|
气动设计数据的知识提取及其在三维涡轮叶片几何中的应用
Knowledge Extraction from Aerodynamic Design Data and its Application to 3D Turbine Blade Geometries |
|
| |
| 课程网址: |
http://videolectures.net/mla09_graening_kefad/
|
| 主讲教师: |
Lars Graening
|
| 开课单位: |
本田欧洲研究所
|
| 开课时间: |
2009-07-20
|
| 课程语种: |
英语
|
| 中文简介: |
在空气动力学设计优化领域应用数值优化方法通常会导致大量的异构设计数据。虽然通常只调查最有希望的结果并用于推动进一步的优化，但研究整个设计数据集的一般方法很少见。我们的目标是从设计数据中提取有关设计形状和性能之间相互关系的综合知识。提取的知识以可用于指导进一步计算以及手动设计和优化过程的方式准备。
对于复杂空气动力学形状的设计，通常使用不同类型的表示，这使得分析整个设计数据集变得困难甚至不可能。我们建议将设计数据转换为离散的非结构化表面网格，从而实现设计的均匀参数化。这使得可以独立于设计和优化过程中使用的表示来分析设计数据。
在离散的非结构化表面网格的基础上，我们提出了一个位移测量，以分析设计之间的局部差异1。该措施提供有关表面改性的数量和方向的信息。我们最近引入了一个框架2，它使用位移数据结合机器学习的统计方法和技术，从手头的数据集中提供有意义的知识。该框架包括许多用于位移分析，灵敏度分析，降维和规则提取的方法。
为了证明所提出的框架的可行性，我们将所提出的方法应用于由本计算优化运行3产生的小型本田涡扇发动机的超低纵横比跨音速涡轮定子叶片的数据集。已经制定了决策树以生成一组参考预定义刀片设计的设计规则。
通过使用自由形变（DMFFD）4技术的直接操纵来修改涡轮叶片几何形状，已经验证了结果。通过运行计算流体动力学（CFD）模拟计算了变形叶片设计的性能。结果表明，建议的框架提供了合理的结果，可以直接转化为设计修改，以指导设计过程。
|
| 课程简介: |
Applying numerical optimisation methods in the field of aerodynamic design optimisation normally leads to a huge amount of heterogeneous design data. While often only the most promising results are investigated and used to drive further optimisations, general methods for investigating the entire design data set are rare. It is our target to extract comprehensive knowledge from the design data concerning the interrelation between the shape and the performance of the design. The extracted knowledge is prepared in a way that it is usable for guiding further computational as well as manual design and optimisation processes.
For the design of complex aerodynamic shapes it is common to use different kinds of representations, what makes it difficult or even impossible to analyse the entire design data set. We suggest the transformation of the design data into discrete unstructured surface meshes and hence result in a homogeneous parametrisation of the designs. This makes it possible to analyse the design data independent of the representation used during the design and optimization process.
On the basis of discrete unstructured surface meshes we propose a displacement measure in order to analyse local differences between designs1. The measure provides information on the amount and the direction of surface modifications. We recently introduced a framework2 that uses the displacement data in conjunction with statistical methods and techniques from machine learning to provide meaningful knowledge from the dataset at hand. The framework comprises a number of approaches for the displacement analysis, sensitivity analysis, dimensionality reduction and rule extraction.
In order to demonstrate the feasibility of the suggested framework, we applied the proposed methods to a data set of a ultra-low aspect ratio transonic turbine stator blade of a small Honda turbofan engine that resulted from a computational optimisation run3. Decision trees have been formulated to generate a set of design rules which refer to a pre-defined blade design.
The results have been verified by means of modifying the turbine blade geometry using direct manipulation of free form deformation (DMFFD)4 techniques. The performance of the deformed blade design has been calculated by running computational fluid dynamic (CFD) simulations. It is shown that the suggested framework provides reasonable results which can directly be transformed into design modifications in order to guide the design process.
|
| 关 键 词: |
空气动力学; 异构设计数据; 位移分析
|
| 课程来源: |
视频讲座网
|
| 最后编审: |
2019-06-28：cjy
|
| 阅读次数: |
62
|
|
图 书 馆