Materials Discovery and Design (eBook)

By Means of Data Science and Optimal Learning

Turab Lookman, Stephan Eidenbenz, Frank Alexander, Cris Barnes (Herausgeber)

eBook Download: PDF

2018 | 1st ed. 2018
XVI, 256 Seiten
Springer International Publishing (Verlag)
978-3-319-99465-9 (ISBN)

Infos

This book addresses the current status, challenges and future directions of data-driven materials discovery and design. It presents the analysis and learning from data as a key theme in many science and cyber related applications. The challenging open questions as well as future directions in the application of data science to materials problems are sketched. Computational and experimental facilities today generate vast amounts of data at an unprecedented rate. The book gives guidance to discover new knowledge that enables materials innovation to address grand challenges in energy, environment and security, the clearer link needed between the data from these facilities and the theory and underlying science. The role of inference and optimization methods in distilling the data and constraining predictions using insights and results from theory is key to achieving the desired goals of real time analysis and feedback. Thus, the importance of this book lies in emphasizing that the full value of knowledge driven discovery using data can only be realized by integrating statistical and information sciences with materials science, which is increasingly dependent on high throughput and large scale computational and experimental data gathering efforts. This is especially the case as we enter a new era of big data in materials science with the planning of future experimental facilities such as the Linac Coherent Light Source at Stanford (LCLS-II), the European X-ray Free Electron Laser (EXFEL) and MaRIE (Matter Radiation in Extremes), the signature concept facility from Los Alamos National Laboratory. These facilities are expected to generate hundreds of terabytes to several petabytes of in situ spatially and temporally resolved data per sample.  The questions that then arise include how we can learn from the data to accelerate the processing and analysis of reconstructed microstructure, rapidly map spatially resolved properties from high throughput data, devise diagnostics for pattern detection, and guide experiments towards desired targeted properties. The authors are an interdisciplinary group of leading experts who bring the excitement of the nascent and rapidly emerging field of materials informatics to the reader. 

