The Development of Surrogate Model Based MDO Platform-ME DOC
A surrogate model approach has attracted a lot of attention in the field of engineering design, due to its significant reduction in the computational time. The kriging model, one of the most widely used surrogate models, provides efficient tools for global optimization because of its stochastic characteristics. Thus, authors had developed helpful design tools using the kriging model, such as the kriging model based GA (Genetic Algorithm), and had developed ANOVA (ANalysis ()f VAriance), and SOM (Self Organization Map) for data mining. However, the developed tools have inconvenience and inefficiency on the whole design works, because of text user interface and separated tools. Thus, the integrated GUI platform MEDOC (Multi-purpose Engineering Design and Optimization Code) has been developed. The main features of MEDOC are supporting multi-OS (Operating System), easy GUI to use, and high calculation efficiency by multi-threading and JNI (Java Native Interface : multi-language coupling). As an example of MDO (Multidisciplinary Design Optimization), the design of the helicopter rotor blade shape for aeroacoustics performance improvements has been conducted by MEDOC. This application demonstrates the efficiency and the usability of MEDOC.
GUI data mining kriging model GA (Genetic Algorithm) ANOVA (Analysis Of Variance) SOM (Self Organization Map) EI (Expected Improvement) Java JNI (Java Native Interface)
Sanghyun Chae Kwanjung Yee Shigeru Obayashi Shinkyu Jeong
Department of Aerospace Engineering, Pusan National University, Busan, Korea Institute of Fluid Science, Tohoku University, Miyagi, Japan
国际会议
2010 Asia-Pacific International Symposium on Aerospace Technology(2010 亚太航空航天技术研讨会 APISAT 2010)
西安
英文
41-44
2010-09-01（万方平台首次上网日期，不代表论文的发表时间）