BSc, BEng, PhD, MAIAA
- Primary position:
- Senior Lecturer
Dr. András Sóbester is a Senior Lecturer in Aeronautical Engineering. His research focus is on developing techniques for the aerodynamic optimization of aircraft (as distilled into the parahprase to the right by aircraft geometry pioneer E. Schmued), as well as on understanding and enhancing the potential of 3d printing in the aerospace design process.
Stratospheric flight (whether on fixed wings or balloon-borne), especially the design of high altitude Unmanned Air Vehicles (UAVs) for scientific applications, is another key area of interest. He leads the ASTRA (Atmospheric Science Through Robotic Aircraft) initiative, which aims to develop high altitude unmanned aircraft for meteorological and Earth science research.
Dr. Sóbester lectures on two modules on the University's Aeronautics and Astronautics course: Aircraft Operations and Mechanics of Flight (first year) and Aircraft Design (third year). He also supervises fourth year group design projects.
Having obtained degrees in Mechanics and Mechanical Engineering (1st) and Design and Manufacture (1st), he joined the University of Southampton as a PhD student in 2000. Upon completion of his doctorate ("Enhancements to Global Design Optimization Techniques") he worked as a Research Fellow in the Computational Engineering and Design research group on a series of industrial research projects for Rolls-Royce and BAE Systems.
His appointment to an academic position in 2007 was followed by the award of a five year Royal Academy of Engineering (RAEng) Research Fellowship. András's RAEng research focused on reducing the environmental impact of passenger airliners through unconventional airframe geometries.
In addition to 20+ journal articles, András is the author of a book exploring the scientific and technological limits of flight at high altitudes, as well as co-author of a Wiley text on statistical modeling for engineering design applications.
"Using sophisticated optimization algorithms to engineer aircraft shapes the air likes to touch"
The University of Southampton's electronic library (e-prints)
Conference or Workshop Item
- Design optimization (global and local search techniques)
- Aircraft design
- Unmanned air vehicles, in particular their use in science missions
- Rapid prototyping methods in aerospace engineering
- High altitude flight
- Aerodynamic shape optimization
- Shape parameterization, geometry dimensionality reduction methods
- Surrogate modelling and optimization methods based on surrogate update schemes
- Design technologies
- Expert systems
- Climate model tuning
The ASTRA initiative…
…investigates new technologies for making low cost observations of the physical parameters of the atmosphere. ASTRA develops and tests platforms capable of delivering scientific instruments to altitudes ranging from the planetary boundary layer (hundreds of meters) to the upper stratosphere (up to 50km).
A key research thrust here is the introduction of new developments in rapid multi-disciplinary design and rapid prototyping into atmospheric platform design, enabling the production of highly bespoke, high performance vehicles around a given payload and mission.
With platform development cycles reduced to weeks, instrument development and deployment can also be accelerated, enabling faster experimental cycles, more rapid deployment and lower operating costs.
A related research interest is the design of stratospheric fixed wing aircraft, in particular of balloon-launched unmanned vehicles. On their own, or as part of a swarm, such aircraft represent a low cost and effective means of sampling pollutant concentrations or other parameters in large blocks of the upper atmosphere.
As part of a Royal Academy-funded Research Fellowship Dr Sóbester investigated novel formulations for the mathematical representation of aircraft external surfaces, as well as for the most efficient use of physics-based performance analyses of the resulting aircraft. The goal of this (ongoing) work is to facilitate multi-disciplinary design optimization studies, which enable the development of high performance aircraft with low environmental impact.
The study included an application of such technologies, centred around reducing environmental noise in communities living under the departure paths of airports by shaping airframe geometries such that they reflect engine noise upwards, away from residential areas.
Surrogate Modelling and Knowledge Capture
A surrogate model is a mathematical formulation used to capture salient information from expensive and/or sparse data (obtained experimentally or through computer simulations). Dr. Sóbester’s work over the last decade has encompassed a number of areas of application for such technologies, ranging from machine learning systems designed to capture physics- or geometry-based knowledge related to an aerospace system to frameworks designed to accelerate the tuning of computational climate models.
