Research Explorer Ruhr: Hosts and Application
Here you find the profiles of the participating professors. You can conveniently apply directly at the end of each profile: please fill in the application form and send it to us via the green button at the end of the document.
Institute for Combustion and Gas Dynamics (IVG)
Mohri group – Tomography
Computed Tomography (CT) is a mathematical tool that is used to reconstruct instantaneous three-dimensional (3D) fields, and can be combined with different types of measurements such as laser absorption, schlieren and emission, to analyse many practical flows and processes in engineering. We combe CT with flame chemiluminescence emission measurements and background-oriented schlieren (BOS) imaging to reconstruct the 3D instantaneous flame shape and refractive index fields, respectively. The data helps better understand combustion processes, and can be used for validation of advanced numerical simulations. We have several CT algorithms, based on different concepts such as Algebraic Reconstruction Technique (ART), probabilistic and Bayesian-based ones. We also have a generic setup of 30 small CCD cameras that are usually used for real flow/flame measurements. The larger number of measurements possible with this setup help improve the reconstruction quality. Accurately determining the camera locations also plays an imporatnt role in the quality of the CT results, and we work on improving camera calibration methods.
The ideal candidate should have very strong numerical skills, in particular fluency in programing languages such as Fortran, C and Matlab. A background in mathematics, computer science and image processing is an absolute necessity. General understanding of fluid mechanics, combustion, and some experience in experimental work on diagnostics and optics in fluid mechanics and related fields of research will be a bonus.
Host Website: Link
Interdisciplinary Centre for Advanced Materials Simulation (ICAMS)
Scale-bridging thermodynamic and kinetic simulation
Phase transformations are phenomena of general importance and play a significant role in all areas of materials processing. In a scale-bridging approach we incorporate first principles calculations of phase stability and transport coefficients and analyze our results with respect to macroscopic properties of condensed matter. Among numerical techniques applied within our department are first-principles methods for phase-stabilities, the CALPHAD method (CALculation of PHAse Diagrams) to calculate phase-stability, the phase-field method to describe phase transformations and microstructure evolution in crystalline materials. Last, but not least, the Lattice-Boltzmann method is applied to solve surface tension driven flow.microstructure of materials and control their macroscopic properties. The research activities in the department focus on the mesoscopic scale of heterogeneous multiphase microstructures. We apply different theoretical methods to investigate the constitutive laws controlling microstructure evolution during various stages of materials processing, ranging from solidification to solid-state transformations during thermal processing.
The candidate should have a sound background in computational mechanics in general. Knowledge about phase-transformation and microstructure evolution in materials including damage and fracture are of advantage.
Host Website: Link
Department of Civil Engineering
Structural Analysis of Plates and Shells
Research in the group of "Structural Analysis of Plates and Shells" at the University of Duisburg-Essen focusses on the development of robust and efficient numerical simulation techniques for solid mechanics and coupled problems with a particular focus on time-dependent problems. We are employing and developing the scaled boundary finite element method (SBFEM). This semi-analytical technique excels in modelling radiation damping and singularities and facilitates the use of structured meshes in the context of image-based analysis.
Our main research interests include:
Dynamic soil-structure interaction
Seismic wave modelling
Numerical modelling of wave-based methods for non-destructive testing and parameter identification
Image-based mesh generation and analysis with applications to material modelling and computational homogenization
Damage and fracture modelling, thermally-induced crack propagation
Non-classical discretization methods for thin-walled structures, solid mechanics and coupled problems
The ideal candidate has obtained a PhD in an area related to the above fields of interest. He or she has a strong background in computational mechanics and can demonstrate very good programming skills. A track record of publications in high-quality international journals would be an advantage.
Host Website: Link
Institute of Hydraulic Engineering and Water Resources Management
Hydraulic Engineering deals with all measures to use water or protect people from the risks and danger of water. Water Resource Management describes all exploitation of water by man. Our institute deals with all matters of water management, flood protection, and the protection and keeping of natural waters.
Our main interests are:
Urban water management
Hydro morphology and morphodynamics
Supervision of river rehabilitation
Water quality management
Water quality modelling and mass flow balancing
Implementation of the Water Framework Directive
Socioeconomics of water management concepts
Flood control and risk management
Provision of the consequences of climate change
The candidate should have a sound background in hydraulic engineering, water resource management and aquatic ecology in general. Knowledge of hydraulic modelling and programming software like R and Phyton is of advantage.
Host Website: Link
Institute for Computer Science and Business Information Systems
Networked Embedded Systems
Networked embedded systems have allowed the seamless integration of technology into everyday life. This embedding of sensing, computation, and actuation has the potential to revolutionise the efficiency of our society at large. In this scenario, we investigate the challenges involved in realising robust systems made of heterogeneous mobile and stationary devices, e.g., smartphones, Internet of Things embedded platforms, wireless robots.
Our focus is on solutions that make these systems work in practice in real-world scenarios. In particular, we perform research on a range of topics spanning (1) the recognition of high-level context information from low-level sensor readings with a focus on spatial and location information, (2) protocols and methodologies supporting the dependable interconnection between wireless embedded devices, (3) approaches to support the natural interaction between robots and humans. To validate our research results, we participate in research and development projects where we provide necessary tools and novel platforms that aid the application development and operation.
We are seeking enthusiastic and committed researchers with a strong background in the design of wireless and/or robotic systems as well as experimental research. The candidate should have a solid publication record in flagship conferences and journals. Applications from related disciplines that could broaden our research portfolio are also welcome.
Host Website: Link
Our group has a strong focus on theoretical foundations of machine learning and data mining methods. Our results cover learning theory, numerical optimization, probability theory, distributed learning, privacy and trustworthiness for classic machine learning techniques (like decision trees, kernel methods, and probabilistic models) as well as deep learning techniques. Our insights reveal the strengths, weaknesses, and resource consumption for each model class and help us to derive error bounds and quality guarantees. We apply our results to practical problems from industry, logistics, and other branches of science like astroparticle physics and medicine. Exemplary results include: integer Markov Random Fields, which allow us to run sophisticated probabilistic models on ultra-low-power hardware; stochastic quadrature, a new technique for probabilistic inference that comes with error guarantees and works with any conditional independence structure; new techniques for Boolean matrix factorization which detect patterns in binary matrices while controlling the false discovery rate; new deconvolution methods for Cherenkov astronomy; rapidly learning kernel-classifiers with conditioned stochastic gradient descent; efficient machine-type communication using multi-metric context-awareness for cars; uplink transmission power prediction for LTE and upcoming 5G networks; ...
The perfect candidate should know the theoretical foundations of basic machine learning and data mining techniques (e.g.: bias-variance decomposition, Mercer's theorem, Hammersley-Clifford theorem, principle of maximum entropy, minimum description length, ...). Her or she should be well-versed in data structures, algorithms, complexity theory, and common programming languages (C++, Java, Python, ...). Using unix command-line tools and standard data mining software shall be easy for her. However, a subset of these skill should suffice to spend two weeks in our group.
Host Website: Link