Allan P. Engsig-Karup
Associate Professor in Scientific Computing, M.Sc.(Eng.), Ph.D.
Faculty Member of Center For Energy Resources Engineering (CERE), DTU
Faculty Member of The Danish Center for Applied Mathematics and Mechanics (DCAMM)

Scientific Computing DTU Compute, Technical University of Denmark Matematiktorvet, Building 303b/108 2800 Kgs.-Lyngby Tel: (+45) 45 25 30 73 Fax: (+45) 45 88 26 73 Skype: a.p.engsigkarup apek @ dtu.dk http://www2.compute.dtu.dk/~apek/

Research interests linked to applications of basic and applied mathematics research for Advanced Simulations and innovative use of Modern Computing Technologies and contribute to sustainability objectives

• Numerical Analysis, Efficient algorithms and Fast solvers.
• High Permance and Heterogeneous Computing on modern many-core hardware (fx. GPU Accelerators).
• Advanced Numerical methods for Simulation, solving ordinary, partial and stochastic differential equations.
• Efficient Uncertainty Quantification Methods.
• Model Order Reduction (MOR) / Reduced Order Modelling (ROM) techniques.
• Mathematical optimization and Learning Algorithms within the emerging area of Scientific Machine Learning for Data-Driven Applications.
• Deep Learning using neural networks and their applications.
• Massively parallel algorithms for massive-scale super computing.
• Scientific and Engineering application (targeting new proof-of-concepts, knowledge generation and scientific simulation), aimed at fundamental research related to sustainable energy research and Center for Energy Resources Engineering (CERE).
• Digital Innovation and Technologies within education and with companies.
• Quantum Computing.

Publications, presentations, etc. or check out my profile at ResearchGate.

Internships with companies and research institutions

• Feel free to contact me if you are interested in internships with research institutions outside Denmark or national/international companies in need of technical skills. Very good opportunities for talented students.

University Teaching

• 02601 Introduction to numerical algorithms
• 02614 High-Performance Computing (contains introduction to programming of many-core GPUs using CUDA / OpenCL)
• 02623 The Finite Element Method for Partial Differential Equations (contains introduction to open source software FENICS and commercial software COMSOL)
  Check out example posters from participants: Posters.
• 02687 Scientific Computing for Ordinary and Partial Differential Equations
• 02689 Advanced Numerical Methods for Differential Equations
  Check out example posters from participants: Posters.
• 02977 Scientific Machine Learning
• Special course in Scientific Machine Learning (Explore data-driven techniques and apply ideas that are scientifically grounded)

Course catalogue at DTU: Catalogue

Organized PhD schools, Workshops and Seminars

2009, DGFEM2009 with Prof. Jan S. Hesthaven.

2010, GPU2010 with Prof. Hendrik Lensch, Robert Strodzka and Timothy Lanfear.

2011, GPU2011with Assoc. Prof. Tim Warburton.

2011, ITSOL2011 with Prof. C. T. Kelley.

2011, Workshop on GPU Programming in Python with PostDoc Andreas Klockner.

2012, DGFEM2012 with Prof. Jan S. Hesthaven.

2013, Seminar on Modern Scientific Computing Trends with Assoc. Prof. Luke Olson and Prof. Xing Cai.

2014, ITSOL2014 with Prof. C. T. Kelley.

2015, Seminar on Modern Scientific Computing Trends with Directeur de Recherche CNRS Olivier Le Maitre

2015-2016 Inspirational seminar series on PDE-constrained optimization

• October, Prof. Anton Evgrafov
• November, Prof. Bijan Mohammedi.
• December, Prof. Roland Herzog.
• January, Prof. Michael Pedersen.

2016, Workshop on How To Scale Scientific Applications from Laptops to Supercomputers with PETSc with Karl Rupp.

2018, Numerical Methods for Uncertainty Quantification UQ2018 with Directeur de Recherche CNRS Olivier Le Maitre.

2020, Model Order Reduction Summer School (MORSS) 2020 organized by EPFL (Ecole polytechnique federale Lausanne), DTU - Technical University of Denmark, a Eindhoven University of Technology (EuroTech Universities Alliance)

2022, PhD school on Scientific Machine Learning with Dr. Christopher Rackauckas, MIT, US.

2022-2023 Inspirational seminar series on Mathematics of Data Science

At the Scientific Computing Section of DTU Compute, we also host Scientific Computing Seminars.

