Advanced nozzle extension design methodology
Nozzle design methodology
Project duration: 2012 to 2013
Project manager: Matěj Lepš, Ph.D.
The project ANED-M was divided into three separate tasks. The first one - statistical fitting of model parameters to measurements - dealt with the probabilistic description of uncertain material parameters based on uncertain experimental data. The objective of this task was to develop and implement a methodology which takes into account expert knowledge, i.e. prior information and to update this information by using noisy experimental data. More specifically, the methodology was applied to the parameter identification of a visco-plastic constitutive material model based on a set of data observed during five different loading tests. The result was a probabilistic description of material properties, which took into account all the available information and allowed for more reliable simulation of structural performance. The partner institution will implement the obtained data as well as the developed methodology into the future launcher design process.
The second task entitled - identification of worst case scenario - dealt with identification of a parameter combination leading to the structural failure. In such a case, parameters comprised all the material properties and geometry or loading parameters of a nozzle extension design and thus the problem was high-dimensional. Moreover, the analysis included the simulation of nozzle extension taking into account different types of failure and was computationally very demanding. The developed methodology allowed to identify some of these worst case parameter combinations which provided an invaluable insight into reliability of the structure.
The last task dealt with geometry optimisation and its goal was to propose a methodology for optimisation of nozzle extension geometry so as to minimise its weight and maximise its reliability with respect to the prescribed uncertain loading. The process was thus multi-objective, high-dimensional and computationally very challenging, but opened the possibility for important savings and technological improvements.
About the Department of mechanics
The department is formed by specialists in the fields of solid, soil and structural mechanics, computer science, applied mathematics, reliability of structures, multi-scale modelling or biomechanics. The scientific group directly involved in the project focuses mainly on solving optimisation and inverse problems in civil engineering or materials science, structural reliability or modelling of heterogeneous materials. This includes development of probabilistic methods from the field of Bayesian statistics with applications in inverse problems under uncertainties, bio-inspired methods like evolutionary algorithms or artificial neural networks for solution of ill-posed or multimodal optimisation problems and different kinds of surrogate models for acceleration of optimisation processes or reliability analysis.
What would you name as main benefits of the project to you and your company?
“The project gave our team an exceptional opportunity to utilize our theoretical knowledge on optimisation and identification problems, and to test the robustness and versatility of the developed methods when applied to complex problems of engineering practice. In addition, establishing cooperation with key partners in the field of future launcher development is of an immense importance to our department. It opened the possibilities for broader cooperation currently resulting into two new projects starting in December 2013.”