By Gabriele Milanese, Johnson Electric GATE
Axial flow fans are widely used within automotive cooling systems that control temperatures of vehicles engine, passengers compartment and auxiliary components.
During last years the design of this classical application is becoming more and more challenging due to the parallel increase of performance requests (e.g. efficiency, noise) and installation constraints.
A fans designer usually starts his work with a preliminary fan geometry definition which is than refined taking into account complex fluid-dynamic, aero-acoustic and other aspects related to the installation environment.
Nowadays these steps can be done by mean of several physical models and different computational tools whose most important feature is the ability to support the designer knowledge in the control of the flow characteristics.
JE Asti work in this area is aimed to continuously update its fans design tools-suite involving self-developed, open-source and commercial software.
Scilab has been recently introduced in the tools-chain, starting with the development of routines which drive and link software devoted to the fan geometry treatment and 3D flow simulations, then using meta-modeling and optimization tool-boxes within the design refinement phase.
Beside conventional techniques, the extensive use of meta-models (Radial Basis Functions) within an innovative interactive approach, greatly contributed to increase the design process level and to reduce the design timing.
In this context the Scilab platform is also providing a significant contribution to the development of a fully integrated graphical environment where different design paths are possible depending on the designer skill and choices.
This work has made possible starting a fruitful interdisciplinary collaboration with the Scilab Enginsoft team (Anna Bassi, Manolo Venturin).
Co-signed paper by Gabriele Milanese – Johnson Electric GATE and Ubaldo Cella – Design Methods
Axial flow fans, Automotive, Meta-models
Presentation not available for confidentiality reasons