Probabilistic and Uncertainty Quantification Methods for Model Verification & Validation (Virtual Classroom)

Topics Covered

Background

• Verification and validation
• Validation metrics
• Validation requirements
• Predictions
• Decisions

Modeling uncertain variables

• Mathematical models for uncertainty (PDF/CDF)
• Data fitting

Propagating uncertainties 

• Sampling methods 
• Analytical methods 

Formulating UQ problems 

• Solution objectives 
• Defining the model 
• Modeling random variables 
• Evaluating results 

Sensitivity analysis 

• Deterministic 
• Probabilistic 
• Global 

UQ for numerical models 

• Uncertain variables related to finite element 
• Modeling spatial and temporal variables 
• Solution approaches 

Response surface models for efficient uncertainty propagation 

• Basic principles 
• Training data bounds/# points in practice 
• Polynomial model fitting 
• Gaussian process concepts 
• Model assessment 

Bayesian statistics for uncertainty quantification and calibration

• Identification and categorization of different types of uncertainty
• Modeling/quantification of uncertainty
• Bayesian analysis
• Bayesian analysis for model calibration
• Evaluation of calibration assumptions

UQ solution strategy examples 

• Turbine blade model overview 
• Model and data uncertainties 
• Solution strategies 
• Results interpretation 

V&V case study

• V&V approach and plan
• Model documentation, phenomena identification, and hierarchies
• V&V and UQ assessments
• Uncertainty reduction
• Model validation and assessment

Erin C. DeCarlo, Ph.D. is a Research Engineer in the Mechanical Engineering Division at Southwest Research Institute. Her research focuses on understanding the impact of uncertainty on design performance and making uncertainty-informed decisions using advanced probabilistic methods. Her work in response surface modeling and Bayesian statistics addresses the fundamental challenges of the computational expense of complex, multidisciplinary simulations, and uncertainty due to limited data in order to guide activities to target and reduce uncertainties.

At Southwest Research Institute, Dr. DeCarlo leads and provides her expertise on projects that involve probabilistic analysis, uncertainty quantification and reduction, and validation of both statistical and physics-based models. She received her Ph.D. in Civil Engineering from Vanderbilt University in 2017 and recently co-led SwRI’s webinar series on “Probabilistic Analysis and Uncertainty Quantification”.
 
 
David Riha, is a Staff Engineer in the Mechanical Engineering Division at Southwest Research Institute.  His technical expertise and interests are concentrated in the area of predicting the probabilistic response and reliability of engineered systems using advanced probabilistic and uncertainty methodologies.  Since 1991, he has developed probabilistic methods and software tools including the NESSUS® probabilistic analysis software. 
 
Mr. Riha provides consulting and leads programs for applied reliability problems, model verification and validation, and uncertainty quantification for various industry and government agencies in areas such as aerospace, automotive, biomechanics, geomechanics, and weapon systems.  He also develops and presents training in the area of probabilistic analysis and design methods, uncertainty quantification, and approaches for model verification and validation. He has taught over 75 courses since 1991.  He has a B.S in aerospace engineering from the University of Texas at Austin and M.S. in mechanical engineering from the University of Texas at San Antonio.
  
Ben H. Thacker, Ph.D., P.E., Executive Director, Materials Engineering Department, Southwest Research Institute, brings over 30 years of expertise in computational mechanics, structural reliability, and computer methods development.  He has been heavily involved in the development and application of probabilistic methods and has applied probabilistic methods to geo-mechanics, biomechanics, and other transient non-linear problems. 
 
Dr. Thacker is an active member of the AIAA Non-Deterministic Technical Committee and the ASME Standards Committee on Verification and Validation.  He has instructed at the “Probabilistic Analysis and Design: Computational Methods and Applications” annual short course at the Southwest Research Institute since its inception.  He received his Ph.D. in Civil Engineering from University of Texas at Austin.