Dr Abhishek Kundu

Trosolwg

Qualifications: PhD, Swansea University, Swansea, UK, 2014. MS, Universite de Sherbrooke, Sherbrooke, Canada, 2010. BE, Jadavpur University, Kolkata, India, Mechanical Engineering, 2007. Research Interest & Expertise: Uncertainty Quantification (UQ) in computational engineering mechanics is an urgent necessity in design, calibration and performance prediction of practical engineering systems to ensure their clean and fail-safe operation. Uncertainty may arise due to a lack or limited information about system parameters, boundary conditions, geometry/configuration of the physical systems or we might be interested in physical phenomenon which are inherently random in nature. Additionally, uncertainty exists in the accumulated data or even due to the inadequate mathematical models used for describing complex physical phenomenon. Thus, the modelling and analysis of engineering systems must quantify and incorporate the sources of uncertainty into the mathematical models/simulators to capture the model variability and confidence in our predictions. This task encompasses a range of serious challenges ranging from theoretical, implementational, computational to even a practical quantitative interpretation of the results. Dr Kundu`s research focuses on developing a stochastic computational framework for total UQ in various practical engineering problems to ensure robust, optimal design of safety-critical systems, evidence-based monitoring of the health and performance of the components and risk mitigation. Dr Kundu works primarily with probabilistic uncertainty descriptors and his expertise include the forward uncertainty propagation problem, multi-level probabilistic design optimization, data assimilation via Bayesian regularization, Kalman filtering techniques, stochastic sensitivity analysis and probabilistic forecasting. The methodologies have been applied to composite and multiscale structural dynamic systems of particular relevance in civil and aerospace applications. Additionally, Dr Kundu is working on condition-based structural health monitoring within a Bayesian framework to estimate their useful life and future performance. Uncertainty analysis in environmental predictions is another area of interest to him, especially with regards to prediction of oceanic and coastal hazards from large-scale waves, storm surge, effects of uncertainties on the future envelope of wave heights and their stochastic sensitivity among others. Dr Kundu is an expert in the real-time adaptive control of vibration and acoustic emission in complex engineering problems. He has worked with acoustic and vibration isolation techniques using both passive and active (with smart materials) techniques, and combining them to obtain an optimal performance over a broad frequency range of operation. His work involved simulating the behavior of coupled multiphysics vibro-acoustic systems and their experimental studies with real-time control to demonstrate vibro-acoustic isolation methodology. He has worked with sensoriactuators or self-sensing techniques which provide a robust, compact control system design. Industrial Relvance: Civil Structures, Aerospace, Environment, Commercial Software Keywords Uncertainty Quantification, Bayesian inference, uncertainty propagation, polynomial chaos, structural dynamics, optimization, machine learning, active control Additional Keywords vibration isolation, noise control, real-time adaptive control, Parallel Computing, Functional analysis, composites, computational methods.
Mechanics, Materials and Advanced Manufacturing


ACHIEVEMENTS: - Zienkiewicz Scholarship, Swansea University, PhD studentship. - Graduate Student Bursary, Universite de Sherbrooke, M.Sc.A. studentship. - Ranked in top 1% in the GATE Examination, India, 2008. INVITED TALKS: - 2014; Uncertainty Quantification and probabilistic design optimization of multiscale engineering systems, Welsh National Research Network in Advanced Engineering and Materials, Cardiff, UK. - 2012; A reduced spectral function approach for stochastic transient dynamics, Workshop on ”Linear Algebra Aspects of Solving PDEs with Random Data”, School of Mathematics, University of Manchester, Manchester, UK. REVIEWER SERVICE: - Mechanical Systems and Signal Processing, Elsevier. - Journal of Vibration and Control, SAGE. - Applied Mathematical Modelling, Elsevier. - Journal of Computational and Applied Mathematics, Elsevier. - Research Grant Application Review: FONDECYT 2014 Initiation into research, an initiative of the Chilean National Science and Technology Commission (CONICYT - Chile). ACADEMIC COLLABORATIONS: - Institute of Risk & Uncertainty, University of Liverpool, U.K. - Technische Universität Braunschweig, Germany. - Swansea University, U.K.