified sampling method And Next possible steps Focusing on Surrogate Models for fast Statistical Models And Applications in Design-Optimization 20