October 2025 Entry

TITLE: Magnetic control of hydrogen flames with real-time data assimilation

DEPARTMENT(S): Department of Mechanical Engineering and Department of Aeronautics

SUPERVISORS:  Dr. Andrea Giusti and Prof. Luca Magri

Transition to zero-carbon aviation requires the use of zero-carbon fuels coupled with aircraft electrification. In next-generation hybrid thermal-electric propulsion, there will be availability of electromagnetic energy onboard, which we will exploit to achieve safe and efficient combustion of hydrogen. We aim to employ magnetic fields to design “flames on demand”, i.e. flames with controlled shapes and reacting characteristics. Magnetic control will counteract the propensity to flashback of hydrogen flames, which is currently a major showstopper for using hydrogen in aero propulsion. Also, magnetic fields could enable modulation of mixing, with the possibility of leading to new ultra-low NOx technologies. In this project, we will develop an innovative real-time multi-physics model that is able to predict the response of reacting gases to magnetic fields. The real-time model will combine (i) data from simulations at atomistic level; (ii) reference experimental data; (iii) a multi-physics solver in the continuum; and (iv) a regularised bias-aware data assimilation method to perform state, parameters, and model error inference (with a scientific machine learning decomposition). The new knowledge and model developed in this project will enable a paradigm shift in the design of propulsion systems with hydrogen, in which the geometry is decoupled from the flame behaviour.

TITLE: Model and atmospheric sensors' fusion for optimal aviation climate impact reduction with scientific machine learning

DEPARTMENT(S): Department of Aeronautics

SUPERVISORS:  Prof. Luca Magri and Dr. Sebastian Eastham

Aviation has significant non-CO2 climate impacts due in part to the formation of contrails. Since contrails only form in specific conditions it should be possible to reroute to avoid their formation, but it is not yet possible to accurately predict those conditions. Although contrail sensors are being developed, these will still have limited accuracy and must be interpreted as part of an imperfect forecasting system.

This project aims to understand the opportunities and limitations of novel sensors in an optimization framework. The goal is to develop an efficient, robust optimization algorithm with scientific machine learning, which can estimate the effectiveness of different re-routing strategies given uncertainties in sensor and model accuracy. Simulators of sensors will need to be created, including a parameterization of accuracy as a function of (e.g.) time of day, location, altitude, and temperature. These will then need to be integrated into a forecasting model which can assimilate the (artificial) sensor data and provide forecasts of contrail-forming regions at different levels of confidence. An aircraft performance model will be needed to predict contrail formation impacts. Finally, different routing strategies will be optimized to find the maximum benefit given different levels of sensor accuracy and given different scales of sensor deployment.

TITLE: Quasi-linear approximation of the aircraft wake

DEPARTMENT(S): Department of Aeronautics

SUPERVISORS:  Prof. Yongyun Hwang and Dr. Kostas Steiros

The formation of contrail cirrus in aircraft exhausts is currently understood to contribute as much to climate change as the CO2 emitted from fossil fuel combustion. Unfortunately, the semi-empirical global models used in contrail prediction are with large uncertainties, partly because they do not capture the complex flow instabilities that appear in the turbulent aircraft wake, whose physics largely determine the expansion and fragmentation of the contrail plume. To model such instabilities, one currently needs to resort to high-fidelity simulations, which are too expensive to be used in contrail prediction protocols.

This project will investigate whether the instabilities that develop in the far wake of aircraft and simplified bluff bodies can be modelled using modelling based on the quasi-linear approximation of the Navier-Stokes equations. Previous work has demonstrated that this theoretical approach can predict the instabilities that develop in turbulent pipe-flow, at a fraction of the cost of high fidelity simulations. The goal of the project will be to apply this approach to turbulent wakes. The project will involve both experiment, where an extensive data-set of turbulent wakes will be acquired from wind tunnels, and mathematical modelling where the measured wake instabilities will be studied using the quasilinear approach.

