Using Statistical Calibration for Model Verification & Validation, Diagnosis of Model Inadequacy and Improving Simulation Accuracy

July 26, 2017, 1:00 PM - 2:00 PM CDT
Presented by Dr. Mark Andrews, SmartUQ's UQ Technology Steward
Dr. Mark Andrews, SmartUQ's UQ Technology Steward, previously worked at Caterpillar in Materials Research, Reliability, and Virtual Product Development.

"Essentially, all models are wrong, but some are useful" – George E. P. Box

The use of simulations to displace physical tests has become essential in accelerating analysis and reducing the cost of research and design. However, simulation results may be substantially different from reality making accurate model calibration and validation critical to achieving desired out comes. Using a method like statistical calibration can characterize model inadequacy by combining aspects of calibration and validation. Statistical calibration can reduce design cycle time and costs by ensuring that simulations are as close to reality as possible and by quantifying how close that really is.

Using case studies, this webinar will walk through the SmartUQ software showing the step-by-step process of statistical calibration and quantifying the uncertainty of the calibration parameters.

SmartUQ Demonstration Webinar: Introduction to SmartUQ version 4.1

August 09, 2017, 1:00 PM - 2:00 PM CDT
Presented by Kevin O'Flaherty, SmartUQ Application Engineer

Uncertainty Quantification may be useful but, like much of statistics, it can seem complicated. This webinar will walk through the requirements and workflow of a typical uncertainty quantification analysis in SmartUQ. Using a simulation based engineering design problem as a case study, we will cover the basics of system selection, design of experiments and sampling, surrogate model/emulator generation, and advanced analytics techniques like uncertainty propagation and sensitivity analysis. We will also outline a few more complex use cases which apply our latest features to solving engineering challenges.

Uncertainty Quantification and Management in Digital Engineering

August 23, 2017, 1:00 PM - 2:00 PM CDT
Presented by Dr. Ed Kraft, SmartUQ technical advisor
He has over 48 years of experience in testing and evaluation in both the private and public sector. He is one of the initiators and principal architects for the Air Force’s Digital Thread/Digital Twin initiative and a strong advocate for the application of Uncertainty Quantification in developing the digital authoritative truth source in support of decision making. He is a distinguished alumnus of both the University of Cincinnati and the University of Tennessee Space Institute, where he received his degrees in aerospace engineering. He is a fellow of the American Institute of Astronautics and Aeronautics (AIAA) and an Arnold Engineering Development Complex fellow.

This webinar illustrates applications of Uncertainty Quantification (UQ) techniques that will be necessary for full implementation of model-based digital engineering. The role of UQ techniques and software, such as SmartUQ, in end-to-end system models, reduced order modeling and probabilistic analysis, the creation of an authoritative digital surrogate “truth source”, and the creation of uncertainty budgets will be discussed.

Using the Department of Defense lifecycle framework for acquisition, operation, and sustainment of systems, approaches will be presented for the continuing collection and application of UQ knowledge over each stage to reduce uncertainties systematically and provide program decision makers with a probabilistic assessment of performance, risk, and costs essential to critical decisions.