Projects
Virtual Power Plant with Artificial intelligence for Resilience and Decarbonisation
Funding Agency: EPSRC
Total Funding: £1,845,327
Duration: 1 May 2023 – 31 Mar 2025
Durham PI: Prof. Hongjiang Sun, Durham Co-Is (Computer Science): Dr. Gagangeet Aujla and Dr. Anish Jindal
Partners: University of Edinburgh, University of Birmingham, Northumbria University
Industry Partners: TNEI, Northumbrian Water, Kinewell Energy, CFMS Services Ltd, Equinor, MithraSol, Durham County Council, National Grid ESO, DVV GL, Siemens pls (UK), Northern Powergrid
Description:This project brings together leading experts in the fields of smart grid, artificial intelligence (AI), built environment, and social science to develop innovative AI solutions for enabling virtual power plants to accelerate the decarbonisation and enhance the power grid resilience to natural hazards and extreme events. Distributed energy resources (DERs), e.g., photovoltaics and batteries, play a critical role in achieving Net Zero. Although DERs have many benefits such as a reduced carbon footprint, they present complex challenges for network operators (incl. grid operation and power flow), creating a major barrier to Net Zero. How best to manage millions of DERs is still an open question, especially for improving grid resilience to natural hazards and extreme events such as storms and heat waves. The intended project outcomes are innovative physics-informed AI solutions for enabling Virtual Power Plants (VPP), capable of aggregating and managing many diverse DERs; not only improving decision-making for network operators but also enhancing the grid resilience to natural hazards and extreme events. These could also lead to reduced energy bills for millions of UK energy consumers, to greater adoption and more efficient management of DERs, and ultimately to achieve the Net Zero goal. Perhaps most importantly, this project will enable us to begin asking how we shall best tackle the Energy Trilemma (i.e., balancing energy affordability, security, and sustainability) by harnessing the power of AI using various data sources, considering not only power systems but also ever-changing climate change and environments. This therefore encourages multidisciplinary research and cross-fertilisation of these research fields.
CHEDDAR: Communications Hub for Empowering Distributed clouD computing Applications and Research
Funding Agency: EPSRC
Total Funding: £ 3,028,049
Duration: 1 July 2023 – 30 Jun 2026
Hub PI: Professor Julie McCann, Vice-Dean (Research) for Imperial’s Faculty of Engineering
Durham PI: Prof. Hongjiang Sun, Durham Co-Is: Dr. Gagangeet Aujla, Dr. Anish Jindal and Dr. Wanqing Tu
Partners: Imperial College London (Lead), University of Glasgow, University of York, University of Leeds, Cranfield University
Weblink: https://gow.epsrc.ukri.org/NGBOViewGrant.aspx?GrantRef=EP/X040518/1
Description: The CHEDDAR: Communications Hub for Empowering Distributed clouD computing Applications and Research lies on this interface between communications fundamentals and an infrastructure that serves our many communities. Bringing together a spectrum of experts not only in communications but across quantum, neuro-symbolic AI, autonomy etc who have worked in application areas from Farms to Space via Smart Cities, and who understand security and privacy, CHEDDAR aims to build a community that will provide the step change in research and be agile to address new opportunities as they arise. CHEDDAR also aims to take a different and inclusive approach to this research which includes clear opportunities for community engagement and dialogue across the UK, and in researcher development providing both technical and soft-skills round inclusivity and responsible research. With a particular focus on security/privacy and sustainability and resilience this Hub aims to accelerate the national scale innovation process in connected computing research and make UK a lighthouse globally. This is necessary as the UK has demonstrated its ability to be world leading in topics such as AI, Autonomy etc., but many EPSRC and UKRI funded advances are set to unleash their £Bn’s of potential on the economy – it requires new communication interfaces to connect them and unleash these benefits, whilst mitigating the new risks that emerge. We aim to develop the pathways to enabling new communication network design and innovation, through connections with the other 2 hubs, with centres of doctoral training and innovation knowledge centres, institutes, government and standards bodies etc. and identify mechanisms for joint capability and facility sharing and cross-fertilize the co-design of connected computing capabilities.
Course Development in Edge Computing and Analytics 2.0
Funding Agency: The British Council
Total Funding: £30,000
Duration: 2022-2024
PI: Dr Anish Jindal
Partners: Cardiff University, UK; Galala University, Egypt; TKH, Egypt
Description: This project aims to develop course on edge computing and analytics 2.0 between UK and Egyptian universities and develop collaboration in order to build edge applications on smart cities, smart grids, water and energy management to achieve SDG goals.
Traffic Management In Smart Cities Using IoT For Reducing Carbon Emissions
Funding Agency: The Royal Society, Scheme: International Exchanges
Total Funding: £12,000
Duration: 2022-2024
PI: Dr Anish Jindal, Co-I: Dr Gagangeet Singh Aujla
Partner: Dr. Luca Foschini, University of Bologna, Italy
Description: This project aims to provide an IoT-based transportation system which will contribute to the good health and well-being of citizens as well as provide better overall traffic flow in densely populated cities which helps the environment by reducing the harmful carbon emissions.
