Introduction to Uncertainty Quantification and Sensitivity Analysis in Engineering

Measuring the level of uncertainty in engineering design and analysis is crucial for ensuring that a system is safe, reliable, and resilient. Modern statistical analysis and machine learning have proven to be invaluable tools in this regard. This talk will explore the significance of uncertainty quantification (UQ) and sensitivity analysis (SA) for engineering design, optimization, and analysis, and how these tools can be aided by the latest advancements in statistical analysis and machine learning. The talk will introduce two distinct approaches to uncertainty UQ and SA - probabilistic and non-probabilistic, to define aleatoric and epistemic uncertainties, respectively. Two advanced statistical learning techniques will be discussed to solve uncertainty propagation problems, namely, polynomial chaos expansion and gaussian process regression. To illustrate the practical benefits of UQ and SA, the talk will conclude with real-world examples from civil and aerospace engineering.

About Pramudita Satria Palar, Ph.D.

Dr. Palar is currently working as Assistant Professor at the Faculty of Mechanical and Aerospace Engineering, Bandung Institute of Technology (Institut Teknologi Bandung - ITB), Indonesia. Prior to his current position, he was a Research Fellow at the Institute of Fluid Science, Tohoku University, Japan. During his postdoctoral position, he also visited Leiden Institute of Advanced Computer Science at Leiden University as a visiting researcher.

Latest Research Collaboration

Currently, Dr. Palar is conducting research together with Prasanti W. Sarli, PhD, where Gaussian Process Regression is applied to solve several structural problems, including predicting fragility curves for large scale and optimizing damper placement to achieve cost-efficiency. This collaboration has been ongoing since 2021.

Further info about Dr. Palar: https://sites.google.com/view/pramudi...