Identification of Convection Constants for Electronic Packages Using Modified Genetic Algorithm and Reduced-Basis Method
Name
HPCES012.pdf
Size
238.96 KB
Format
Adobe PDF
Checksum (MD5)
bf4163f552f2f29104906ec2c29a14c1
Author(s)
Yang, Zhenglin
Lee, Jung Hong
Liu, Guirong
Patera, Anthony T.
Lam, Khin Yong
Date Issued
January 2002
Series/Report no.
High Performance Computation for Engineered Systems (HPCES);
Abstract
A new inverse analysis method is presented to identify parameters of heat convection in microelectronic packages. This approach adopts a modified Micro Genetic Algorithm (µGA) in finding the global optimum of parameters. A reduced-basis approach is introduced in the forward heat transfer analysis so as to significantly improve the efficiency in the calculation. Different identification procedures are employed to identify heat convection coefficients of a typical microelectronic package. Comparisons between different algorithms are performed. Results show that the use of the reduced-basis method together with the modified µGA outperforms the conventional GAs significantly. The presented method of coefficient identification is ideal for practical applications. It is efficient enough even for online analysis of both forward and inverse problem.
Subjects
inverse analysis
microelectronic package
modified µGA
parameter identification
reduced-basis method
Persistent DSpace Link