Research and Innovation projects

• Efficient Added Mass Calculations for Large and Complex Floating Offshore Structures (2023)
  PhD student, Jens Haa¥kon Visbech Christesene( COWIFONDEN / (DTU Compue, principal investigator, main supervisor)
• New advanced simulation techniques for wave energy converters (2022-2025)
  PhD student, Jens Haa¥kon Visbech Christesenn (DTU Compute / DTU Mechanics, main supervisor)
• Interactive auditory virtual reality using reduced order modeling (2022)
  PhD student, Hermes Sampedro Llopis (DTU Electro / DTU compute / Ecophon / RambÃll / Virsabi / Treble technologies / Danish Sound Cluster, co-supervisor)
  RB-SEM virtual acoustics (proof-of-concept)
  PhD thesis: Reduced basis modelling for accelerated numerical room acoustics simulations
• Cloud-based architectural acoustics simulations (CBAAS) (2022)
  PhD student, Anders Dalsgaard Melander (DTU Compute / DTU Electro / Treble Technologies / Trivium / FF4EUROHPC, main supervisor)
• High-Performance Computing for Computational Fluid Dynamics for Free Surface Flows (2021-2023)
  PhD student, Anders Dalsgaard Melander (DTU Compute, main supervisor)
• Accurate and Fast Auditory Virtual Reality Implementation (2019-2022)
  PhD student, Nikolas Borrel-Jensen (DTU Electro / DTU Compute / Eindhoven University of Technology / EuroTech, co-supervisor)
  PhD thesis: Accelerated methods for computing acoustic sound fields in dynamic virtual environments with moving sources
  New position: Software developer, Meta, Redmond
• Robust Surrogate Modelling for Antenna Design Applications (2020-2023)
  Industrial PhD student, Sabine Fie Hansen (DTU Compute / TICRA / EPFL, main supervisor)
• Estimation, Simulation and Control for Optimal Containment of COVID-19 (2020)
  PostDoc, Kristian Meyer (DTU Compute / AAU / Novo Nordisk foundation, main supervisor)
  Research Assistant, Anders Dalsgaard Melander (DTU Compute / AAU / Novo Nordisk foundation, main supervisor)
• Acoustic Virtual Reality for Architectural Design (2019-2022)
  Industrial PhD student, Hermes Sampedro Llopis (DTU Electro / DTU Compute / EPFL / Rambø¸ll / Ecophon, co-supervisor)
  Talk given at the Model Order Reduction Summer School (MORSS) 2020: RB models for real-time wave-based virtual acoustics simulations
• Extension of a Fast Potential Flow Solver to Fully-Nonlinear Wave Loading on Offshore Structures (2017-2020)
  PhD student, Jacob Bjarke Hansen Hicks (DTU Mechanics / DTU Compute / DHRTC, co-supervisor)
  PhD Thesis: Development of a high-order potential flow solver for nonlinear wave-structure interaction
• Architecture acoustics: an improved design process using integrated hybrid room acoustic simulations (2016-2019)
  Industrial PhD student, Finnur Pind Jorgensson (DTU Electro / DTU Compute / Henning Larsen Architects / Innovation Fund Denmark, co-supervisor)
  Won Danish Sound Research Talent Award 2018, Link to pitch, Link to motivation for the research, at event organized by the Danish Sound Network.
  PhD thesis: Wave-based Virtual Acoustics
  New position: PostDoc at University of Iceland and CEO of the startup Treble Technologies, Iceland
• Multi-fIdelity Decision making tools for Wave Energy SysTems (MIDWEST) (2015-2018)
  PhD Student, Umberto Bosi (INRIA / RISE / DTU Compute, EU funded, co-supervisor)
  PhD thesis: A unified spectral/hp element depth-integrated Boussinesq model for nonlinear wave-floating body interaction
• Optimising Oil Production by Novel Technology Integration (OPTION) (2014-2018)
  Reduced Order Modelling for Efficient Reservoir-Well Simulations
  PostDoc, Hani Akbari (DTU Compute/Lloyds-Ods/ Innovation Fund Denmark project, Main supervisor)
  Multi-scale Reservoir-Well Simulations (DTU Compute/Lloyds-Ods/Innovation Fund Denmark, main supervisor, 4 months project)
  PhD student, Casper Schytte Hemmingsen
  Multi-scale Reservoir-Well Simulations (DTU Compute/Lloyds-Ods/Innovation Fund Denmark, main supervisor, 3 months project)
  PostDoc, Nikolas Karvounis
• Added Resistance of Ultra-Slow Ships (2013-2016)
  PhD Student, Stavros Kontos (DTU Mechanics/FORCE Technology/DTU Compute, EU-funded, co-supervisor)
  PhD thesis: Robust Numerical Methods for Nonlinear Wave-Structure Interaction in a Moving Frame of Reference
• Efficient Large-Scale Reservoir Simulation on Modern Many-Core Hardware (2013-2016)
  Industrial PhD Student, Max la Cour Christensen (DTU Compute/LLoyds-ods/Tie-Line Technology/CERE, main supervisor)
  PhD thesis: Multilevel Methods for Reservoir Simulation (Link to PDF)
  Video of PhD Defense is found here: LINK
  New position: PostDoc at the Danish Hydrocarbon Research and Technology Centre (DHRTC).
• Uncertainty Quantification for Engineering Applications (2011-2014)
  PhD Student, Daniele Bigoni (DTU Compute, main supervisor)
  PhD thesis: Uncertainty Quantification with Application to Engineering Problems
  Position after PhD: PostDoc at Department of Aeronautics and Astronautics, Massachusetts Institute of Technology. Cambridge, USA
• Real-time ship hydrodynamics (2012-2013)
  PostDoc, Ole Lindberg (DTU Mechanics/DTU Compute/FORCE Technology/The Danish Maritime Cluster)
• Desktop Scientific Computing on Consumer Graphics Cards (2010-2013)
  PhD Student, Stefan Glimberg (DTU Compute, main supervisor)
  PhD thesis: Designing Scientic Software for Heterogeneous Computing With application to large-scale water wave simulations
  Position after PhD: Senior Consultant, Lloyds Register Consulting, Hellerup, Denmark.
• Modelling Nonlinear Wave Interaction with Floating Ocean Energy Devices (2010-2013)
  PhD Student, Torben Christiansen (DTU Mechanics/DTU Compute, co-supervisor)
• Multiscale simulations of wave forces on ocean energy devices (2008-2010)
  PhD student, Ole Lindberg (DTU Mechanics/DTU Informatics, co-supervisor)
  PhD thesis: Multiscale Simulation of Breaking Wave Impacts
  Position after PhD: Software developer at FORCE Technology, Kongens Lyngby, Denmark.
• Direct solutions to the exact Laplace problem (2006-2008, PostDoc, DTU Mechanics)
• Unstructured Nodal DG-FEM Solution of high-order Boussinesq-type equations (2003-2006, PhD Thesis, DTU Mechanics)

