Advances in Uncertainty Quantification and Optimization Under Uncertainty with Aerospace Applications: Proceedings of the 2020 UQOP International Conference (Space Technology Proceedings, 8)

✍ Scribed by Massimiliano Vasile (editor), Domenico Quagliarella (editor)

• Publisher: Springer
• Year: 2022
• Tongue: English
• Leaves: 448
• Edition: 1st ed. 2021
• Category: Library

✦ Synopsis

The 2020 International Conference on Uncertainty Quantification & Optimization gathered together internationally renowned researchers in the fields of optimization and uncertainty quantification. The resulting proceedings cover all related aspects of computational uncertainty management and optimization, with particular emphasis on aerospace engineering problems.

The book contributions are organized under four major themes:

• Applications of Uncertainty in Aerospace & Engineering
• Imprecise Probability, Theory and Applications
• Robust and Reliability-Based Design Optimisation in Aerospace Engineering
• Uncertainty Quantification, Identification and Calibration in Aerospace Models

This proceedings volume is useful across disciplines, as it brings the expertise of theoretical and application researchers together in a unified framework.

✦ Table of Contents

Preface
Contents
Part I Applications of Uncertainty in Aerospace & Engineering (ENG)
From Uncertainty Quantification to Shape Optimization: Cross-Fertilization of Methods for Dimensionality Reduction
1 Introduction
2 Design-Space Dimensionality Reduction in Shape Optimization
2.1 Geometry-Based Formulation
2.2 Physics-Informed Formulation
3 Example Application
4 Concluding Remarks
References
Cloud Uncertainty Quantification for Runback Ice Formations in Anti-Ice Electro-Thermal Ice Protection Systems
Nomenclature
1 Introduction
2 Modelling of an AI-ETIPS
2.1 Computational Model
2.2 Case of Study
3 Cloud Uncertainty Characterization
4 Uncertainty Propagation Methodologies
4.1 Monte Carlo Sampling Methods
4.2 Generalized Polynomial Chaos Expansion
5 Numerical Results
6 Concluding Remarks
References
Multi-fidelity Surrogate Assisted Design Optimisation of an Airfoil under Uncertainty Using Far-Field Drag Approximation
1 Introduction
2 Multi-fidelity Gaussian Process Regression
3 Aerodynamic Computational Chain
4 Far-Field Drag Coefficient Calculation
5 Deterministic Design Optimisation Problem
6 Probabilistic Design Optimisation Problem
7 Optimisation Pipeline
8 Results
8.1 Deterministic Optimisation
8.2 Probabilistic Optimisation
9 Conclusion
References
Scalable Dynamic Asynchronous Monte Carlo Framework Applied to Wind Engineering Problems
1 Introduction
2 Monte Carlo Methods
2.1 Monte Carlo
2.2 Asynchronous Monte Carlo
2.3 Scheduling
3 Wind Engineering Benchmark
3.1 Problem Description
3.2 Source of Uncertainty
3.3 Results
4 Conclusion
References
Multi-Objective Optimal Design and Maintenance for Systems Based on Calendar Times Using MOEA/D-DE
1 Introduction
2 Methodology and Description of the Proposed Model
2.1 Extracting Availability and Economic Cost from Functionability Profiles
2.2 Multi-Objective Optimization Approach
2.3 Building Functionability Profiles
3 The Application Case
4 Results and Discussion
5 Conclusions
References
Multi-objective Robustness Analysis of the Polymer Extrusion Process
1 Introduction
2 Robustness in Polymer Extrusion
2.1 Extrusion Process
2.2 Robustness Methodology
2.3 Multi-objective Optimization with Robustness
3 Results and Discussion
4 Conclusion
References
Quantification of Operational and Geometrical Uncertainties of a 1.5-Stage Axial Compressor with Cavity Leakage Flows
1 Motivation and Test Case Description
1.1 Geometry and Operating Regime
1.2 Uncertainty Definition
Correlated Fields at the Main Inlet
Secondary Inlets
Rotor Blade Tip Gap
2 Uncertainty Quantification Method
2.1 Scaled Sensitivity Derivatives
3 Simulation Setup and Computational Cost
4 Results and Discussion
4.1 Non-deterministic Performance Curve
4.2 Scaled Sensitivity Derivatives
5 Conclusions
References
Can Uncertainty Propagation Solve the Mysterious Case of Snoopy?
1 Introduction
2 Background
3 Methodology
3.1 Dynamics Modelling