Carbon-Intelligent Computing for handling Uncertainties in Data Centers
Funding Agency: Durham SeedCorn Award
Total Funding: £11,552
Duration: 2022-2023
PI: Dr Gagangeet Singh Aujla, Co-I: Dr Anish Jindal, Co-I: Prof. Hongjian Sun
Partners: Prof. Omer Rana, Cardiff University, UK and Dr. Angelos Marnerides, University of Glasgow, UK
Description: A typical Data Center (DC) consumes 41.11 TWh energy equivalent to ~12.13% of total energy generated in the UK (UKERC, 2020). During COVID-19, individuals and organisations are shifting towards the use of digital platforms which introduced more data traffic in the DCs, and hence, further increasing their corresponding energy consumption. The project aims to design a carbon-intelligent computing framework to reduce the energy-based carbon footprints and realise the goal of carbon neutral DCs. As carbon-neutral energy carries various uncertainties such as intermittency, therefore, the proposed research would determine and control these uncertainties, and optimise the energy consumption of DCs using Artificial Intelligence (AI) to operate on 100% carbon-neutral energy.
Programmable, secure and resilient IoT communication networks for smart grid
Funding Agency: The Royal Society, Scheme: International Exchanges
Total Funding: £12,000
Duration: 2022-2024
PI: Dr Anish Jindal
Partner: Dr. Nurul Sarkar, Auckland University of Technology, New Zealand.
Description: This project aims to provide an Programmable, secure and resilient IoT communication networks for smart grid.
Resilient Self and Dynamic Data Stream Spanning Structures
Funding Agency: The Royal Society, Scheme: International Exchanges
Total Funding: £12,000
Duration: 2022-2024
PI: Dr Amitabh Trehan
Partner: Dr. Michael Elkin, Ben-Gurion University of the Negev, Israel.
Description: This project aims to provide an Resilient Self and Dynamic Data Stream Spanning Structures.
ACUTE (Algorithms for Computing with Uncertainty – Theory and Experiments)
Funding Agency: EPSRC
Total Funding: £401,205 (EP/S033483/) and £191,339 (EP/S033483/2)
Duration: EP/S033483/1: 01/11/2019-19/09/2021, EP/S033483/2: 20/09/2021-31/10/2023
PI: Prof. Thomas Erlebach, Co-I: Dr Amitabh Trehan
PostDoc: Konstantinos Dogeas (since September 2022)
Visiting Researchers and Collaborators:
-
Evripidis Bampis, Sorbonne Université, Paris
-
Christoph Dürr, Sorbonne Université, Paris
-
Nicole Megow, University of Bremen
-
Jens Schlöter, University of Bremen
Ex-Staff: Michael Hoffmann, University of Leicester (co-PI, until June 2021) and Murilo Santos de Lima (post-doc, until April 2021)
Website: https://sites.google.com/view/thomas-erlebach/home/acute
Description:How much information should we collect before making a decision? This question underlies the research area of computing with explorable uncertainty. For example, assume that we want to build a network connecting a set of branch offices. For any two locations A and B of branch offices, we have an estimate of the cost for building a link between A and B based on the distance between them. The exact cost of building a link between A and B can be determined by further investigations, but these investigations take time and cost money. If we knew the exact link cost for every pair of locations, we could determine the cheapest way of building the network using a known algorithm for the “minimum spanning tree” problem. The approach of first determining for all pairs of locations the exact cost of building a link between them is not efficient, however: It will take a long time to determine all the exact link costs, and the costs for obtaining that information will be significant. It is therefore desirable to find efficient methods for selecting in a clever way the pairs of locations for which the link costs need to be determined, while still achieving the goal of being able to build a cheap network with the information gained. Algorithms for computing with explorable uncertainty solve such problems: They specify a strategy for selecting the pairs of locations for which exact information should be determined until sufficient information has been gained to determine the best possible network to be built. More generally, computing with explorable uncertainty deals with problems where part of the input is uncertain (known only approximately) but can be obtained at a cost using a query operation.
Previous work on computing with uncertainty has focused on the setting where queries are made one by one sequentially (which may take a long time) and where the goal is to make as few queries as possible while ensuring that sufficient information is obtained to solve the problem optimally. This leaves open the question of how the queries should be selected if a number of queries can be made at the same time in parallel, which is realistic in many applications (for example, in the application outlines above, the exact costs of building links between several pairs of branch office locations could be determined in parallel). Another direction that has not yet been sufficiently considered is the setting where the goal is to optimize a combination of the query cost and the cost of the solution determine in the end. The project aims to take research in computing with explorable uncertainty to the next level by addressing these open questions and developing new algorithms that work provably well in the described scenarios. Methods developed in the project can potentially be useful to any decision-making scenarios where additional information about the input data of a problem is available in principle and can be obtained at a cost.