Active and recent research areas in new technologies, computational mathematics and their applications

All highlights are build from scratch with an aim to deliver state-of-the-art approaches using advanced numerical methods as a part of research effort in my group.

	Research in paradigm shifts in scientific computing using modern parallel programming paradigms and emerging many-core and heterogeneous architectures ranging from work desktops to the largest super-computing clusters in the world.
	Novel algorithms and high-performance computing for fast (possibly real-time) calculations to pave the way for novel marine and naval hydrodynamics calculations.
	Novel and efficient spectral (high-order) and adaptive algorithms for uncertainty quantification of problems with high dimensionality, data science, machine learning and stochastic simulations.
	Massively parallel and scalable algorithms such as multigrid and multi-level algorithms for scientific engineering applications, e.g. reservoir simulation and hydrodynamics.
	Novel application proof-of-concepts for engineering analysis, e.g. offshore engineering, computational fluid dynamics, room acoustics and vibro-acoustics.
	Research in efficient and robust high-order unstructured numerical methods and time-dependent models.

Scientific Software Engineering

Drivers of the research are scientific software for scientific investigation.

• Unstructured High-Order Boussinesq-type model - a flexible coastal engineering research tool for simulation of nonlinear free surface waves from deep to shallow water.
• OceanWave3D - a very efficient open souce coastal engineering research tool for simulation of nonlinear free surface waves from very deep to shallow water.
• For wave propagation and breaking wave description, a coupling of OceanWave3D with OpenFoam exists in the Waves2Foam package.
• MarineSEM - Spectral Element Solvers for Marine Hydrodynamics applications
• DTU Compute GPULab Library - a high-performance GPU-based software C++ library for fast development of massively parallel and scalable scientific applications on heterogeneous distributed many-core hardware
• DYnamics Train SImulation (DYTSI)
• DTU UQ Library - a Python library for Uncertainty Quantification using high-order Spectral methods
• TensorToolbox
• SpectralToolbox
• PyORTHPOL
• UQToolbox
• A Matlab package for the construction of two-dimensional Quadrilateral Element solvers using Nodal Discontinuous Galerkin Finite Element Methods. The package works together with the easy-to-use Matlab framework provided at nodal-dg due to Prof. Tim Warburton and Prof. Jan S. Hesthaven.
• Advanced simulations using Spectral Element Methods
  • A DTU-SEM Toolbox for the generic construction and fast proto-typing of high-order Spectral/hp Element Methods for partial differential equations in any dimensions.
    Some of the basic are described in course notes of dtu course 02623:
    