3.2 Using the TDA Structure to Solve ODE
3.3 Performing Numerical Analysis
3.4 Propagator Implementation and Validation
3.5 Monte-Carlo Estimation
4 Results and Discussion
4.1 Performing Numerical Analysis on the Trajectory of Snoopy
4.2 Computing Snoopy's Trajectory
4.3 Estimating the Probability of Snoopy's Presence
5 Conclusions and Future Work
References
Part II Imprecise Probability, Theory and Applications (IP)
Robust Particle Filter for Space Navigation Under EpistemicUncertainty
1 Introduction
2 Filtering Under Epistemic Uncertainty
2.1 Imprecise Formulation
2.2 Expectation Estimator
2.3 Bound Estimator
3 Test Case
3.1 Initial State Uncertainty
3.2 Observation Model and Errors
3.3 Results
4 Conclusions
References
Computing Bounds for Imprecise Continuous-Time Markov Chains Using Normal Cones
1 Introduction
2 Imprecise Markov Chains in Continuous Time
2.1 Imprecise Distributions over States
2.2 Imprecise Transition Rate Matrices
2.3 Distributions at Time t
3 Numerical Methods for Finding Lower Expectations
3.1 Lower Expectation and Transition Operators as Linear Programming Problems
3.2 Computational Approaches to Estimating Lower Expectation Functionals
4 Normal Cones of Imprecise Q-Operators
5 Norms of Q-Matrices
6 Numerical Methods for CTIMC Bounds Calculation
6.1 Matrix Exponential Method
6.2 Checking Applicability of the Matrix Exponential Method
6.3 Checking the Normal Cone Inclusion
6.4 Approximate Matrix Exponential Method
7 Error Estimation
7.1 General Error Bounds
7.2 Error Estimation for a Single Step
7.3 Error Estimation for the Uniform Grid
8 Algorithm and Examples
8.1 Parts of the Algorithm
8.2 Examples
9 Concluding Remarks
References
Simultaneous Sampling for Robust Markov Chain Monte Carlo Inference
1 Introduction
2 Markov Chain Monte Carlo
3 Simultaneous Sampling
4 Markov Chain Monte Carlo for Imprecise Models
5 Practical Implementation
6 Linear Representation for Exponential Families
7 Computer Representation of the Credal Sets
8 Credal Set Merging
9 Discussion
Reference
Computing Expected Hitting Times for Imprecise Markov Chains
1 Introduction
2 Existence of Solutions
3 A Computational Method
4 Complexity Analysis
References
Part III Robust and Reliability-Based Design Optimisation in Aerospace Engineering (RBDO)
Multi-Objective Robust Trajectory Optimization of Multi-Asteroid Fly-By Under Epistemic Uncertainty
1 Introduction
2 Problem Formulation
3 Lower Expectation
3.1 Minimizing the Expectation
3.2 Estimating the Expectation
4 Multi-Objective Optimization
4.1 Control Mapping for Dimensionality Reduction
Deterministic Control Map
Max-Min Control Map
Min-Max Control Map
4.2 Threshold Mapping
5 Asteroid Tour Test Case
6 Results
6.1 Control Map and Threshold Map
6.2 Lower Expectation
6.3 Expectation and Sampling Methods
6.4 Execution Times
7 Conclusions
References
Reliability-Based Robust Design Optimization of a Jet Engine Nacelle
1 Introduction
2 Definition of Aeronautical Optimization Under Uncertainties
2.1 Nacelle Acoustic Liner and Manufacturing Tolerances
2.2 Nacelle Acoustic Liner FEM Model
3 Adaptive Sparse Polynomial Chaos for Reliability Problems
3.1 Basic Formulation of Adaptive PCE
3.2 Adaptive Sparse Polynomial Chaos Expansion
3.3 Application of Adaptive PCE to Reliability-Based Optimization
4 Reliability-Based Optimization of the Engine Nacelle
4.1 Optimization Platform
4.2 Optimization Results
5 Conclusion
References
Bayesian Optimization for Robust Solutions Under Uncertain Input
1 Introduction
2 Literature Review
3 Problem Definition
4 Methodology
4.1 Gaussian Process
4.2 Robust Bayesian Optimization
Direct Robustness Approximation
Robust Knowledge Gradient
4.3 Stochastic Kriging
5 Experiments
5.1 Benchmark Problems
Test Functions
Experimental Setup
5.2 Results
Latin Hypercube Sampling
Stochastic Kriging
Uncontrollable Input
6 Conclusions
References
Optimization Under Uncertainty of Shock Control Bumps for Transonic Wings
1 Introduction
2 Gradient-Based Robust Design Framework
2.1 Motivation
2.2 Surrogate-Based Uncertainty Quantification
2.3 Obtaining the Gradients of the Statistics
2.4 Optimization Architecture
2.5 Application to Analytical Test Function
