Biswadip Shome

Design of Experiments Enabled CFD Approach for Optimizing Cooling Fan Performance

Increasing demands on engine power to meet increased load carrying capacity and adherence to emission norms have necessitated the need to improve thermal management system of the vehicle. The efficiency of the vehicle cooling system strongly depends on the fan and fan-shroud design and, designing an optimum fan and fan-shroud has been a challenge for the designer. Computational Fluid Dynamics (CFD) techniques are being increasingly used to perform virtual tests to predict and optimize the performance of fan and fan-shroud assembly. However, these CFD based optimization are mostly based on a single performance parameter. In addition, the sequential choice of input parameters in such optimization exercise leads to a large number of CFD simulations that are required to optimize the performance over the complete range of design and operating envelope. As a result, the optimization is carried out over a limited range of design and operating envelope only. In this paper, a Design of Experiments (DoE) based CFD approach has been used to optimize the fan and fan-shroud design of a cooling pack system. The input design variables of Fan Immersion ratio, Fan to Core distance and Shroud Chamfer Length ratio were considered in this study. The performance output variables of mass flow rate, fan power, and velocity uniformity in the radiator core were predicted. The Central Composite Design (CCD) based DoE approach was used to design the layout of the CFD simulations with the goal of maximizing airflow through the fan, minimize the fan power requirement, and maximize the velocity uniformity in the radiator core. The results from these designed set of CFD simulations were used to generate a response surface that linked the input and output variables with a 2 nd order accuracy transfer function which was then used to optimize the fan and fan-shroud design.

An Enhanced Additive Correction Multigrid Method

ABSTRACT The additive correction multigrid (ACM) method, as originally proposed, does not exhibit good multigrid convergence performance for convection-diffusion equations. A new, enhanced ACM methodology is proposed which uses higher-order restriction operators for explicit formulation of coarse-grid coefficients. The enhanced ACM method offers significant improvement in convergence performance for convection-diffusion equations. The enhanced ACM method is demonstrated to be robust and offers enhanced convergence performance for both convection-dominated and diffusion-dominated cases.