A robust and efficient algorithm is presented to build multiresolution models (MRMs) of arbitrary meshes without requirement of subdivision connectivity. To overcome the sampling difficulty of arbitrary meshes, edge c...A robust and efficient algorithm is presented to build multiresolution models (MRMs) of arbitrary meshes without requirement of subdivision connectivity. To overcome the sampling difficulty of arbitrary meshes, edge contraction and vertex expansion are used as downsampling and upsampling methods. Our MRMs of a mesh are composed of a base mesh and a series of edge split operations, which are organized as a directed graph. Each split operation encodes two parts of information. One is the modification to the mesh, and the other is the dependency relation among splits. Such organization ensures the efficiency and robustness of our MRM algorithm. Examples demonstrate the functionality of our method.展开更多
在汽车气动外形优化设计中,往往需要大量的高精度CFD数据作为支撑。然而,高精度CFD数据获取难度大、成本高。为了缓解汽车气动优化设计中气动特性评估精度和效率之间的矛盾,根据迁移学习与数据融合的思想,提出了一种基于多精度深度神经...在汽车气动外形优化设计中,往往需要大量的高精度CFD数据作为支撑。然而,高精度CFD数据获取难度大、成本高。为了缓解汽车气动优化设计中气动特性评估精度和效率之间的矛盾,根据迁移学习与数据融合的思想,提出了一种基于多精度深度神经网络(multi-fidelity deep neural network, MFDNN)的汽车外形优化设计方法,以减少优化设计中所需的高精度数据个数,从而有效提升优化速度、降低优化成本。将所发展的优化方法应用于快背式MIRA标准模型减阻优化设计中,优化结果表明,该方法能够充分融合不同精度数据所蕴含的知识,加速气动外形优化进程,提升优化效率。以收敛用时作为评价指标,在取得相近或更优优化结果的前提下,基于多精度神经网络的优化框架的收敛速度是基于单精度神经网络的离线优化框架的5.85倍,是基于单精度神经网络的在线优化框架的2.81倍。展开更多
Network economy had changed manufacturing environme nt at all. Open global market offer more choice to customer, and it become changea ble and unpredictable as consumers’ needs become more and more characteristic an ...Network economy had changed manufacturing environme nt at all. Open global market offer more choice to customer, and it become changea ble and unpredictable as consumers’ needs become more and more characteristic an d diversified. Various new technology coming forth and application accelerate th e rapid change of the market. The manufacturing enterprises were compelled t o change their strategy by the variability of the market, and time has been put to the all-important place. There is a need driven by the market to set up a ne twork design and manufacturing mode which have rapid market responsiveness. In order to meet the need for network manufacturing, the organization and manage ment of manufacturing enterprise need a completely innovation, next generation o f manufacturing system must have the character such as digitization, flexibility , agility, customization and globalization and so on. As for an enterprise in au to industry, how to gather together the orders through the distribution, and rap id produce the product which can meet the customer’s need, it is the key that th e contemporary enterprises succeed in the competitive market. The competitive market requires rapid product development. Close cooperation amo ng the designers will accelerate the product development by shortening the devel opment cycle, improving the product quality and reducing the investment. It has been emphasized in the methodology of concurrent engineering (CE). But sometimes those partners are distributed in the world, so there is a need for an importan t technology contribution to collaborative engineering, and supporting distribut ed designers for rapid product development. This paper focuses on a collaborative design system: Product Digit Collaborative Design System (PDCDS). The solution of PDCDS can make it more efficient and rel iable to visit teledata as well as we can get it from local database. It will be ease to get the newest design process information aided by PDCDS, and it will h ave higher efficiency by collaborative work. Comparing with other traditional Pr oduct Data Management (PDM) software system, PDCDS have some new characters such as group, dynamicness, synchronization or asynchronism working mode, and the hi story recorder is needed, and it also surport Webservice.展开更多
文摘A robust and efficient algorithm is presented to build multiresolution models (MRMs) of arbitrary meshes without requirement of subdivision connectivity. To overcome the sampling difficulty of arbitrary meshes, edge contraction and vertex expansion are used as downsampling and upsampling methods. Our MRMs of a mesh are composed of a base mesh and a series of edge split operations, which are organized as a directed graph. Each split operation encodes two parts of information. One is the modification to the mesh, and the other is the dependency relation among splits. Such organization ensures the efficiency and robustness of our MRM algorithm. Examples demonstrate the functionality of our method.
文摘在汽车气动外形优化设计中,往往需要大量的高精度CFD数据作为支撑。然而,高精度CFD数据获取难度大、成本高。为了缓解汽车气动优化设计中气动特性评估精度和效率之间的矛盾,根据迁移学习与数据融合的思想,提出了一种基于多精度深度神经网络(multi-fidelity deep neural network, MFDNN)的汽车外形优化设计方法,以减少优化设计中所需的高精度数据个数,从而有效提升优化速度、降低优化成本。将所发展的优化方法应用于快背式MIRA标准模型减阻优化设计中,优化结果表明,该方法能够充分融合不同精度数据所蕴含的知识,加速气动外形优化进程,提升优化效率。以收敛用时作为评价指标,在取得相近或更优优化结果的前提下,基于多精度神经网络的优化框架的收敛速度是基于单精度神经网络的离线优化框架的5.85倍,是基于单精度神经网络的在线优化框架的2.81倍。
文摘Network economy had changed manufacturing environme nt at all. Open global market offer more choice to customer, and it become changea ble and unpredictable as consumers’ needs become more and more characteristic an d diversified. Various new technology coming forth and application accelerate th e rapid change of the market. The manufacturing enterprises were compelled t o change their strategy by the variability of the market, and time has been put to the all-important place. There is a need driven by the market to set up a ne twork design and manufacturing mode which have rapid market responsiveness. In order to meet the need for network manufacturing, the organization and manage ment of manufacturing enterprise need a completely innovation, next generation o f manufacturing system must have the character such as digitization, flexibility , agility, customization and globalization and so on. As for an enterprise in au to industry, how to gather together the orders through the distribution, and rap id produce the product which can meet the customer’s need, it is the key that th e contemporary enterprises succeed in the competitive market. The competitive market requires rapid product development. Close cooperation amo ng the designers will accelerate the product development by shortening the devel opment cycle, improving the product quality and reducing the investment. It has been emphasized in the methodology of concurrent engineering (CE). But sometimes those partners are distributed in the world, so there is a need for an importan t technology contribution to collaborative engineering, and supporting distribut ed designers for rapid product development. This paper focuses on a collaborative design system: Product Digit Collaborative Design System (PDCDS). The solution of PDCDS can make it more efficient and rel iable to visit teledata as well as we can get it from local database. It will be ease to get the newest design process information aided by PDCDS, and it will h ave higher efficiency by collaborative work. Comparing with other traditional Pr oduct Data Management (PDM) software system, PDCDS have some new characters such as group, dynamicness, synchronization or asynchronism working mode, and the hi story recorder is needed, and it also surport Webservice.
