Electrical Engineering, Mathematics & Computer Science (EEMCS)

Over the past years, the penetration rates of renewable generation with intermittent character like solar and wind energy have been rapidly growing within the different energy systems. Combined operation of storage within specific applications of (DC) energy systems (including data centers, ships, electric aircrafts, houses, EV charging hubs, neighbourhoods, and districts among others) is one of the key solutions for energy transition. For example, the distribution grid infrastructure does not have adequate flexibility to cope up with the large increase in demand that is anticipated (the use of electric vehicles and solar energy is expected to drastically increase). System integration of the storage will also enable new research areas such as all electric ship and aircraft and will have an influence on decreasing the peak power of the generators in the various systems.
To solve these challenges, it is required to understand the physics of different storage devices, to develop effective integration strategies and to build modular systems in which the power electronic converters will be integrated together with the storage devices to enable system integration. Additionally, the research is required to cover different storage technology alternatives and to encompass the combined operation of storage within specific applications. Therefore a multidisciplinary approach that spans from the understanding of aging and electrochemical processes to the implementation of control algorithms is essential.
 

Fully aware of the challenges that mankind faces with climate changes and energy transition, the mission of Photovoltaic Materials and Devices group at TU Delft is to deploy photovoltaic (PV) systems everywhere providing green electricity for the sustainable electrification of society. As any surface can be nowadays endowed with PV technology, X-Integrated PV (X-IPV) systems comprise all possible forms of photovoltaic systems, which can upgrade with PV technology both urban and open environments. The PVMD group works on the whole PV value chain, spanning from atomistic scale up to kilometer scale. At the PVMD group new solar energy materials are studied and integrated into high-efficiency multi-junction solar cells; novel PV-powered multifunctional building elements are conceived and implemented in customized PV systems; advanced opto-thermo-electrical modelling describe the operation of solar cells, modules and systems; and sustainability and possible recycling of PV technologies are assessed. Leveraging these research areas at the PVMD group, the multi-faceted discipline of X-IPV systems emerges. On the one hand, this new discipline tailors PV technologies and adapts PV modules for implementing innovative system solutions in both urban and open environments. On the other hand, it studies the impact of massive PV penetration on the development of smart cities, the electrification of industrial processes, the uptake of electrical mobility, the electrification of domestic activities, the sustainable agriculture and farming as well as on the stability of the electrical grid.

The energy transition towards carbon-free technologies is accompanied by a second transition that affects all sectors: the digital transformation. Digital technologies such as the Internet-of-Things, blockchain, cloud computing, or peer-to-peer networks are also changing the power system. Distributed systems, consensus-based agreements, and the ability to work with large numbers of active players are new degrees of freedom in designing and operating power systems. Future designs can make use of this freedom using distributed data for supporting the distributed operating paradigms rather than centralized paradigms of the past.
This Technology Fellowship shall cover data-driven methods (from Machine Learning and AI), the science, and the application of future control toward autonomous, distributed control paradigms in power systems. Topics of interest are - among others - self-organizing power systems, cyber- physical and cognitive energy agents, holarchies, meshed Microgrids, data-driven preventive-corrective controls and market-driven methods, transparency vs. privacy, and mastering the complexity of such digitalized systems. In the end, all standard operations of a power system (dispatching, N-1 risk assessment, scheduling, etc.) shall be done in a distributed way. Special focus is put on covering both the physical and the computational parts.

Quantum computers and quantum sensors promise performance well beyond those offered by their classical counterparts by exploiting quantum effects. At the Quantum & Computing Engineering (QCE) department, we aim at bringing those systems out of the lab into real applications by engineering and scaling the current proof of concepts up to the complexity necessary to tackle practical problems. We address those challenges over the full computer/sensor architecture stack, including the integration of quantum and classical devices, their (cryogenic) electrical interface, the microarchitecture, and the algorithms. Collaborations with TU-Delft experts in classical and quantum disciplines are highly encouraged, both within the department/faculty and with other faculties and institutes, e.g., Applied Sciences and QuTech. We aim to strengthen our team with new faculty members and groups that can reinforce ongoing research themes or bring in unique expertise, such as system and device modelling.

This theme is foundational for the realization of the omnipresent computation platforms, from simple IoT nodes to complex high performance computing clusters, that surround us and improve our life quality, in terms of, e.g., health care, access to information, safety, while diminishing environmental damage, e.g., anthropogenic climate change. In synergy with CMOS and emerging fabrication technologies, computer architecture innovations attempt to fulfill our computation hungry civilization demands. However, demands are continuously increasing while performance approaches saturation regime, which calls for Alternative Computer Architectures and Paradigms.  This theme investigates alternative roads to overcome hurdles towards energy effective, high performance, reliable, and secure hardware platform realizations based on traditional CMOS and/or emerging technologies. It contributes to computer architecture fundaments by introducing and promoting new concepts (e.g., reconfigurable, in-memory, neuromorphic computing, edge computing) by pioneering work in computation with emerging technologies and materials such as memristors, spintronics and graphene.

This theme focuses on developing appropriate design mechanisms to be incorporated in integrated circuits (ICs)  in order to guarantee the required quality, reliability, security, etc. Developing, designing and deploying of such ICs is by far more than producing a correct functioning IC in a simulation environment.  It rather requires the design of dedicated  countermeasures to be deployed both at time zero (during manufacturing) and  in field (during the lifetime of the application) in order not only to facilitate the manufacturability of such integrate circuits, but also to ensure key performance indicators of such chips including yield, quality, reliability, security,  lifetime management, etc.

