Project description
Enhancing resilience in natural gas networks
Natural gas production and transmission (NGPT) networks are vulnerable to (partial) failures, leading to supply disruptions and methane emissions. These networks require enhanced resilience. Supported by the Marie Skłodowska-Curie Actions (MSCA) programme, the ResilientGas project aims to develop a new framework using Markov Decision Process (MDP) and Deep Reinforcement Learning (DRL) to optimise the resilience of integrated NGPT network systems. The project focuses on strengthening European NGPT networks and aims to benefit gas producers and transmission operators. This research will be conducted at Politecnico di Milano in Italy, with additional industrial expertise from ARAMIS SRL.
Objective
Natural Gas Production and Transmission (NGPT) network systems are vulnerable to (partial) failures, leading to natural gas supply disruptions and methane emissions, thus, calling for enhanced resilience. This project proposes to develop a novel framework for modelling and optimising the resilience of integrated NGPT network systems by employing Markov Decision Process (MDP) frameworks that identify optimal policies for system operation and maintenance. This novel approach allows for incorporating the dynamics of the system, including network topology and functionality changes upon operation and maintenance interventions; thus significantly improves on present-day practices. The MDP model is solved by Deep Reinforcement Learning (DRL) algorithms integrating a novel network capacity model developed using Graph Neural Networks (GNN). For illustration, I consider a European NGPT network comprising of offshore production platforms, subsea pipelines and onshore pipelines exporting the Norwegian gas to Europe. The project outcomes can be beneficial to gas producers and gas transmission operators whose optimal decisions contribute to more resilient natural gas supply for the benefit of Society. The success of ResilientGas hinges on advanced training on network resilience modelling, and on computational intelligence methods such as DRL and GNNs. I will thus need to compliment my own skills with a scientific tradition that is best studied with the team at Politecnico di Milano in Italy, where I will be under the direct supervision of Prof. Enrico Zio, whose research team is internationally recognised to lead the development of advanced tools and techniques for reliability, risk and resilience. A six-month placement in ARAMIS SRL (www.aramis3d.com) a cutting-edge R&D company, is foreseen to complement the research training with industrial practice and first-hand experience in deploying project outcomes, enabling me to position myself as a top researcher in this field.
Fields of science
- natural sciencescomputer and information sciencesartificial intelligencemachine learningreinforcement learning
- engineering and technologyenvironmental engineeringenergy and fuelsfossil energynatural gas
- engineering and technologyelectrical engineering, electronic engineering, information engineeringinformation engineeringtelecommunicationstelecommunications networks
- natural scienceschemical sciencesorganic chemistryaliphatic compounds
- natural sciencescomputer and information sciencesartificial intelligencecomputational intelligence
Keywords
Programme(s)
- HORIZON.1.2 - Marie Skłodowska-Curie Actions (MSCA) Main Programme
Funding Scheme
HORIZON-TMA-MSCA-PF-EF - HORIZON TMA MSCA Postdoctoral Fellowships - European FellowshipsCoordinator
20133 Milano
Italy