Automatic Optimization of Air Conduct Design Using Experimental Data and Numerical Results

Abstract

The development of a new product is a process that goes through several phases, each of them requiring the ca- pitalization of data, information, knowledge and experiences gathered from previous projects. This knowledge is kept by a limited number "Experts", and is not necessarily captured in a practical, reusable way, which translates into a loss of time and a delay for the projects. Our approach involves exporting the client's specifications from a PLM platform and con- structing a generic product, an automotive air conduct, from them. This generic product is then fed through an optimization cycle to determine the best position and geometrical shape for all its elements. This multi-criteria optimization requires several constraints and objective functions in order to improve the part quality and process reliability needed expected by the Original Equipment Manufacturer (OEM). The need to have a design process less dependent on personal "expert" knowledge was the driving idea behind this work. The automatic optimization of the design of an air conduct still remains a challenge and the purpose of this work is to contribute to the development of an automatic design tools capable of doing this.

Keywords

• design for manufacturing
• optimization
• knowledge based engineering
• simulation