Optimal Design for Noise Reduction of the Engine Cooling Fan in Automobile Using DACE Model(Abstract)

摘要

The objective of this research is optimal design for noise reduction of the engine cooling fan in automobile using DACE (Design and Analysis of Computer Experiments). For this purpose, first of all this paper propose a simulation model that will allow design engineers to estimate the fan noise based on the Wright's and Wu's theories, given the characteristic dimensions and working conditions of the fan. Determining the total sound pressure level is not easy to characterize the noise performance of the axial flow fan, because the fan noise spectrum of components have both narrow and broadband noise according to noise source. Within the narrow components, discrete frequency noise predominates with axial flow fans and is periodic in character, consisting of the fundamental blade passage frequency and many higher harmonics. Broadband noise is due to vortex shedding from the leading and trailing edges of a fan. This noise is broadband in character, ie, with no discrete frequency components, which for typical axial flow fans used in road vehicles is of relatively low intensity. The components are caused by different generated mechanism. Therefore, in order to improve the noise prediction performance of the fan, it is necessary to be able to predict the simulation model based on the fan characteristic parameters. A simulation model has been developed to estimate the noise spectrum with Wright's and Wu's methods. To validate this simulation model, the calculated noise spectra are compared with the measured ones of a cooling fan. And it is difficult to reduce the time and the cost for the development of the low noise fan system because the method to reduce the fan noise has been applied by means of trial and error after the design of the engine compartment is finished. Further, experimental design methods are inadequate in computer experiments because these are considered along with the error term. Therefore, we consider Kriging based on DACE model to improve the performance of the system with a low cost and time savings. In this study, Kriging based on DACE model is proposed for an optimal design scheme to reduce noise of the engine cooling fan. The experimental design we used is the orthogonal arrays. For Kriging, non-linear optimization problems have to be solved, that is, one is for estimating an optimal correlation parameter, and the other is for finding an optimal level for the Kriging estimate. In this study, meta-heuristic techniques such as genetic algorithms and simulated annealing are applied to solve the optimization problems. So, we can find an optimal level of the engine cooling fan by optimization of the Kriging.