TITLE: Physics-aware, Operation-aware Reinforcement Learning for Fuel-saving Airline Fuel Loading Strategy

DEPARTMENT(S): Department of Aeronautics

SUPERVISORS:  Dr. Rhea Liem

More accurate, operation-specific flight fuel planning has emerged as a strategy for reducing aviation fuel consumption. Airline dispatchers usually load extra contingency fuel -- on top of trip and regulatory fuel -- at their discretion based on weather conditions and airway congestion, usually on the basis of worst-case scenarios. The excessive fuel loads increase their payload and thus their fuel consumption. We aim to derive a new fuel planning strategy that can accurately predict the amount of fuel required under different fuel contingency requirements and minimise the fuel penalty from carrying excessive fuel. The model must be able to represent the interdependence between the amount of fuel carried and the amount of fuel burned during a flight under certain aircraft and mission configurations, which cannot be achieved by data-driven models alone. In this project, we will develop a physics-aware, operation-aware reinforcement learning (RL) technique, which complements data-driven and statistical models with physics-based models. To establish a realistic simulated environment for RL, we will combine physically-consistent modelling of aircraft fuel consumption with a stochastic representation of operational variation. To ensure flight safety, the solution methods must adhere to current regulatory fuel requirements. The benefits of this new method will be compared against the statistical contingency fuel method that is currently used by airlines.

PROJECT 5

TITLE: Infosymbiotics for sustainable aviation: on non-CO2 emissions

DEPARTMENT(S): Department of Aeronautics

SUPERVISORS:  Prof. Laura Mainini and Dr. Sebastian Eastham

Different energy vectors and alternative technologies are being explored to enable a cleaner air transportation network. However several questions remain about the true environmental impacts of those solutions across space and time scales, and how impacts are changed by design and operational decisions. This problem is often treated in a linear fashion: quantify impact, estimate sensitivity to some upstream choices, and adapt. Such static framings cannot adapt or properly exploit capabilities such as live reporting, rapid atmospheric simulation, Earth observing systems, and digital twins. This project will instead address impact characterization and decision tasks as symbiotic problems: accordingly, aircraft and fleets are conceived as multi-sensing systems whose measurements can inform operational decisions at scale and actively mitigate impacts. Infosymbiotics approaches will be advanced to integrate adaptive models and data acquisition as mutually informed dynamic systems; methods will be developed at the intersection of scientific computing and machine learning for decision-oriented sensing, data acquisition/assimilation, uncertainty reduction, aerospace engineering, and Earth system science. The project will show how aircraft design and operations choices on timescales of years can enable a downstream feedback loop, with shorter-term decisions made in a data assimilation framework which allows and empirically verifies a reduction in environmental impact.

TITLE: Actionable SAF Pathways: multi-criteria assessment of production and impact on aviation

DEPARTMENT(S): Centre for Environmental Policy and Department of Aeronautics

SUPERVISORS:  Prof. Niall Mac Dowell and Prof. Laura Mainini

There are multiple pathways to producing sustainable aviation fuel (SAF), each presenting their own advantages and criticalities from a technology, economic, environmental, and policy perspective. However, open questions remain about the impact of specific choices of fuel alternatives on airport logistics (storage, handling, refuelling) as well as on the design and operation of aircraft and fleets. Within this context, the purpose of this PhD project is to perform a comparative analysis of the different SAF pathways through the lens of impact on aircraft design, operations, and logistics. The project will build up on a system perspective and multidisciplinary modelling and simulation frameworks will be developed to capture the implications of fuel alternatives/choices at scale. Computational search methods and multisource learning schemes will be advanced to ease the formal and algorithmic tractability of those large-scale comparative assessments, which could also account for technology and availability uncertainties associated with different SAF pathways.

TITLE: Ιnteraction of contrails from SAF and H2 combustion with aircraft trailing vortices: High fidelity simulations and reduced order modelling.