3 Application to the Robust Design of Shock Control Bumps: Problem Definition
3.1 Test Case
3.2 Numerical Model
3.3 Parametrization of Shock Control Bumps
3.4 Optimization Formulations
4 Results
4.1 Single-Point (Deterministic) Results
4.2 Uncertainty Quantification
4.3 Robust Results
5 Conclusions
References
Multi-Objective Design Optimisation of an Airfoil with Geometrical Uncertainties Leveraging Multi-Fidelity Gaussian Process Regression
1 Introduction
2 Design Optimisation Problem of Airfoil
3 Solvers
4 Multi-Fidelity Gaussian Process Regression
5 Uncertainty Treatment
6 Multi-Objective Optimisation Framework for Airfoil Optimisation Under Uncertainty
7 Results
8 Conclusion
References
High-Lift Devices Topology Robust Optimisation Using Machine Learning Assisted Optimisation
1 Introduction
2 Machine Learning Assisted Optimisation
2.1 Surrogate Model
2.2 Classifier
3 Quadrature Approach for Uncertainty Quantification
4 Problem Formulation
4.1 Optimisation Design Variables
4.2 High-Lift Devices Robust Optimisation Problem
Original Objective Function
Artificial Objective Function
5 Optimisation Setup
6 Results
7 Conclusions and Future Work
References
Network Resilience Optimisation of Complex Systems
1 Introduction
2 Evidence Theory as Uncertainty Framework
3 System Network Model
4 Complexity Reduction of Uncertainty Quantification
4.1 Network Decomposition
4.2 Tree-Based Exploration
4.3 Combined Method
5 Optimisation Approach
6 Resilience Framework
7 Application
8 Results
9 Conclusions
References
Gaussian Processes for CVaR Approximation in Robust Aerodynamic Shape Design
1 Introduction
2 Robust Design and CVaR Risk Function
3 Risk Function Approximation
3.1 Gaussian Processes
3.2 Training Methodology
4 Numerical Analysis Tools
5 Design Application Example
5.1 Optimisation Problem Setup
5.2 Optimisation Process and Robust Design Results
6 Conclusions
References
Part IV Uncertainty Quantification, Identification and Calibration in Aerospace Models (UQ)
Inference Methods for Gas-Surface Interaction Models: From Deterministic Approaches to Bayesian Techniques
1 Introduction
2 Plasma wind Tunnel Experiments
2.1 Heterogeneous Catalysis
2.2 Thermochemical Ablation
3 Deterministic Approaches to the Inference of Model Parameters
3.1 Heterogeneous Catalysis
3.2 Thermochemical Ablation
4 Bayesian Approaches to the Inference of Model Parameters
4.1 Bayes Theorem
4.2 Heterogeneous Catalysis
4.3 Thermochemical Ablation
5 Conclusions
References
Bayesian Adaptive Selection Under Prior Ignorance
1 Introduction
2 Model
3 Posterior Computation
3.1 Selection Indicators
3.2 Regression Coefficients
4 Illustration
4.1 Synthetic Datasets
4.2 Real Data Analysis
5 Conclusion
References
A Machine-Learning Framework for Plasma-Assisted Combustion Using Principal Component Analysis and Gaussian Process Regression
1 Introduction
2 Reactor Model and Ignition Simulations
3 PCA-Based Gaussian Process Regression
4 Results
4.1 Principal Component Analysis
4.2 Combination of PCA with Gaussian Process Regression
5 Conclusion
References
Estimating Exposure Fraction from Radiation Biomarkers: A Comparison of Frequentist and Bayesian Approaches
1 Introduction
2 Methodology
3 Simulation
4 Estimation of Exposed Fraction
5 Discussion
Appendix
References
A Review of Some Recent Advancements in Non-Ideal Compressible Fluid Dynamics
1 Introduction
2 Non-Ideal Oblique Shock Waves
3 NICFD Computational Model Accuracy Assessment
4 Bayesian Inference of Fluid Model Parameters
5 Conclusions
References
Dealing with High Dimensional Inconsistent Measurements in Inverse Problems Using Surrogate Modeling: An Approach Based on Sets and Intervals
1 Introduction
2 Identification Strategy and Outlier Detection Method
3 Results
3.1 Application with the Set-Valued Inverse Method When Measurements Are in a Small Amount
3.2 Application with the Set-Valued Inverse Method When Measurements Are in a Large Amount
4 Summary
References
Stochastic Preconditioners for Domain Decomposition Methods
1 Introduction
2 Acceleration of the Schwarz Method
3 Acceleration of Schur Complement Based Methods
4 Conclusions and Perspectives
References
Index