This theme focuses on the fundaments of Network Science. Networks are everywhere! We all know mobile networks, the Internet, logistics networks, energy networks, transport networks and on-line social networks, while besides these man-made networks the most complex network is the human brain.  This theme aims  at understanding  the graph structure of networks and the processes on networks as well as the processes that create time-varying networks such as human mobility (e.g., transportation) and biological networks (e.g., diffusion  of epidemics).
The theme investigate geometric representations of networks, epidemic spread on networks, spectra of graphs and network algorithms. In addition, we apply our mathematical knowledge to the design and control of critical infrastructures, such as telecommunication networks and power grids, to make these networks robust, resilient, efficient and reliable.

Next-generation (XG) wireless sensing and communication systems share the same need to transmit and receive EM waves, resulting in similar (RF/mm-wave) front-end technology and system architecture. The XG theme within the department aims to address the industrial and societal challenges associated with a lot of applications, such as communication systems for 5G and beyond, massive MIMO, wideband/high-speed data links, (cognitive) carrier aggregation, sensing systems for use in space, environmental and safety applications, automotive radar, distributed sensing, high-accuracy localization, and object recognition.

The ASSys theme centres around developing (distributed) sensor systems capable of autonomous operation, including IoT systems, sensors for robots, autonomous drones, environmental monitoring, or for operation in harsh environments. Specific scientific challenges in this domain encompass the realization of ultra-low power interface circuits and the integration of digital circuitry for signal processing and machine learning. 

The theory of partial differential equations is the key to understanding the qualitative and quantitative behaviour of nonlinear systems arising in the natural sciences and at the same time plays a fundamental role in other parts of mathematics such as differential geometry and applications in finance. We have a long and internationally recognised tradition in PDE, ranging from fundamental aspects to modelling and the development of efficient numerical schemes. This theme unites many currently existing research lines, and we aim at expanding it into its key research theme by investing in three areas at the forefront of modern research in PDE, in strengthening research in foundations of mathematical finance and novel numerical opportunities.

The ‘data explosion’ leads to many challenging problems in the field of data science. One challenge concerns the modelling and processing of complex and dynamic data structures and drawing inference from these. Also fundamental and computational aspects of statistical learning algorithms, including issues of causal inference and fairness in Artificial Intelligence, require deep mathematical insights we want to further develop, combined with data available in many areas of application.

Due to the continuing increase in computing power and the development of new computer architectures, new opportunities emerge in the field of Computational Mathematics. To optimally use the capacities of modern computer architectures, it is necessary to have deep insight into the characteristics of this hardware as well as into state-of-the-art numerical procedures. Current developments also enable us to simulate many sets of coefficients for the used partial differential equations. This opens up possibilities to approximate solutions of stochastic partial differential equations. These problems are an order of magnitude harder than the underlying
deterministic partial differential equations. To analyse the increasing amount of data and detect patterns, computational methods on highly tuned GPU clusters are needed.
 

With our society becoming increasingly more digital, the security of information transmission is crucial. With the foreseen arrival of the quantum computer, this task will be even more complex. To build the theory needed in order to ensure information security, we will use and further develop our expertise in discrete optimisation, combinatorics, graph theory, algorithms, algebraic discrete mathematics and coding theory.

Stochastic Modelling in Science and Engineering

Stochastic models to describe complex systems, from the natural sciences or an engineering context, have been very effective in the past decades. DIAM is strong in the field of interacting particle systems. In physics, these give accurate descriptions of phenomena such as metastability, phase transitions, shape of (growing) interfaces and self-organized criticality. In biology, they capture phenomena such as the spread of an infection, the evolutions of allelic traits of a population in the presence of mutation and selection, and transport of molecular motors within a cell. Stochastic models are also developed to better understand fundamental phenomena in engineering applications, related to spatial statistics, shape constrained high dimensional models, extreme value theory and dependence modelling. These are closely connected to risk quantification, which is needed to support decision making. Research in this research theme has important interfaces with research areas including physics, biology, materials science, meteorology, climate science, finance, economics and health.

This theme encompasses devising, implementing and evaluating fundamental (classical and data-intensive) AI solutions to improve the machine-level processes involving humans and supporting them in making decisions. Research topics of interest include, but are not restricted to: interactive and hybrid intelligence, intelligent agents and multiagent systems, computer vision and machine learning, user modeling and search and recommendation.

This theme focuses on enabling humans to efficiently handle large quantities of data, but also to get insight into and analyse complex (big, unstructured, distributed, heterogeneous, web) data and the AI decisions, actions and their consequences forthcoming from data collection and processing. This research also generates scientific foundations of data-driven, human-centric AI tech- nology. Research topics of interest include, but are not restricted to: data management and engineering, bioinformatics, content generation and data visualization, and data analysis and process modeling in complex networks.

This theme brings together the expertise in the areas of software engineering, programming languages, and software reverse engineering and is aimed at developing new ways to safeguard key quality attributes of complex software systems, such as robustness, reliability, and evolvability. To that end, this theme has combined foundational (e.g. program semantics), empirical (software analytics), and AI (model learning, search-based software engineering) research. Research topics of interest include, but are not restricted to: software analytics, AI for software engineering, language engineering, and software verification.

This theme aims at devising, implementing and analyzing fun- damental algorithmic and systems concepts for distributed and networked systems. Focus has been on distributed trust, in particular on the problems related to scalability and consistency in existing blockchain-based approaches, Internet of Things, with special attention to low-power devices, big data processing, which addresses the complexity and scalability challenges in the data streaming and data processing components of these systems, and visible light communication, aiming at unlocking the bandwidth of the visible light spectrum for wireless communication.

This theme targets foundational computer science challenges in security and privacy, namely rigorously showing that a computer system is inherently (in)secure, and the maturing of design-for-security and privacy methodologies. Research topics of interest include, but are not restricted to:  AI for intrusion detection and prevention, AI-based testing for secure systems, security of AI, and privacy protection.