Aerodynamic Design Optimization & Big Data Analysis Lab

한국어

Researches

Researches



⊙ Design Knowledge Mining and Big Data Analysis


 During the design process using Evolutionary Algorithms, tremendous design data are obtained. From these data, it is possible to extract invaluable knowledge about the design problem. Recently, the process of t is called ‘Data Ming’ or ‘Big Data Analysis’.

 For example, there are 1024 hull forms which are obtained through the low drag hull form design. (obj1, obj2 and obj3 are wave drag coefficient at Fn = 0.22, 0.27 and 0.305, respectively). Each gray circle in Fig. 1 corresponds to each hull form. 



bigdata1.jpg  

Figure 1 Design Exploration Results




 To analyze these data, we applied ANOVA (Analysis of Variance) and Self-Organizing Map (SOM). The result of ANOVA and SOM are shown in Fig. 2.

 


bigdata2.jpg bigdata3.jpg     


(a)  ANOVA

 

 

       bigdata4.jpg


  (b)  SOM


Figure 2 Data Mining with ANOVA and SOM



      Based on the information obtained from the data mining, the following design knowledge can be acquired.

 

(1) A narrower waterline in the bow region (small DV10 and DV11) is a necessary condition for a non-dominated solution.


(2) A wider section shape (large DV4, DV7, DV8, and DV12) is preferable for low-speed performance, while a narrower section shape (small DV4, DV7, DV8, and DV12) is necessary for high-speed performance.



bigdata5.jpg