USE OF GENETIC ALGORITHMS IN THERMAL PROPERTY ESTIMATION: PART II - SIMULTANEOUS ESTIMATION OF THERMAL PROPERTIES

Abstract

This two-part study is on the use of genetic algorithms (GAs) to design experiments and develop estimation methodologies for the determination of thermal properties; Part I is focused on the development of an improved GA, called an extended elitist genetic algorithm (EEGA), and on the implementation of this algorithm to optimize experimental designs used in thermal property estimation, while Part II is directed toward the application of this algorithm to the simultaneous estimation of thermal properties. In Part II the EEGA is used to minimize a least squares objective function containing calculated and measured temperatures. While the EEGA was shown to be an effective strategy for the optimization of experiments in Part I, its potential for use in the estimation of thermal properties is shown here in Part II through the use of case studies. In addition, the effect of the choice of the criterion used to optimize the experimental designs on the accuracy of the property estimates was analyzed for one of the case studies. These case studies include the simultaneous estimation of two, three, and five thermal properties, with some of them being highly correlated. Correlation or near-correlation among simultaneously estimated properties can be a limiting factor of commonly used gradient-based methods. The results from the analysis of the case studies demonstrate that the proposed GA is a useful toot in the simultaneous estimation of correlated thermal properties.

Notes

Address correspondence to Professor Elaine P. Scott, Department of Mechanical Engineering, Virginia Polytechnic Institute and State University, Randolph Hall, Blacksburg, VA 24061-0238. E-mail: [email protected]