DEPARTMENT(S): Department of Mechanical Engineering and Department of Aeronautics

SUPERVISORS:  Prof. George Papadakis and Dr. Stelios Rigopoulos

Contrails are clouds formed behind aircraft due to ice nucleating on particles at the engine exhaust and have been identified as the greatest impact of aviation on climate change. Sustainable Aviation Fuels (SAF) and hydrogen (H2) have been proposed as alternatives to the standard jet fuel to reduce CO2 emissions and there is an urgent need to understand better the mechanisms of contrail formation from these fuels. The proposed PhD research will build upon the work that will be carried out in the recently funded NERC project “Contrails from SAF and H2 combustion; from lab experiments to global mitigation policy”. In the proposed PhD project, we will focus on the vortex regime, which is the interaction of the exhaust jet with the aircraft trailing vortices. We will perform high resolution Large Eddy Simulations and develop parameterised reduced order models (ROMs) of the interaction process for different ambient conditions (humidity, temperature etc). The long-term objective is to provide accurate contrail parametrisations that can be included in global weather and aviation system models to reduce current uncertainty and guide global mitigation efforts (these aspects are studied by our NERC project collaborators).

TITLE: Co-Design of eVTOL Systems

DEPARTMENT(S): Department of Aeronautics

SUPERVISORS:  Dr. Urban Fasel and Dr. Sophie Armanini

Recent advances in electric propulsion may enable new forms of mobility and help to tackle urgent environmental challenges. Electric vertical-take-off-and-landing (eVTOL) aircraft, in particular, have become a key focus in industry and academia. Thanks to zero in-air emissions, they promise to (1) replace highly climate-impacting regional and commuter aircraft, (2) alleviate ground traffic and reduce emissions in metropolitan areas, and (3) improve connectivity in hard-to-access regions. However, many challenges must be met before eVTOLs can take to the skies. eVTOLs have complex dynamics involving multiple flight modes, which need to be handled safely and efficiently. Moreover, eVTOLs will need to operate in confined spaces (cities), close to critical infrastructure, and above densely populated areas, implying stringent safety requirements.

This project aims to advance eVTOL design and operations research, contributing to more sustainable short- to medium-range transport, while also ensuring a high level of intra- and inter-city connectivity. The closely coupled design and operation of eVTOLs will be investigated by means of computational modeling approaches for co-design problems. This will allow incorporation of operational cost efficiency, regulatory aspects, and sustainability and environmental compatibility directly into the preliminary design of eVTOLs. The obtained insight will contribute to the definition of the future air mobility ecosystem.

TITLE: Dynamics and control of stratospheric vehicle for atmospheric sensing 

DEPARTMENT(S): Department of Aeronautics

SUPERVISORS: Prof. Rafael Palacios &  Prof. Thulasi Mylvaganam

Solar electric high altitude pseudo satellites (HAPS) have the potential to offer satellites like services but at a fraction of the cost and thus provide a new capability for atmospheric research. Although it is possible to close the delicate energy balance using state of the art power systems, it requires extremely light weight span loaded high aspect ratio aerostructures. This brings a degree of structural flexibility which is orders of magnitude more complex that traditional aviation when considering flight envelopes, stability modes, and flight controller, across an altitude range of sea level to 80,000ft. Current high-performance HAPS have extremely restricted operating envelopes due to the vulnerability of these structures to meteorological conditions. The aim of the project is to embrace the potential of flexible structures with complex distributions of sensors and actuators to reduce structural loadings and dramatically increase operating envelopes. Unlocking this technology has huge potential beyond the particular vehicle class, as advanced gust load alleviation systems are needed for wing aspect ratios above 10. The project will build on simulation expertise at Imperial to combine aerostructural modelling of high aspect ratio structures, controller design for optimal performance, and flight testing on industry-sponsored instrumented scale flight test vehicles. The industrial partner is Voltitude Ltd.