Quantification of Uncertainty: Methods, Applications, Challenges

The lecture deals with the ability of quantifying uncertainty in technical systems. Students should be able to describe, quantify and evaluate uncertainty in real technical systems.

Content:

1. to recognize the different types of uncertainty in real technical systems and to independently develop an approach to quantify and evaluate them.

2. conduct a sensitivity analysis to determine the key parameters of a model using methods such as Sobol's method and Morris screening.

3. quantify the uncertainty of models using Markov Chain Monte Carlo methods and interpret the results.

4. propagate the quantified uncertainty through a mode.

5. to buil quick replacement models of complex, computationally intensive models in order to make these also accessible to uncertainty quantification. Methods are such as Gaussian process regression (GPR), support vector machines (SVM), polynomial chaos expansion (PCE) are covered.

6. evaluate the probability of failure of a system based on the quantified uncertainty using subset simulation.

Availability Winter Semester, Wednesdays 9:50 – 11:30 AM
Place L101/264K
Lecture Dr.-Ing. Robert Feldmann
Exam date TBD
Coursework Master MPE und AE
WS:V2
Credit Points; 4