  • Advanced simulations and contributions to Nektar++: An efficient h to p finite element framework
  • Learn about advanced simulations at the Nektar++ Workshop 2016.
• Massively parallel computing on many-core hardware
  • Some work related to massively parallel comnputing on modern many-core hardware are described in chapters of Designing Scientific Applications on GPUs
    
  • Developing code for execution on laptops as well as supercomputer, e.g. the European No. 1. Supercomputer LUMI.
    Example dambreak application based on Shallow Water Equations in 2D using a nodal DG-FEM solver with shock-capturing.

Poster highlights of research

• Physics-informed neural networks for 1D sound field predictions with parameterized sources and impedance boundaries. Joint with DTU Electro, 2021.
• Room acoustic simulations using the spectral element method. Joint with DTU Electro and Henning Larsen Architects, Denmark.
• Near-wellbore modeling in ECLIPSE with Computational Fluid Dynamics. Joint with Lloyds Register Consulting, Denmark.
• Parallel Algebraic Multiscale Linear Solver for Reservoir Models. Joint with Lloyds Register Consulting, Denmark.
• Adaptive spectral tensor-train decomposition for the construction of surrogate models. Joint with MIT, USA. Presented at SIAM CSE, 2015.
• A Spectral Element Method for Nonlinear and Dispersive Water Waves. Joint with Chalmers University of Technology, Sweden and MIT, USA. Presented at SIAM CSE, 2015.
• Robust Ship Hydrodynamics via Novel Methods. Joint with FORCE Technology, Denmark. Presented at SIAM CSE, 2015.
• Mixed Finite Element Multilevel Approach Enables Fast and Accurate Reservoir Simulations. Joint with Lawrence Livermore National Laboratory, US, and Lloyds Register Consulting, Denmark. 2013.
• Nonlinear multigrid for reservoir simulation. Joint with Schlumberger, UK. Presented at SPE Mathematical Methods in Fluid Dynamics and Simulation of Giant Oil and Gas Reservoirs Conference, 2012.
• Development of a GPU-accelerated MIKE21 Solver for Water Wave Dynamics. Joint with DHI Group, Denmark. Presented at the Conference Multicore-Challenge III, 2012.

Supervision in Research projects (selected and recent)

