Gastprofessor Drahomir Novak
Gastprofessor Drahomir Novak
Wir freuen uns, Prof. Drahomir Novak (Institute of Structural Mechanics, Brno University of Technology, Czech Republic) als Gastprofessor an der BOKU University begrüßen zu dürfen. Im Rahmen seines Aufenthalts hält Prof. Novak vier Gastvorlesungen zu probabilistischer Tragwerksanalyse, Zuverlässigkeit, nichtlinearer Modellierung und Surrogatmodellen. Nutzen Sie die Gelegenheit zum wissenschaftlichen Austausch und herzlich willkommen an der BOKU University.
Bio – Drahomir Novak
Drahomir Novak serves as Professor at the Institute of Structural Mechanics, Brno University of Technology, Czech Republic. He is a member of the Engineering Academy of the Czech Republic (elected in 2009), the Czech Society of Mechanics, the Czech Standardization Committee on Loading and Reliability, the International Association IABMAS and FRAMCOS, and is an honorary member of EUROSTRUCT.
He is an active reviewer for several international journals and a member of fib (COM8 TG 8.4 Life Cycle Cost; COM3 TG 3.3 Assessment and Evaluation Procedures for Existing Structures) as well as a member of the Scientific Committee of the International Association on Life-Cycle Civil Engineering (IALCCE).
During his professional career, Prof. Novak has held various visiting positions and lectures at institutions around the world, including Kyoto University, Japan; University of Innsbruck, Austria; BOKU University, Austria; Kasetsart University, Thailand; Northwestern University, USA; and Hohai University and Chuzhou University, China.
Prof. Novak will present four lectures as follows
| No. | Lecture Title | Time & Date | Location |
|---|---|---|---|
| 1 | Shear strength code models uncertainties and reliability of design assessment based on probabilistic simulation via Excel-FReET interface | 16:00–18:00, 06.10.2026 | Exner Haus 01/12 |
| 2 | Computational simulation of shear strength of steel and GFRP-reinforced concrete members: Material parameters identification and nonlinear fracture mechanics modelling using ATENA software | 16:00–18:00, 08.10.2026 | Exner Haus 01/12 |
| 3 | How to eliminate high computational burden I: Surrogate modelling based on artificial neural networks and polynomial chaos expansion | 16:00–18:00, 27.10.2026 | Exner Haus 01/12 |
| 4 | How to eliminate high computational burden II: Semi-probabilistic approaches and safety formats – ECoV techniques | 16:00–18:00, 28.10.2026 | Exner Haus 01/12 |
Lecture 1 – Shear strength code models uncertainties and reliability of design assessment based on probabilistic simulation via Excel-FReET interface
Shear strength probabilistic assessment of concrete members with steel and GFRP reinforcement is performed. The shear strength is analyzed by modelling with low-fidelity models – analytical formulas based on two main approaches to predict the shear strength of reinforced concrete beams with and without shear reinforcement: the modified compression field theory and the truss model.
Based on a comparison of these low-fidelity analytical models and experimental data, model uncertainties can be evaluated. The aim of the analysis performed is to verify the existing code analytical formulas for shear strength calculation using stochastic models, to perform uncertainty propagation, sensitivity analysis and model uncertainty assessment. The code models of EN 1992-1-1, ACI 318 and fib Model Code 2010 are examined with respect to uncertainties involved and the reliability of the design value determination.
6.10.2026 16.00-18.00, Exner Haus 01/12
Lecture 2 - Computational simulation of shear strength of steel and GFRP-reinforced concrete members: Material parameters identification and nonlinear fracture mechanics modelling using ATENA software
The aim is to demonstrate potential and discuss important aspects of computational modelling of T-shaped beams reinforced by longitudinal steel and glass fiber reinforced polymer (GFRP) bars. The computational model is based on the finite element method and employs a non-linear fracture-plastic material model.
Particular attention is given to bond between the concrete and reinforcement, because proper bond modelling is extremely important when using GFRP reinforcement. The results of experimental investigation are compared to results of numerical modelling in the form of load-deflection curves, deflection of beams and crack patterns. A material parameters identification approach based on artificial neural networks will be presented.
8.10.2026 16.00-18.00, Exner Haus 01/12
Lecture 3 - How to eliminate high computational burden I: Surrogate modelling based on artificial neural networks and polynomial chaos expansion
This topic focuses on two surrogate modelling techniques and their potential for stochastic analysis of engineering structures. The first technique is polynomial chaos expansion (PCE), a representative from the field of uncertainty quantification (UQ). PCE is often employed in applications with a limited number of samples since it achieves high accuracy and allows powerful post-processing without additional cost. This can yield statistical moments or Sobol sensitivity coefficients, which is a frequent requirement of stochastic analysis.
The second technique is based on artificial neural networks (ANN), a well-known machine-learning model used in many mathematical and engineering fields. ANN is a powerful signal-processing model that can efficiently handle large numbers of samples, which determines its common use in big-data analysis. Both models have great potential for use in the field of stochastic structural analysis, each with its specific advantages and disadvantages.
27.10.2026 16.00-18.00, Exner Haus 01/12
Lecture 4 - How to eliminate high computational burden II: Semi-probabilistic approaches and safety formats – ECoV techniques
It is highly time consuming to perform fully probabilistic analysis of large non-linear models with many stochastic input variables. A solution can be represented by a semi-probabilistic approach focused on determination of the design value of resistance, which is able to greatly reduce the number of needed nonlinear calculations.
The key ingredient in semi-probabilistic design and assessment of structures is an estimation of the coefficient of variation (ECoV). In recent years, development of ECoV methods has been an area of interest for many researchers and engineers, and there are several known approaches.
28.10.2026 16.00-18.00, Exner Haus 01/12