The aim of the project is to improve shape design in the aeronautical industry by developing mathematical and engineering tools using supercomputers. In situations where the numerical simulations of aircraft by computers are in current use, optimization methods provide a powerful aid to engineers to improve their designs.
Subsonic and transonic airfoil and wing optimization procedures have been developed to treat optimization problems namely drag minimization, lift maximization or target pressure recovery with possible non linear constraints to satisfy geometrical requirements and control aerodynamic characteristics in off-design conditions. The most innovative outcomes of this project concern progress accomplished in design with Euler solvers, fast one pass inverse methods for rotational flows, parametrisation of non linear surfaces, hierarchical multi level method for control variables and automatic adaptive remeshing.
Results obtained from a workshop indicate quite large discrepancies both in design due to the quality of the flow analysis solver and the parametrisation of the shape and also in efficiency with the choice of the optimiser. These comparisons can provide useful guidelines for choosing optimization or inverse methods with inviscid potential or Euler flows.
It is clear that access to practical 3-D applications including viscous effects in an industrial environment requires still an important effort. In particular cost effective designs of good quality achieved with the above methodologies and their associated algorithms should have to be implemented in a near future on parallel architectures.
The partners have experimented with optimum design methods in the past. Now that computers are more powerful, it is believed that systematic 2-dimensional and 3-dimensional optimization of aircraft configurations can be implemented by industry. The aim of this study is to validate this point and to develop the tools necessary for this implementation.
The study will cover optimal design for the compressible inviscid flow equations produced by advanced numerical techniques. Particular attention will be devoted to the development of fast algorithms adapted to the problem (which usually contains several non-linear constraints), using modern minimization algorithms, to obtain reliable adaptive mesh generators when the shape of the configuration being studied is changed and to write user-friendly interactive graphic tools to monitor the generated shapes. Simple engineering methods will be also investigated as inverse problem techniques. At the end of this project, a workshop will take place to validate the developed methodologies on a few test cases including wing fuselage design, multi-airfoil design and air intake design.
Funding SchemeCSC - Cost-sharing contracts
8300 AD Emmeloord