The multi-disciplinary character at Onera
DOOM
Multi-disciplinary tooled optimization approach
Context and objectives
Today, the design of aeronautic systems relies on the use of increasingly refined modeling tools, but the scope of which is generally limited to either a discipline or a sub-system. In this context, the search for the "best" system is more often the result of trade-offs between teams where each one specializes in a discipline or a sub-system and the overall optimization performance, such as the coming together of models and tools within a systematic or computerized process, and still comes across methodological, organizational and technical difficulties.
The main obstacles are of various types:
- The number of calculations needed to optimize the system in one go is too high for it to be compatible with the industrial design deadlines.
- The capabilities and modeling available in each discipline or sub-system are controlled by specific services that have difficulty in collaborating closely with each other.
- The models used are essentially representations replete with errors of the systems or sub-systems under consideration.
MDO (Multidisciplinary Design Optimization) is intended for all projects in which the notion of a complex system appears and scientific disciplines interact. Within the scope of ONERA's activities, this means aerial vehicles (aircrafts and helicopters, with or without pilot, for civil or military applications), missiles, launchers and satellites. Thus, most of ONERA's large industrial clients are interested by the MDO techniques and would like to profit from an appraisal of the capabilities of the techniques coming within MDO's scope and their applicability according to the problem handled and the defined design objectives.
With the DOOM PRF (Multi-disciplinary tooled optimization approach), Onera aims to build on this appraisal by grouping together the already acquired knowledge and capabilities, and completing them with works focused on proposed techniques (actual MDO methodologies, model reduction methods, optimization algorithms, etc.) and then bringing them together with applications representative in the fields of aircraft and missile preliminary design.
Results
Onera's multi-disciplinary capability is seen strengthened through this methodological appraisal.
In addition to the participants' own skills in the project, this includes editing reference manuals for each of the main scientific themes addressed: MDO formulations, optimization algorithms, model reduction techniques; manuals with targeted experiments pursued in order to highlight the main features. Lastly, the implementation of these techniques on concrete application scenarios led to the production of a user manual proposing an MDO process construction methodology in which each of the main steps is based on the corresponding reference manuals.
The PhD work conducted at the DCPS and experiments on various computing environments using freeware or based on commercial products have also led to the acquisition of capabilities in developing and installing tools as a statement of requirements pertaining to the MDO platform.
The body of knowledge and capabilities built up also includes those of men and women who have been involved in the project: this network is now able to bring the MDO technologies to their colleagues. An operation by Onera aimed at spreading this "MDO culture" internally will be continued following in the project's wake.
-> full size
Figure 1 - MDF and CO methodologies applied to the same test scenario - Implementation and experimentation with ModelCenter
Figure 2 - 3 ways of designing/optimizing a missile with the same objective: 3 original results obtained by applying the same MDO formulation (MDF)
Prospects
Through this project, Onera acquired know-how it can use to help the industrial world to integrate MDO technologies into its system design process. The preferred fields of action are still naturally those belonging to the main systems or sub-systems of aeronautics and space: aircrafts, missiles, helicopters, launchers, satellites, propulsive units, etc.
These skills also strengthen Onera's capabilities in:
- designing (preliminary level) systems in order to help make some choices or to confirm the feasibility of innovative ideas before they are transferred to the industry,
- restoring performances of existing systems,
- highlighting the expected benefits of a new device as being capable of integrating its model within a multi-disciplinary process to ensure that all interactions have been taken into account.
Key words
Multi-disciplinary design - Optimization - Reduction and approximation of models - Distributed design platforms
Top
