Dr. Biswarup Bhattacharyya, Department of Civil Engineering, IIT Hyderabad gave an online talk on Uncertainty Quantification of Dynamical Systems on 28 January.
The lecture focused on the importance of accounting for uncertainty in the analysis of engineering systems, particularly dynamical systems, where inherent variability and modelling assumptions can significantly influence system response and reliability.
Dr. Bhattacharyya began the session by introducing the fundamental sources of uncertainty in engineering problems, categorizing them into aleatory uncertainty, associated with inherent randomness, and epistemic uncertainty, arising from lack of knowledge, modelling assumptions, and measurement errors. He emphasized the need to identify and treat these uncertainties appropriately to ensure realistic prediction of system behaviour and to mitigate potential risks in engineering applications. The lecture further provided an overview of uncertainty quantification (UQ) and its relevance to both linear and nonlinear dynamical systems. Subsequently, the concept of uncertainty propagation was discussed, explaining how uncertainties in input parameters affect the response of dynamical systems. Dr. Bhattacharyya elaborated on commonly adopted approaches for uncertainty propagation, including Monte Carlo Simulation, highlighting its robustness and general applicability. He also discussed the associated computational challenges and introduced surrogate modelling techniques as an efficient alternative. The advantages of surrogate models in reducing computational cost while retaining adequate accuracy were explained using examples from dynamical systems, along with an introduction to online surrogate models for adaptive uncertainty quantification.
In the concluding part of the lecture, Dr. Bhattacharyya summarized the key advantages of surrogate-based approaches and their applicability in modern engineering problems such as structural dynamics, digital twins, and reliability analysis. The session concluded with an interactive Q&A, during which participants raised questions related to the practical implementation of UQ methods, the selection of appropriate modelling strategies, and trade-offs between accuracy and computational efficiency. Dr. Bhattacharyya addressed these queries with clear explanations and practical insights. The lecture, in conclusion, substantially strengthened the participants’ understanding of uncertainty quantification in dynamical systems and offered a clear conceptual framework along with practical insights for applying UQ methodologies in advanced engineering analysis and research.
YouTube Link: https://youtu.be/TT59Q-MKBDk
January 2026