• Development of a high-order spectral element method for simulation of water waves using Incompressible Navier-Stokes Equations (M.Sc. project, Aug 2022 - Jan 2023)
• Transient vibro-acoustics using the discontinuous Galerkin finite element method (M.Sc. project, Feb - Aug 2022)
• Development of a high-order pseudospectral method for accurate simulation of water waves using Incompressible Navier Stokes Equations (M.Sc. project, Sep 2021 - Mar 2022)
• Machine Learning for Marine Engine Fault Detection (M.Sc. project, Apr 2020 - Sep 2020)
• Massively Parallel DG-FEM Simulator for 3D Room Acoustics (M.Sc. Project, Oct 2019 - April 2020)
  Project was awarded the DANSIS Graduate Prize 2020.
• Sparse Grid Methods for High Dimensional Function Approximation (M.Sc. Project, Feb 2019 - July 2019)
• Music Composing with Artifical Intelligence (Minor project, Feb 2019 - June 2019)
• Optimal Control of Nonlinear PDE Systems Using Reduced Order Modeling and Neural Networks (M.Sc., Aug 2018 - Jan 2019)
• High-Performance Computing (PhD Project, Apr 2018 - May 2018)
• Data-driven solutions of nonlinear partial differential equations through physics-informed neural networks (B.Sc., Feb 2018 - Jun 2018)
• Reduced Basis Methods for Parametrized Partial Differential Equations (PhD Project, Dec 2017 - Jan 2018)
• Application of the Spectral Element Method for an engineering hydrodynamic application (project, Jan 2018 - Jan 2018)
• Boundary conditions in wave-based room acoustic simulation methods (M.Sc. Project, Jan 2018 - Jun 2018)
• Advanced Numerical Methods for Differential Equations (project, Oct 2017 - Jan 2018)
• Applied Machine Learning for Prediction (Project, Sep 2017 - Dec 2017)
• Spectral Element Modelling of Wave-Body Interactions (M.Sc., Aug 2017 - Feb 2018)
• Function approximation using Artificial Neural Networks (B.Sc., Feb 2017 - Jun 2017)
• Robust Spectral Element Methods for Free Surface flows with Structures (M.Sc., Mar 2016 - Sep 2016)
• Numerical methods for solving Navier-Stokes equations (M.Sc., Mar 2016 - Aug 2016)
• Multi-scale Finite Volume Method for accelerating reservoir simulations (M.Sc., Sep 2015 - Feb 2016)
• Multi-level algorithms for uncertainty quantification (M.Sc., Sep 2015 - Feb 2016)
• Massively Parallel Nonlinear Multigrid on Modern Architectures (M.Sc., Mar 2015 - Sep 2015)
• Robust massively parallel free surface simulation using the Spectral Element Method (M.Sc., Feb 2015 - Aug 2015)
• Mathematical Techniques for Reduced Order Modelling (M.Sc., Nov 2014 - April 2015)
• Spectral element modelling of wave-floating body interactions (M.Sc., Mar 2015 - Sep 2015)
• Free Surface Hydrodynamics using Isogeometric Analysis (B.Sc., Feb 2015 - Jun 2015)
• Scientific software solution for visualization of large data sets (Minor project, Feb 2015 - Jun 2015)
• Hybrid Coupling of Modern Finite Element Methods (M.Sc., Nov 2014 - April 2015)
• Mesh Adaptive Techniques for Finite Element Solvers (B.Sc., Sep 2014 - Jan 2015)
• Financial Modelling using PDEs (Project, Aug 2014 - Dec 2014)
• Software design for portable and scalable scientific calculations on modern and emerging heterogeneous many-core architectures (M.Sc., Jan 2014 - Nov 2014)
• Implementing exact methods for discrete optimization on a GPU (M.Sc., Feb 2014 - Jun 2014)
• Implementing metaheuristics using GPU programming (M.Sc., Feb 2014 - Jun 2014)
• Advanced Techniques for Investigating Structures in Computational Fluid Dynamics (M.Sc., Dec 2012 - May 2013)
• Spectral Methods for Uncertainty Quantification (M.Sc., Dec 2012 - May 2013)
• Study in Modern Uncertainty Quantification Methods (M.Sc., Dec 2012 - May 2013)
• Nonlinear Multigrid for Efficient Reservoir Simulation (M.Sc., Mar 2012 - Sep 2012)
  Project was awarded the DANSIS Graduate Prize 2013.
• Feasibility study of the parareal algorithm (M.Sc., Feb 2012 - Aug 2012)
  Award for best presentation at Facing the Multicore Challenge III in 2012.
• GPU-Acceleration of Linear Algebra using OpenCL (B.Sc., Feb 2012 - Jun 2012)
• Development of a GPU-accelerated MIKE 21 Solver for Water Wave Dynamics (B.Sc., Feb 2012 - Jun 2012)
• Curving Dynamics in High Speed Trains (M.Sc., Feb 2011-Aug 2011)
• Fusion Plasma Thermal Transport Radial and Poloidal Profile Modeling (M.Sc., Sep 2010-Jun 2011)
• Acceleration of a non-linear water wave model using a GPU (M.Sc., May 2010-Dec 2011)
• Implementation of Lattice Boltzmann Method on GPU (Special course, Feb 2010 - Jun 2010)
• Marine simulation and rendering (M.Sc., Sep 2009 - Mar 2010)
• Simulation of launch and landing of small rocket (Minor project, Feb 2009-Jun 2009)

Community service

2010, Co-founder (with colleagues at Scientific Computing Section) of GPULAB with strong focus on Scientific GPU Computing, fundamental aspects of high-performance and heterogenous computing on modern many-core architectures and software development for proofs of concepts related to next-generation scientific applications using modern programming paradigms (CUDA/OpenCL) and tools. GPUlab was designated Nvidia CUDA Teaching Center May 2012 and Nvidia CUDA Research Center (PI: Allan P. Engsig-Karup) since November 2012.

I am a member of the Scientific Council of Danish Center for Applied Mathematics and Mechanics (DCAMM).