Inhalt

Preface 6
Contents 9
Contributors 14
1 Dimensions, Bits, and Wows in Accelerating Materials Discovery 16
1.1 Introduction 16
1.2 Creativity and Discovery 18
1.3 Discovering Dimensions 20
1.4 Infotaxis 21
1.5 Pursuit of Bayesian Surprise 23
1.6 Conclusion 26
References 26
2 Is Automated Materials Design and Discovery Possible? 30
2.1 Model Determination in Materials Science 31
2.1.1 The Status Quo 31
2.1.2 The Goal 31
2.2 Identification of the Research and Issues 32
2.2.1 Reducing the Degrees of Freedom in Model Determination 32
2.2.2 OUQ and mystic 34
2.3 Introduction to Uncertainty Quantification 36
2.3.1 The UQ Problem 36
2.4 Generalizations and Comparisons 39
2.4.1 Prediction, Extrapolation, Verification and Validation 39
2.4.2 Comparisons with Other UQ Methods 40
2.5 Optimal Uncertainty Quantification 42
2.5.1 First Description 43
2.6 The Optimal UQ Problem 46
2.6.1 From Theory to Computation 46
2.7 Optimal Design 51
2.7.1 The Optimal UQ Loop 51
2.8 Model-Form Uncertainty 55
2.8.1 Optimal UQ and Model Error 55
2.8.2 Game-Theoretic Formulation and Model Error 56
2.9 Design and Decision-Making Under Uncertainty 57
2.9.1 Optimal UQ for Vulnerability Identification 57
2.9.2 Data Collection for Design Optimization 58
2.10 A Software Framework for Optimization and UQ in Reduced Search Space 59
2.10.1 Optimization and UQ 59
2.10.2 A Highly-Configurable Optimization Framework 60
2.10.3 Reduction of Search Space 61
2.10.4 New Massively-Parallel Optimization Algorithms 64
2.10.5 Probability and Uncertainty Tooklit 65
2.11 Scalability 68
2.11.1 Scalability Through Asynchronous Parallel Computing 68
References 69
3 Importance of Feature Selection in Machine Learning and Adaptive Design for Materials 74
3.1 Introduction 75
3.2 Computational Details 77
3.2.1 Density Functional Theory 77
3.2.2 Machine Learning 78
3.2.3 Design 78
3.3 Results 79
3.4 Discussion 88
3.5 Summary 91
References 92
4 Bayesian Approaches to Uncertainty Quantification and Structure Refinement from X-Ray Diffraction 95
4.1 Introduction 95
4.2 Classical Methods of Structure Refinement 97
4.2.1 Classical Single Peak Fitting 97
4.2.2 The Rietveld Method 98
4.2.3 Frequentist Inference and Its Limitations 100
4.3 Bayesian Inference 101
4.3.1 Sampling Algorithms 103
4.4 Application of Bayesian Inference to Single Peak Fitting: A Case Study in Ferroelectric Materials 104
4.4.1 Methods 106
4.4.2 Prediction Intervals 107
4.5 Application of Bayesian Inference to Full Pattern Crystallographic Structure Refinement: A Case Study 108
4.5.1 Data Collection and the Rietveld Analysis 109
4.5.2 Importance of Modelling the Variance and Correlation of Residuals 110
4.5.3 Bayesian Analysis of the NIST Silicon Standard 111
4.5.4 Comparison of the Structure Refinement Approaches 111
4.5.5 Programs 113
4.6 Conclusion 114
References 115
5 Deep Data Analytics in Structural and Functional Imaging of Nanoscale Materials 117
5.1 Introduction 118
5.2 Case Study 1. Interplay Between Different Structural Order Parameters in Molecular Self-assembly 120
5.2.1 Model System and Problem Overview 120
5.2.2 How to Find Positions of All Molecules in the Image? 121
5.2.3 Identifying Molecular Structural Degrees of Freedom via Computer Vision 122
5.2.4 Application to Real Experimental Data: From Imaging to Physics and Chemistry 126
5.3 Case Study 2. Role of Lattice Strain in Formation of Electron Scattering Patterns in Graphene 129
5.3.1 Model System and Problem Overview 129
5.3.2 How to Extract Structural and Electronic Degrees of Freedom Directly from an Image? 130
5.3.3 Direct Data Mining of Structure and Electronic Degrees of Freedom in Graphene 131
5.4 Case Study 3. Correlative Analysis in Multi-mode Imaging of Strongly Correlated Electron Systems 135
5.4.1 Model System and Problem Overview 135
5.4.2 How to Obtain Physically Meaningful Endmembers from Hyperspectral Tunneling Conductance Data? 136
5.5 Overall Conclusion and Outlook 140
References 141
6 Data Challenges of In Situ X-Ray Tomography for Materials Discovery and Characterization 143
6.1 Introduction 144
6.2 In Situ Techniques 147
6.3 Experimental Rates 150
6.4 Experimental and Image Acquisition 155
6.5 Reconstruction 159
6.6 Visualization 160
6.7 Segmentation 162
6.8 Modeling 165
6.9 In Situ Data 166
6.10 Analyze and Advanced Processing 167
6.11 Conclusions 170
References 172
7 Overview of High-Energy X-Ray Diffraction Microscopy (HEDM) for Mesoscale Material Characterization in Three-Dimensions 180
7.1 Introduction 180
7.1.1 The Mesoscale 181
7.1.2 Imaging Techniques 182
7.2 Brief Background on Scattering Physics 184
7.2.1 Scattering by an Atom 185
7.2.2 Crystallographic Planes 187
7.2.3 Diffraction by a Small Crystal 188
7.2.4 Electron Density 190
7.3 High-Energy X-Ray Diffraction Microscopy (HEDM) 191
7.3.1 Experimental Setup 191
7.3.2 Data Analysis 192
7.4 Microstructure Representation 194
7.5 Example Applications 196
7.5.1 Tracking Plastic Deformation in Polycrystalline Copper Using Nf-HEDM 196
7.5.2 Combined nf- and ff-HEDM for Tracking Inter-granular Stress in Titanium Alloy 199
7.5.3 Tracking Lattice Rotation Change in Interstitial-Free (IF) Steel Using HEDM 200
7.5.4 Grain-Scale Residual Strain (Stress) Determination in Ti-7Al Using HEDM 202
7.5.5 In-Situ ff-HEDM Characterization of Stress-Induced Phase Transformation in Nickel-Titanium Shape Memory Alloys (SMA) 203
7.5.6 HEDM Application to Nuclear Fuels 204
7.5.7 Utilizing HEDM to Characterize Additively Manufactured 316L Stainless Steel 205
7.6 Conclusions and Perspectives 207
7.6.1 Establishing Processing-Structure- Property-Performance Relationships 209
References 211
8 Bragg Coherent Diffraction Imaging Techniques at 3rd and 4th Generation Light Sources 215
8.1 Introduction 216
8.2 BCDI Methods at Light Sources 223
8.3 Big Data Challenges in BCDI 224
8.4 Conclusions 226
References 226
9 Automatic Tuning and Control for Advanced Light Sources 228
9.1 Introduction 229
9.1.1 Beam Dynamics 231
9.1.2 RF Acceleration 233
9.1.3 Bunch Compression 234
9.1.4 RF Systems 235
9.1.5 Need for Feedback Control 237
9.1.6 Standart Proportional Integral (PI) Control for RF Cavity 238
9.2 Advanced Control and Tuning Topics 243
9.3 Introduction to Extremum Seeking Control 244
9.3.1 Physical Motivation 245
9.3.2 General ES Scheme 247
9.3.3 ES for RF Beam Loading Compensation 249
9.3.4 ES for Magnet Tuning 251
9.3.5 ES for Electron Bunch Longitudinal Phase Space Prediction 253
9.3.6 ES for Phase Space Tuning 257
9.4 Conclusions 260
References 260
Index 263

Produkt-Details

• Erscheint lt. Verlag: 22.9.2018
• Reihe/Serie: Springer Series in Materials Science
• Reihe/Serie: Springer Series in Materials Science
• Zusatzinfo: XVI, 256 p. 98 illus., 88 illus. in color.
• Verlagsort: Cham
• Sprache: englisch
• Themenwelt: Mathematik / Informatik ► Informatik ► Datenbanken
• Themenwelt: Mathematik / Informatik ► Mathematik
• Themenwelt: Naturwissenschaften ► Physik / Astronomie ► Allgemeines / Lexika
• Themenwelt: Naturwissenschaften ► Physik / Astronomie ► Theoretische Physik
• Themenwelt: Technik ► Maschinenbau
• Schlagworte: Automated materials design • Combinatorial materials science • Data-driven materials design • Data-driven materials science • Data Optimization Analysis for Facilities • Functionality-driven materials design • Large data sets and materials
• ISBN-10: 3-319-99465-4 / 3319994654
• ISBN-13: 978-3-319-99465-9 / 9783319994659