Primary research group: Computational Engineering and Design
Balloon-borne stratospheric instrument platform demonstrating rapid prototyping technologies
One of ASTRA’s goals is to improve access to extreme altitudes
Unconventional airframe geometries for fan noise shielding
Global sea ice extent, as predicted by a numerical climate simulation
Mr Christopher Paulson
Chris is investigating the place of additive manufacturing techniques in the development process of small unmanned aircraft. Technologies such as Selective Laser Sintering (SLS) have the potential to speed up the development process by becoming an integral part of the Multi-Disciplinary Design Optimization process through the ability to create several flying prototypes through the process. These iterations, which thus augment the the computational analysis through flight test data, will lead to a better final design, which, in some cases, can be a 3d printed air vehicle itself.
Mr Christopher Crispin
One of the goals of the ASTRA initiative is to develop a Massive Atmospheric Volume Instrumentation System (MAVIS). At the centre of MAVIS lies a novel concept for an atmospheric sensing system: a fleet of small, very light, instrumented gliders are released en masse from a high altitude meteorological balloon over the environment to be observed. During their autopilot-guided descent along paths optimized for sampling efficiency, they collect a dense set of readings, which can subsequently be converted into an accurate map of the quantity (e.g., pollutant concentration) being observed. Chris Crispin is looking into how to optimize the efficiency of the sampling through evolving efficient autopilot algrithms for the individual aircraft, as well as exploiting any emerging behaviour exhibited by the glider swarm.
Dr Stephen Powell
Stephen tackled a complex multi-disciplinary design optimization problem: that of finding the optimum locations of gas turbine engine nacelles mounted above the wings of an A320/B737-scale commercial airliner. Such an installation has numerous advantages. It permits very high bypass ratios, reduces the risk of ingesting foreign objects on take-off, etc. The greatest advantage, however, is that it mitigates the noise nuisance on communities on the ground by virtue of the airframe reflecting the much of the broadband fan noise upward, away from the ground. The key challenge here was to come up with effective and efficient ways of combining physics-based computational analyses (computational fluid dynamics) and physical experiments (acoustic scale model measurements) into a single design framework.
Dr Dong Li
Dong's work was centred around the idea of using machine learning techniques to capture - and later deploy - design knowledge in the context of automated parageometry generation for multi-disciplinary design optimization processes. Dong used support vector regression to create models of geometry plausibility, which were then used to increase the efficiency of the search for the optimal design by steering the optimizer away from areas where the resulting design would be likely to fail.
Mr. Aditya Deshpande
Manufacturing uncertainty quantification in the context of the design and manufacture of gas turbine engine turbine blades. Lead supervisor Prof Andy J Keane.
Mr. Liam Kelly
Automated design of complex structural components using topology optimization techniques. Lead supervisor Prof Andy J Keane.
Ms. Giang Tran
Bayesian data fusion techniques (in particular Co-Kriging-type emulators) in Earth System modeling. Lead supervisor Dr Kevin Oliver (School of Ocean and Earth Sciences).
Mr. Stephan Langmaak
Activity-based parametric cost models with applications in gas turbine engine design. Lead supervisor Prof James Scanlan.
An innovative proposal to design an aircraft by computer and produce it through 3D printing has won an Engineering PhD student a £3,000 prize.
Investigating the potential of additive manufacturing in the development of UAVs
Stephen Powell's PhD considered the noise reduction potential of installing engines above the wings of airliners
Aircraft Operations and Mechanics of Flight (first year, module leader)
This part one specialist module covers topics such as air law, airline economics, air safety, the regulatory bodies of aviation, aviation weather and airport design in the first semester. In the second semester the fundamentals of flight mechanics are introduced, with an opportunity for the students to experience some fundamental phenomena in the University's flight simulator.
Aircraft Design (third year, module leader)
This is a synthesis module, which pulls together a number of aircraft engineering science topics studied in years 1 and 2 (mechanics of flight, fluid dynamics, wing aerodynamics, structures, design, etc.) and gives student teams the opportunity to undertake their first real aircraft design exercise.
A Flying Laboratory Course, a one week series of lectures and test flights undertaken on board the National Flying Laboratory (G-NFLA), is also part of this module and it is an exciting opportunity for students to experience some key phenomena related the static and dynamic stability of aircraft on board this specially instrumented turboprop (see sidebar for images).
Atmospheric Science UAV - Group Design Project (fourth year, supervisor)
Typically constructed around a real science mission, this GDP challenges a group of fourth year students to design, build and flight test an unmanned air system capable of collecting measurements in the atmosphere.
Dr András Sóbester
Computational Engineering and Design Group
Engineering Centre of Excellence
University of Southampton
Room Number: 176/5005