open menu
MENU

CAP (Computer Aided Principle)

CAP tool unveils why the optimum is optimal: this innovative analytic method, proposed by Mr. Qiang Yu, Assistant Professor at Faculty of Engineering at Yokohama National University, supports the extraction of fundamental design principles.

The Computer Aided Principle (CAP) within modeFRONTIER® is an innovative analytic method that supports the feasibility evaluation of product concept features. Once the optimal configuration of one component has been identified, the feasibility of another component may be compromised further along in the design process.

The CAP method allows the user to pull out important information regarding the basic principles underlying the physical behaviour of the design, even with strong non-linearity cases. Extracting relevant figures like correlative physical parameters helps in understanding the essence of the phenomena involved. Even when a modification of a component’s design occurs during the detailed design step, such knowledge will allow engineers to develop a comprehensive design plan and avoid changing aspects not related to the component.

 

How it works

While conventional sensitivity and contribution analyses focus on effects of a single variable, CAP sheds light on the relationship between variables and their values by focusing on the effects, (called trends), of variable combinations.

The CAP approach first considers different output categories in order to determine a input behaviour pattern. Starting from the traditional definition of the design problem, the parametric study and the optimization cycle, CAP tool extracts basic design behaviour patterns following the steps below:

  • Hierarchical cluster analysis: classification of designs on the basis of characteristic values, providing a graphic representation of trends and variations affecting those trends;
  • Determining the Principle: building from the visual representations, CAP helps designers interpret the physical meaning of the trends underlying the combinations of variables analyzed.

Sign in